This paper addresses the evolving landscape of electricity markets in Europe, with a focus on the integration of Renewable Energy Communities as introduced by the Renewable Energy Directive 2018/2001. Residents within a postal code area are highly incentivized to join a community, which enables them to exchange energy among themselves at lower procurement costs. Thereby, energy management systems optimize the operation of respective energy systems, with electrical load forecasting playing a key role. Given that prosumers may switch between communities on a daily basis, electricity demands within these groups will vary, leading to data that is non-stationary, discontinuous as well as non-identical and independently distributed. To encounter this issue, we propose a sophisticated forecasting model that applies federated learning, using informations from various distributed communities to learn domain-invariant features. To achieve this, we initially utilize synthetic electrical load time series at district level and aggregate them to profiles of Renewable Energy Communities with dynamic portfolios. Subsequently, we develop a forecasting model that accounts for the composition of residents of a Renewable Energy Community, adapt data pre-processing in accordance with the time series process, and detail a federated learning algorithm that incorporates weight averaging and data sharing. Following the training of various experimental setups, we ultimately evaluate their effectiveness. The findings suggest that our proposed framework is capable of effectively forecast non-stationary and discontinuous time series and that it can be applied to new, unseen data through the integration of knowledge from multiple sources.

Convolutional Neural Networks (CNNs) present drawbacks for modeling geometric transformations, such as scaling and rotation, caused by the convolution operation’s locality. Deformable convolution (DCON), a mechanism that substitutes standard convolution, increasing the receptive field to capture relevant features, is a promising approach to solve this drawback and improve the robustness of CNNs. However, the optimal way to replace the standard convolution with its deformable counterpart in a CNN model is unclear. In this study, we clarify this aseveration by conducting several experiments using deformable convolutions applied in the layers that conform a small four-layer CNN model. We also use deformable convolutions on the four-layers of several ResNet CNNs with depths 18, 34, 50, and 101. The models were tested in binary balanced classes with 2D data for image classification: Cats & Dogs, EyesPACS, Spyders & Chickens, and Shapes. After this testing, we evaluated DCON in 3D data for action recognition: UCF101 and Human2 (a dataset we compiled to control movement, clothing and background). The contribution of this research lies in a guideline to use DCON. It can be summarized as follows: if DCON is used on the first layers of the proposal of model (with simple features), the computational resources expressed as the quantity of Flops will tend to increase and produce bigger misclassification than the standard CNN. However, if the DCON is used at the end layers, the quantity of Flops used in the training and testing will decrease, and the classification accuracy will improve by up to 20% about the base model. Moreover, it gains robustness when using deformable convolutions because it can adapt to the region of interest. Also, the best kernel size of the DCON is three. It showed better results than size five. In the last case, the quantity of Flops increase quadratically, but their performance does not increase significantly. With these results, we propose a guideline to use the DCON and contribute to understanding the impact of DCON on the robustness of CNNs.

Deep learning has been widely used in various tasks such as computer vision, natural language processing, and predictive analysis, recommendation systems in the past decade. However, practical scenarios often lack labeled data, posing challenges for traditional supervised methods. Semi-supervised classification methods address this by leveraging both labeled and unlabeled data to enhance model performance, but they face challenges in effectively utilizing unlabeled data and distinguishing reliable information from unreliable sources. This paper introduces ReliaMatch, a semi-supervised classification method that addresses these challenges by using a confidence threshold. It incorporates a curriculum learning stage, feature filtering, and pseudo-label filtering to improve classification accuracy and reliability. The feature filtering module eliminates ambiguous semantic features by comparing labeled and unlabeled data in the feature space. The pseudo-label filtering module removes unreliable pseudo-labels with low confidence, enhancing algorithm reliability. ReliaMatch employs a curriculum learning training mode, gradually increasing training dataset difficulty by combining selected samples and pseudo-labels with labeled data. This supervised approach enhances classification performance. Experimental results show that ReliaMatch effectively overcomes challenges associated with the underutilization of unlabeled data and the introduction of error information, outperforming the pseudo-label strategy in semi-supervised classification.

This paper introduces a novel framework for zero-shot learning (ZSL), i.e., to recognize new categories that are unseen during training, by distilling knowledge from foundation models. Specifically, we first employ ChatGPT and DALL-E to synthesize reference images of unseen categories from text prompts. Then, the test image is aligned with text and reference images using CLIP and DINO. Finally, the predicted logits are aggregated according to their confidence to produce the final prediction.Experiments are conducted on multiple datasets, including CIFAR-10, CIFAR-100, and TinyImageNet. The results demonstrate that our model can significantly improve classification accuracy compared to previous approaches, achieving AUROC scores above 96\% across all test datasets. Our code is available at https://github.com/1134112149/MICW-ZIC.

Gastric cancer is the fifth most common and fourth deadliest cancer worldwide, with a bleak 5-year survival rate of about 20%. Despite significant research into its pathobiology, prognostic predictability remains insufficient due to pathologists’ heavy workloads and the potential for diagnostic errors. Consequently, there is a pressing need for automated and precise histopathological diagnostic tools. This study leverages Machine Learning and Deep Learning techniques to classify histopathological images into healthy and cancerous categories. By utilizing both handcrafted and deep features and shallow learning classifiers on the GasHisSDB dataset, we conduct a comparative analysis to identify the most effective combinations of features and classifiers for differentiating normal from abnormal histopathological images without employing fine-tuning strategies. Our methodology achieves an accuracy of 95% with the SVM classifier, underscoring the effectiveness of feature fusion strategies. Additionally, cross-magnification experiments produced promising results with accuracies close to 80% and 90% when testing the models on unseen testing images with different resolutions.

The dynamics of surface and sub-surface water events can lead to slope instability resulting in anomalies such as slough slides on earthen levees. Early detection of these anomalies by a remote sensing approach could save time versus direct assessment. We have implemented a supervised Mahalanobis distance classification algorithm for the detection of slough slides on levees using complex polarimetric Synthetic Aperture Radar (polSAR) data. The classifier output was followed by a spatial majority filter post-processing step which improved the accuracy. The effectiveness of the algorithm is demonstrated using fully quad-polarimetric L-band Synthetic Aperture Radar (SAR) imagery from the NASA Jet Propulsion Laboratory’s (JPL’s) Uninhabited Aerial Vehicle Synthetic Aperture Radar (UAVSAR). The study area is a section of the lower Mississippi River valley in the southern USA. Slide detection accuracy of up to 98 percent was achieved, although the number of available slides examples was small.

Commercial banks usually classify customers according to their credit reports when making loans. In this study, we put our focus on classifying customers based on their credit reports from the People's Bank of China (PBC). Since there are no target labels of users in the credit report of the People's Bank of China, we put forward the fuzzy clustering method for the initial label, and then Construct ant colony search to optimize intelligent recognition. Finally, this study uses SVM, BP neural network, and random forest to classify users and compare their results. The research re-sults indicate that using ant colony clustering algorithm and random forest for classification is the most effective method with the PBC credit reports.

Hyperspectral images (HSI) contain abundant spectral information. Efficient extraction and utilization of this information for image classification remain prominent research topics. Previously, hyperspectral classification techniques primarily relied on statistical attributes and mathematical models of spectral data. Deep learning classification techniques have recently been extensively utilized for hyperspectral data classification, yielding promising outcomes. This study proposes a deep learning approach that uses polarization feature maps for classification. Initially, the polar coordinate transformation method is employed to convert the spectral information of all pixels in the image into spectral feature maps. Subsequently, the proposed Deep Context Feature Fusion Network (DCFF-NET) is utilized to classify these feature maps. The model is validated using three open-source hyperspectral datasets: Indian Pines, Pavia University, and Salinas. The experimental results indicated that DCFF-NET achieves excellent classification performance. Experimental results on three public HSI datasets demonstrate that the proposed method accurately recognizes different objects with an overall accuracy (OA) of 86.68%, 94.73%, and 95.14% based on the pixel method, and 98.15%, 99.86%, and 99.98% based on the pixel-patch method.

Climate change is increasing air temperatures and altering the precipitation and hydrological re-gime on a global scale. Challenges arise when assessing the impacts of climate change on the local scale for water resources management proposes, especially for low mountain headwater catch-ments that not only serve as important water towers for local communities, but also have distinct hydrological characteristics. Until now, no low flow or hydrological drought studies have been carried out on the Lauter River. This study is unique in that it compares the Lauter River, a transboundary Rhine tributary with a nearby station on the River Rhine just below its confluence at the French-German border. The Lauter catchment is a mostly natural, forested catchment, however, its water course has been influenced by past and present cultural activities. Climate change disturbances cascade through the hydrologic regime right down to the local scale. As we are expecting more low flow events, the decrease in water availability could cause conflicts be-tween different water user groups in the Lauter catchment. However, the choice of different methods for identifying low flow periods may cause confusion for local water resources manag-ers. Using flow rate time series of the Lauter River between 1956 and 2022 we compare for the first time three low flow identification methods: the variable threshold method (VT), the fixed threshold method (FT), and the Standardized Streamflow Index (SSI). Similar analyses are applied and compared to the adjacent Maxau station on the Rhine River for the same time period. This study aims at 1) interpreting the differences amongst the various low flow identification methods and 2) revealing the differences in low flow characteristics of the Lauter catchment compared to the Rhine River. It appears that the FT reacts faster to direct climate or anthropogenic impacts whereas VT is more sensitive to indirect factors such as decreasing subsurface flow which is typical for small headwater catchments, such as the Lauter where flow dynamics react faster to flow disturbances. Abnormally low flow during the early Spring in tributaries such as the Lauter can help predict low flow conditions of the Rhine River during the following half year and espe-cially for the summer. The results could facilitate Early Warning of hydrological droughts and drought management for water users in the Lauter catchment and further downstream along some of the Rhine.

Non-linear activation functions are one of the main parts of deep neural network architectures. The choice of the activation function can affect model speed, performance and convergence. Most popular activation functions don't have any trainable parameters and don't alter during the training. We propose different activation functions with and without trainable parameters. Said activation functions have a number of advantages and disadvantages. We'll be testing the performance of said activation functions and comparing the results with widely known activation function ReLU. We assume that the activation functions with trainable parameters can outperform functions without ones, because the trainable parameters allow the model to "select'' the type of each of the activation functions itself, however, this strongly depends on the architecture of the deep neural network and the activation function itself.

Polarimetric Synthetic Aperture Radar (PolSAR) images encompass valuable information that can facilitate extensive land cover interpretation and generate diverse output products. Extracting meaningful features from PolSAR data poses challenges distinct from those encountered in optical imagery. Deep Learning (DL) methods offer effective solutions for overcoming these challenges in PolSAR feature extraction. Convolutional Neural Networks (CNNs) play a crucial role in capturing PolSAR image characteristics by leveraging kernel capabilities to consider local information and the complex-valued nature of PolSAR data. In this study, a novel three-branch fusion of complex-valued CNN named Shallow to Deep Feature Fusion Network (SDF2Net) is proposed for PolSAR image classification. To validate the performance of the proposed method, classification results are compared against multiple state-of-the-art approaches using the Airborne Synthetic Aperture Radar (AIRSAR) datasets of Flevoland, San Francisco, and ESAR Oberpfaffenhofen dataset. The results indicate that the proposed approach demonstrates improvements in OA, with 1.3% and 0.8% enhancement for the AIRSAR datasets and 0.5% improvement for the ESAR dataset. The analyses conducted on the Flevoland data underscore the effectiveness of the SDF2Net model, revealing a promising OA of 96.01% even with only 1% sampling ratio.

Deep neural networks with powerful auto-optimization tools have been widely applied in various research fields, e.g., NLP and computer vision. However, existing neural network architectures typically are constructed using different inductive biases, e.g., preconceptions, expecting to decrease parameter search space during training, reduce computational cost or introduce expert knowledge in the neural network design. As an alternative, Multilayer Perceptron (MLP) provides much better freedom for exploration, has a lower inductive bias than convolutional neural networks (CNNs), and offers good flexibility in learning complex patterns. Even though, such neural architectures are commonly built in a flat Euclidean space, which is not necessarily the optimal space for any data, and is especially not good for modeling hierarchical correlations. Hyperbolic neural networks (HNNs) have gained attention for their ability to capture hierarchical structures present in complex data types like graphs. Recently, there has been an increasing interest to extend HNNs to computer vision tasks, motivated by the observations that images possess rich hierarchical relations. However, this is generally applied by employing a Euclidean backbone for learning higher-level semantic representations and only incorporating a hyperbolic classifier for classification, which, we argue, does not make full use of the advantage of hyperbolic space. Considering the recovery of the attention-free Multilayer Perceptron(MLP), in this paper, we extend it to non-Euclidean space and propose a novel architecture, named Hyperbolic Res-MLP (HR-MLP), that leverages fully hyperbolic layers to learn feature embeddings and perform image classification in an end-to-end fashion. With the help of the proposed Lorentz cross-patch and cross-channel layers, we can directly perform operations in the hyperbolic domain with fewer parameters, making it faster to train and providing comparatively better performance than its Euclidean counterpart. Experiments on CIFAR10, CIFAR100, and MiniImageNet demonstrate a comparable and superior performance when compared to Euclidean baselines. Our code is available at (https://github.com/Ahmad-Omar-Ahsan/HR-MLP)

Manual interpretation of these huge amounts of image volumes are susceptible to inter-reader variability and human error. Thus, accurate automated CAD scheme is highly desirable in clinical pathological diagnosis. In this research, plethora of machine learning paradigms (e.g. feature extraction, dimensionality reduction and supervised classification methods) were explored, evaluated, compared and analyzed to identify the optimal pathway for brain MR images (normal vs neoplastic) binary classification task. External validation dataset was used to test the generalizability of the optimal predictive models implemented. Relevant and informative features were selected to construct cross-validated decision tree and eventually simple rule set was built based on the decision tree. The experimental results show that almost all pattern recognition paradigms achieve high accuracy with careful selection of number of attributes. LDA+ELM with 55 features are the optimal pipelines which achieve perfect classification when training and test data are of same source; and achieving (accuracy=97.5%, AUC=0.989, sensitivity=95% and specificity=100%) under balanced test dataset; (accuracy=99.5%, AUC=0.988, sensitivity=95% and specificity=100%). Cross-validated decision tree model also shows comparable performance: accuracy=98.8%, AUC=99.1%, sensitivity=99.6% and specificity=98.2%. Three highly relevant and robust attributes are visualized and selected for construction of decision tree models and finally a rule sets are read directly off the decision tree. This rule sets can potentially serve as fast and accurate classification algorithm.

White matter hyperintensity (WMH) has been considered the primary biomarker from small-vessel cerebrovascular disease to Alzheimer’s disease (AD) and has been reported for its correlation of brain structural changes. To perform WMH related analysis with brain structure, both T1-weighted (T1w) and (Fluid Attenuated Inversion Recovery(FLAIR) are required. However, in a clinical situation, it is limited to obtain 3D T1w and FLAIR images simultaneously. Also, the most of brain segmentation technique supports 3D T1w only. Therefore, we introduced the semi-supervised learning method that can perform brain segmentation using FLAIR image only. Our method achieved a dice overlap score of 0.86 for brain tissue segmentation on FLAIR, with the relative volume difference between T1w and FLAIR segmentation under 4.8%, which is just as reliable as the segmentation done by its paired T1w image. We believe our semi-supervised learning method has a great potential to be used to other MRI sequences and provide encouragement to people who seek brain tissue segmentation from a non-T1w image.

In Indonesia, types of meat are still identified manually due to increasing prices for beef and buf-falo ahead of Idul Fitri. Traders mix beef with pork to thwart the beef scam. Haar wavelet and GLCM, GLCM with angles of 0°, 45°, 90°, and 135°, as well as matrices using contrast, correlation, energy, homogeneity, and entropy, are used in feature extraction. The following are the results of meat image classification testing using k-NN, Haar wavelet, and GLCM: The k-NN algorithm shows superiority in identifying fresh (99%), frozen (99%), and rotten (96%) meat texture images with the most fantastic accuracy results in every situation. Meanwhile, GLCM routinely provides good results, especially regarding the texture image of fresh meat (183.21) and rotten meat (115.79). However, despite delivering less throughput than k-NN and GLCM, Haar wavelets are still helpful, especially when dealing with fresh meat texture images (89,96).

Hyperspectral image (HSI) classification is a mechanism of analyzing differentiated land cover in remotely sensed hyperspectral images. In the last two decades, a number of different types of classification algorithms have been proposed for classifying hyperspectral data. These algorithms include supervised as well as unsupervised methods. Each of these algorithms has its own limitations. In this research, three different types of unsupervised classification methods are used to classify different datasets i-e Pavia Center, Pavia University, Cuprite, Moffett Field. The main objective is to assess the performance of all three classifiers K-Means, Spectral Matching, and Abundance Mapping, and observing their applicability on different datasets. This research also includes spectral feature extraction for hyperspectral datasets.

Popular geo-computer vision works make use of aerial imager with sizes ranging from 64 × 64 to 1024 × 1024 pixels without any overlap, although the learning process of deep learning models can be affected by the reduced semantic context or the lack of information near image boundaries. In this work, the impact of three image sizes (256 × 256, 512 × 512, and 1024 × 1024 pixels) and two image overlap levels (no overlap and 12.5% overlap) on the performance of road classification models was statistically evaluated. For this, two convolutional neural networks used in various tasks of geospatial object extraction were trained (using the same hyperparameters) on a large dataset (containing aerial image data covering 8650 km² of the Spanish territory that was labelled with binary road information) under twelve different scenarios, each scenario featuring a different combination of tile size and overlap. To assess their generalisation capacity, the performance of all resulting models was evaluated on a data from novel areas covering approximately 825 km². The performance metrics obtained were analysed using appropriate descriptive and inferential statistical techniques to evaluate the impact of the performance at distinct levels of the fixed factors (tile size, tile overlap, neural network architecture). Statistical tests were applied to study the main and interaction effects of the fixed factors on the performance. A significance level of 0.05 was applied to all the null hypothesis tests. The results were highly significant for the main effects (p-values lower than 0.001), while the two-way and three-way interaction effects among them had different levels of significance. The results indicate that the training of road classification models on images with a higher tile size (more semantic context) and tile overlap (additional border context and continuity) significantly impacts their performance. The best model was trained on a dataset featuring tiles with a size of 1024 × 1024 pixels and a 12.5% overlap and achieved a loss value of 0.0984, an F1 score of 0.8728, and an ROC-AUC score of 0.9766, together with an error rate of 3.5% on the test set.

State-of-the-art remote sensing scene classification methods employ different Convolutional Neural Network architectures for achieving very high classification performance. A trait shared by the majority of these methods is that the class associated with each example is ascertained by examining the activations of the last fully connected layer, and the networks are trained to minimize the cross-entropy between predictions extracted from this layer and ground-truth annotations. In this work, we extend this paradigm by introducing an additional output branch which maps the inputs to low dimensional representations, effectively extracting additional feature representations of the inputs. The proposed model imposes additional distance constrains on these representations with respect to identified class representatives, in addition to the traditional categorical cross-entropy between predictions and ground-truth. By extending the typical cross-entropy loss function with a distance learning function, our proposed approach achieves significant gains across a wide set of benchmark datasets in terms of classification, while providing additional evidence related to class membership and classification confidence.

Image segmentation and classification still represent an active area of research since no universal solution can be identified. Established segmentation algorithms like thresholding are problem specific, treat well the easy cases and mostly relied on single parameter i.e intensity. Machine learning approaches offer alternatives where predefined features are combined into different classifiers. On the other hand, the outcome of machine learning is only as good as the underlying feature space. Differential geometry can substantially improve the outcome of machine learning since it can enrich the underlying feature space with new geometrical objects, called invariants. In this way, the geometrical features form a high-dimensional feature space for each pixel, where original objects can be resolved. Alternatives based on the geometry of the image scale-invariant interest points have been exploited successfully in the field of computer vision. Here, we integrate geometrical feature extraction based on signal processing, machine learning, and input relying on domain knowledge. The approach is exemplified on the ISBI 2012 image segmentation challenge data set. As a second application, we demonstrate powerful image classification functionality based on the same principles, which was applied to the HeLa and HEp-2 data sets. Obtained results demonstrate that feature space enrichment properly balanced with feature selection functionality can achieve performance comparable to deep learning approaches.

Celiac disease is a gluten-sensitive immune-mediated enteropathy of the small intestine that occurs in genetically predisposed individuals. Abnormal immune response results in mucosal inflammation, villous atrophy, and crypt hyperplasia. This study was a proof-of-concept exercise that used a convolutional neural network to classify hematoxylin and eosin (H&E) histological images of celiac disease, normal small intestine control, and non-specified duodenal inflammation; 7294, 11,642, and 5966 images, respectively. The trained network classified celiac disease images with high performance (accuracy 99.7%, precision 99.6%, recall 99.3%, F1-score 99.5%, and specificity 99.8%). Interestingly, when the same network (already trained for the 3 class images), analyzed duodenal adenocarcinoma (3723 images), the new images were classified as duodenal inflammation in 63.65%, small intestine control in 34.73%, and celiac disease in 1.61% of the cases. Finally, when the network was retrained using the 4 histological subtypes of images, all performance parameters were above 99% for celiac disease. In conclusion, the convolutional neural network (CNN)-based deep neural system was able to classify medical histological images with high performance. Narrow artificial intelligence (AI) is designed to perform tasks that typically require human intelligence, but it operates within limited constraints and is task specific.

Egyptian cotton is one of the most important commodities to the Egyptian economy and is renowned globally for its quality, which is currently graded by manual inspection. This has several drawbacks including significant labour requirement, low inspection efficiency, and influence from inspection conditions such as light and human subjectivity. This current work uses a low-cost colour vision system, combined with machine learning to predict the cotton lint grade of the cultivars Giza 86, 97, 90, 94 and 96. Unsupervised and supervised machine learning approaches were explored and compared. Three different supervised learning algorithms were evaluated: linear discriminant analysis, decision trees and ensemble modelling. The highest accuracy models (77.3-98.2%) used an ensemble modelling technique to classify samples within the Egyptian cotton grades: Fully Good, Good, Fully Good Fair, Good Fair and Fully Fair. The unsupervised learning technique k-means showed that human error is more likely to occur when classifying lint belonging to the higher quality grades and underlined the need for an intelligent system to replace manual inspection.

Recently, there has been considerable research on deepfake detection. However, most existing methods face challenges in adapting to the advancements in new generative models within unknown domains. In addition, the emergence of new generative models capable of producing and editing high-quality images, such as Diffusion, Consistency, and LCM, poses a challenge for traditional deepfake training models. These advancements highlight the need for adapting and evolving existing deepfake detection techniques to effectively counter the threats posed by sophisticated image manipulation technologies. In this paper, our objective is to detect deepfake videos in unknown domains using unlabeled data. Specifically, our proposed approach employs Meta Pseudo Labels (MPL) with Supervised Contrastive Learning, so-called SupCon-MPL, allowing the model to be trained on unlabeled images. MPL involves the simultaneous training of both a Teacher model and a Student model, where the Teacher model generates Pseudo Labels utilized to train the Student model. This method aims to enhance the adaptability and robustness of deepfake detection systems against emerging unknown domains. Supervised Contrastive Learning utilizes labels to compare samples within similar classes more intensively, while encouraging greater distinction from samples in dissimilar classes. This facilitates the learning of features in a diverse set of deepfake images by the model, consequently contributing to the performance of deepfake detection in unknown domains. When utilizing the ResNet50 model as the backbone, SupCon-MPL exhibited an improvement of 1.58% in accuracy compared to traditional MPL in known domain detection. Moreover, in the same generation of unknown domain detection, there was a 1.32% accuracy enhancement, while in the detection of post-generation unknown domains, there was an 8.74% increase in accuracy.

Scene classification plays an important role in the intelligent processing of high-resolution satellite (HRS) remotely sensed image. In HRS image classification, multiple features, e.g. shape, color, and texture features, are employed to represent scenes from different perspectives. Accordingly, effective integration of multiple features always results in better performance compared to methods based on a single feature in the interpretation of HRS image. In this paper, we introduce a multi-task joint sparse and low-rank representation model to combine the strength of multiple features for HRS image interpretation. Specifically, a multi-task learning formulation is applied to simultaneously consider sparse and low-rank structure across multiple tasks. The proposed model is optimized as a non-smooth convex optimization problem using an accelerated proximal gradient method. Experiments on two public scene classification datasets demonstrate that the proposed method achieves remarkable performance and improves upon the state-of-art methods in respective applications.

Autonomous image recognition has numerous potential applications in the field of planetary science and geology. For instance, having the ability to classify images of rocks would allow geologists to have immediate feedback without having to bring back samples to the laboratory. Also, planetary rovers could classify rocks in remote places and even in other planets without needing human intervention. Shu et al. classified 9 different types of rock images using a Support Vector Machine (SVM) with the image features extracted autonomously. Through this method, the authors achieved a test accuracy of 96.71%. In this research, Convolutional Neural Networks(CNN) have been used to classify the same set of rock images. Results show that a 3-layer network obtains an average accuracy of 99.60% across 10 trials on the test set. A version of Self-taught Learning was also implemented to prove the generalizability of the features extracted by the CNN. Finally, one model has been chosen to be deployed on a mobile device to demonstrate practicality and portability. The deployed model achieves a perfect classification accuracy on the test set, while taking only 0.068 seconds to make a prediction, equivalent to about 14 frames per second.

Abstract:The objects of fine-grained image categories(e.g., bird species) are various subclass under different categories. Because the differences between subclass are very subtle and most of them are concentrated in multiple local areas, the task of fine-grained image recognition is very challenging. At the same time, some fine-grained networks tend to focus on a certain region when judging the target category, resulting in the lack of other auxiliary regional features. To this end, Inception V3 is used as the backbone network, and an enhanced and complementary fine-grained image classification network is designed. While adopting the method of reinforcement learning to obtain more detailed fine grain image features, the complementary network can obtain the complementary discriminant area of the target through the method of attention erasure to increase the network's perception of the overall target. Finally, experiments are conducted on CUB-200-2011, FGVC Aircraft and Stanford dogs three open datasets. The experimental results show that the proposed model has better performance.

Accuracy and computational cost are the main challenges of deep neural networks in image recognition. This paper proposes an efficient ranking reduction to binary classification approach using a new feed-forward network and feature selection based on ranking the image pixels. Preference net (PN) is a novel deep ranking learning approach based on Preference Neural Network (PNN), which uses new ranking objective function and positive smooth staircase (PSS) activation function to accelerate the image pixels’ ranking. PN has a new type of weighted kernel based on spearman ranking correlation instead of convolution to build the features matrix. The PN employs multiple kernels that have different sizes to partial rank image pixels’ in order to find the best features sequence. PN consists of multiple PNNs’ have shared output layer. Each ranker kernel has a separate PNN. The output results are converted to classification accuracy using the score function. PN has promising results comparing to the latest deep learning (DL) networks using the weighted average ensemble of each PN models for each kernel on CFAR-10 and Mnist-Fashion datasets in terms of accuracy and less computational cost.

Reemergence of deep Neural Networks (CNNs) has lead to high-performance supervised learning algorithms for the Electro-Optical (EO) domain classification and detection problems. This success is possible because generating huge labeled datasets has become possible using modern crowdsourcing labeling platforms such as Amazon Mechanical Turk that recruit ordinary people to label data. Unlike the EO domain, labeling the Synthetic Aperture Radar (SAR) domain data can be a lot more challenging and for various reasons using crowdsourcing platforms is not feasible for labeling the SAR domain data. As a result, training deep networks using supervised learning is more challenging in the SAR domain. In the paper,we present a new framework to train a deep neural network for classifying Synthetic Aperture Radar (SAR) images by eliminating the need for huge labeled dataset. Our idea is based on transferring knowledge from a related EO domain problem, where labeled data is easy to obtain. We transfer knowledge from the EO domain through learning a shared invariant cross-domain embedding space that is also discriminative for classification. To this end, we train two deep encoders that are coupled through their last year to map data points from the EO and the SAR domains to the shared embedding space such that the distance between the distributions of the two domains is minimized in the latent embedding space. We use the Sliced Wasserstein Distance (SWD) to measure and minimize the distance between these two distributions and use a limited number of SAR label data points to match the distributions class-conditionally. As a result of this training procedure, a classifier trained from the embedding space to the label space using mostly the EO data would generalize well on the SAR domain. We provide theoretical analysis to demonstrate why our approach is effective and validate our algorithm on the problem of ship classification in the SAR domain by comparing against several other learning competing approaches.

Cancer diagnosis and classification are pivotal for effective patient management and treatment planning. In this study, we present a comprehensive approach utilizing ensemble deep learning techniques to analyze breast cancer histopathology images. Our datasets are based on two widely employed datasets from different centers for two different tasks: BACH and BreakHis. Within the BACH Dataset, we deployed an ensemble strategy incorporating VGG16 and ResNet50 architec-tures to achieve precise classification of breast cancer histopathology images. Introducing a novel image patching technique, to preprocess a high-resolution image, which facilitates focused analysis of localized regions of interest. The annotated BACH dataset encompasses 400 WSIs across four distinct classes: Normal, Benign, In Situ Carcinoma, and Invasive Carcinoma. In addition, we employed the BreakHis dataset, utilizing VGG16, ResNet34, and ResNet50 models to classify mi-croscopic images into eight distinct categories (four benign and four malignant). For both models we leveraged a five-fold cross-validation approach for rigorous training and testing. Preliminary ex-perimental results indicate a Patch classification accuracy of 95.31% (on BACH dataset) and WSI image classification accuracy of 98.43% (BreakHis). This research significantly contributes to on-going endeavors in harnessing artificial intelligence to advance breast cancer diagnosis, potentially fostering improved patient outcomes and alleviating healthcare burdens.

Timely and accurate acquisition of information on distribution of the crop planting structure in Loess Plateau of eastern Gansu Province, one of the most important agricultural areas in western China, is crucial for promoting fine management of agriculture and ensuring food security. In this study, the Object-Based Image Classification (OBIC), the Random Forest (RF) and Convolutional Neural Network (CNN) models were employed to classify the crop planting structure of four representative test areas in Qingyang City in a precise manner. Firstly, different optimal segmentation scales for various crops were selected using the Estimation of Scale Parameter 2 (ESP2) tool and the Ratio of Mean Difference to Neighbors(ABS) to Standard Deviation (RMAS)model. The images were then segmented through multiresolution segmentation combined with the Canny Edge Detection algorithm. Secondly, the L1 regularized logistic regression model was utilized to select and optimize 39 spatial feature factors including spectral, textural, geometric, and index features, in conjunction with phenological factors. Lastly, under the multi-level classification framework, the Random Forest (RF) classifier and Convolutional Neural Network (CNN) model, combined with object-based multiresolution segmentation, was used to classify the crop planting structure. The findings show that: Thanks to the Canny Edge Detection algorithm, we can obtain a more complete boundary of the segmented objects and improve the separability. The optimal segmentation scales for corn, vegetables, and buckwheat were found to be 55, 70, and 35, respectively, while wheat and apple had optimal segmentation scales of 65. In addition to phenological characteristics, the number of selected spatial features for corn, vegetables, buckwheat, wheat, and apple were 9, 7, 16, 12, and 10, respectively. The CNN model demonstrated high consistency with the RF model in the classification results, but the accuracy of RF model is higher than that of CNN model on the whole. The overall accuracy of classification using RF model in four test areas registered 91.93%, 94.92%, 89.37% and 90.68%, respectively. This paper introduced crop phenological factors, effectively improving the extraction precision of shattered agricultural planting structure in the Loess Plateau of eastern Gansu Province. Its findings have important application value in crop monitoring, management, food security and other related fields.

Target recognition and fine-grained ship classification in remote sensing face the challenges of high inter-class similarity and sample scarcity. A transfer fusion-based ship image harmonization algorithm is proposed to overcome these challenges. This algorithm designs a feature transfer fusion strategy based on the combination of region aware instantiation and attention mechanism. Implement adversarial learning through image harmony generator and discriminator module to generate realistic remote sensing ship harmony images. Furthermore, the do-main encoder and domain discriminator modules are responsible for extracting feature representations of foreground and background, and further aligning ship foreground with remote sensing ocean background features through feature discrimination. Compared with other advanced image conversion techniques, our algorithm delivers more realistic visuals, improving classification accuracy for six ship types by 3% and twelve types by 2.94%, outperforming Sim2RealNet. Finally, a mixed dataset containing data augmentation and harmonizing samples and real data was proposed for the fine-grained classification task of remote sensing ships. Evaluation experiments were conducted on 8 typical fine-grained classification algorithms, and the accuracy of fine-grained classification for all categories of ships was analyzed. The experimental results show that the mixed dataset proposed in this paper effectively alleviates the long tail problem in real datasets, and the proposed remote sensing ship data augmentation framework performs better than state-of-the-art data augmentation methods in fine-grained ship classification tasks.

This paper discusses the computer-aided (CAD) classification between Hepatocellular Carcinoma (HCC), i.e., the most common type of liver cancer, and Liver Abscess, based on ultrasound image texture features and Support Vector Machine (SVM) classifier. Among 79 cases of liver diseases, with 44 cases of HCC and 35 cases of liver abscess, this research extracts 96 features of Gray-Level Co-occurrence Matrix (GLCM) and Gray-Level Run-Length Matrix (GLRLM) from the region of interests (ROIs) in ultrasound images. Three feature selection models, i) Sequential Forward Selection, ii) Sequential Backward Selection, and iii) F-score, are adopted to determine the identification of these liver diseases. Finally, the developed system can classify HCC and liver abscess by SVM with the accuracy of 88.875%. The proposed methods can provide diagnostic assistance while distinguishing two kinds of liver diseases by using a CAD system.

Synthetic Aperture Radar has been extensively used in numerous fields and can gather a wealth of information about the area of interest. This large-scene data-intensive technology puts a high value on automatic target recognition (ATR) which can free the utilizers and boost the efficiency. Recent advances in artificial intelligence have made it possible to create a deep learning-based SAR ATR that can automatically identify target features from massive input data. In the last 6 years, intensive research has been conducted in this area, however, most papers in the current SAR ATR field used recurrent neural network (RNN) and convolutional neural network (CNN)-varied models to deepen the regime’s understanding of the SAR images. To equip SAR ATR with updated deep learning technology, this paper tries to apply a lightweight vision transformer (LViT)-based model to classify SAR images. The entire structure was verified by an open- accessed SAR data set and recognition results show that the final classification outcomes are robust and more accurate in comparison with referred traditional network structures without even using any convolutional layers.

Crop disease classification has always been a critical and persistent problem in the field of agricultural and forestry sciences, where often we do not have access to a sufficient number of samples to know the distribution of real-world samples. How to make full use of the existing data is the starting point of our thinking. To address this problem, this paper proposes a supervised image augmentation method Negative Contrast, which uses the contrast images of existing disease samples after removing disease areas as negative samples for image augmentation when samples are relatively scarce. Numerous experiments have shown that several classical models using this augmentation method have improved in disease classification of four crops, rice, wheat, corn, and soybean, with a maximum accuracy improvement of 30.8%. In addition, the comparative analysis of attentional heat map shows that the model using Negative Contrast is more accurate and intense on the area of interest of diseases, and thus reflects better generalization ability in real-world disease classification. Our dataset and codes can be found in https://www.kaggle.com/datasets/w970704112/corn-wheat-rice-soybean and https://github.com/hiter0/contrastaug .

Convolutional neural networks (CNN) have attracted much attention as a commonly used method for hyperspectral image (HSI) classification in recent years, however, CNNs can only be applied to Euclidean data and have limitations in dealing with relationships due to the limitations of local feature extraction. However, each pixel of a hyperspectral image contains a set of spectral bands that are correlated and interact with each other, and the methods used to process Euclidean data cannot effectively obtain these correlations. In contrast, the graph convolutional network (GCN) can be used in non-Euclidean data, but usually leads to oversmoothing and ignoring local detail features due to the need for superpixel segmentation processing to reduce computational effort. To overcome the above problems, we constructed a network a fusion network based on GCN and CNN, which contains two branches: a graph convolutional network based on superpixel segmentation and a convolutional network with added attention mechanism. The graph convolu-tional branch can extract the structural features and capture the relationships between the nodes, and the convolutional branch can extract the detailed features in the local fine region. Owing to the fact that the features extracted from the two branches are different, the classification performance can be improved by fusing the complementary features extracted from the two branches. To vali-date the proposed algorithm, experiments were conducted on three widely used datasets, namely Indian Pines, Pavia University, and Salinas, and the overall accuracy of 98.78% was obtained in the Indian Pines dataset, and the overall accuracy of 98.99% and 98.69% was obtained in the other two datasets. The results showed that the proposed fusion network can obtain richer features and achieve high classification accuracy.

The task of weakly supervised point cloud semantic segmentation has received widespread attention and also has been widely used in autonomous driving, robotics, and modern industries. Due to the high dimensionality and large volume of point cloud data, many technical difficulties appeal to weakly supervised semantic processing, especially segmentation. However, weakly-supervised schemes usually provide only partial labelling information of the underlying point cloud data and thus need to effectively extract local features, geometric information, and contextual relationships only using these limited labelled data for supervised learning. To improve weakly-supervised semantic segmentation, we propose a novel segmentation network through the boundary-based feature aggregation based on a K-NN algorithm with down-sampling operation, and also introduce the smoothness loss and Siamese loss for effective segmentation. The experiments on public datasets demonstrate that our presented segmentation method exceeds most of the existing fully supervised and weakly-supervised methods in terms of mIoU. Specifically, our network has high segmentation accuracy on the labels of objects with similar geometrical structures, such as ceiling, wall, floor, chair and table, reaching 91.2%, 98.8%, 83.3%, 75.3% and 80.2%, respectively. Extensive experiments also illustrate its robustness, effectiveness, and generalization of the proposed weakly supervised segmentation network.

"Finding fresh water in the ocean of data." is a challenge that all deep learning domains struggle with, especially in the area of hyperspectral image analysis. As hyperspectral remote sensing technology advances by leaps and bounds, there are increasing amounts of hyperspectral images(HSIs) can be available. Whereas in fact, these unlabeled HSIs are powerless to be used as material to driven a supervised learning task due to the extremely expensive labeling costs and some unknown regions. Although learning-based methods have achieved remarkable performance due to their superior ability to represent features, at the cost, these methods are complex, inflexible and tough to carry out transfer learning. In this paper, we propose the "Instructional Mask AutoEncoder"(IMAE), which is a simple and powerful self-supervised learner for HSI classification that uses a transformer-based mask autoencoder to extract the general features of HSIs through a self-reconstructing agent task. Moreover, we utilize the metric learning to perform an instructor which can direct the model focus on the human interested region of the input so that we can alleviate the defects of transformer-based model such as local attention distraction, lack of inductive bias and tremendous training data requirement. In downstream forward propagation, instead of global average pooling, we employ a learnable aggregation to put the tokens into fullplay. The obtained results illustrate that our method effectively accelerates the convergence rate and promotes the performance in downstream task.

To solve the problem that the common long-tailed classification method does not use the semantic features of the original label text of the image, and the difference between the classification accuracy of most classes and minority classes is large, the long-tailed image classification method based on enhanced contrast visual language trains the head class and tail class samples separately, uses text image to pre-train the information, and uses enhanced momentum contrast loss function and RandAugment enhancement to improve the learning of tail class samples. On the ImageNet-LT long-tailed dataset, the enhanced contrastive visual-language based long-tailed image classification method has improved all class accuracy, tail class accuracy, middle class accuracy, and F1 values by 3.4%, 7.6%, 3.5%, and 11.2%, respectively, compared to the BALLAD method. The difference in accuracy between the head class and tail class is reduced by 1.6% compared to the BALLAD method. The results of three comparative experiments indicate that the long-tailed image classification method based on enhanced contrastive visual-language has improved the performance of tail classes and reduced the accuracy difference between majority and minority classes.

This review provides an overview of eXplainable AI (XAI) methods for oncological ultrasound image analysis and compares their performance evaluations. A systematic search of Medline Embase and Scopus between March 25 and April 14 2024 identified 17 studies describing 14 XAI methods, including visualization, semantics, example-based, and hybrid functions. These methods primarily provided specific, local, and post-hoc explanations. Performance evaluations focused on AI model performance, with limited assessment of explainability impact. Standardized evaluations incorporating clinical end-users are generally lacking. Enhanced XAI transparency may facilitate AI integration into clinical workflows. Future research should develop real-time methodologies and standardized quantitative evaluative metrics.

Attribute reduction is a core technique in the rough set domain and an important step in data preprocessing. Researchers have proposed numerous innovative methods to enhance the capability of attribute reduction, such as the emergence of multi-granularity rough set models, which can effectively process distributed and multi-granularity data. However, these innovative methods still have many shortcomings, such as how to deal with complex constraints and how to perform multi-angle effectiveness evaluations. Based on the multi-granularity model, this study proposes a new method of attribute reduction, namely using multi-granularity neighborhood information gain ratio as the measurement criterion. This method combines both supervised and unsupervised perspectives, and by integrating multi-granularity technology with neighborhood rough set theory, constructs a model that can adapt to multi-level data features. This model can select the optimal granularity level or attribute set according to the requirements. Finally, the proposed method is experimented with on 15 UCI datasets against six other attribute reduction algorithms. The experimental outcomes, upon analysis and comparative review, substantiate the reliability and consistency of the attribute reduction approach introduced in this study.

One of the most important challenges in the Machine and Deep Learning areas today is to build good models using small datasets, because sometimes it is not possible to have large ones. Several techniques have been proposed in the literature to address this challenge. This paper aims at studying the different available Deep Learning techniques and performing a thorough experimentation to analyze which technique or combination thereof improves the performance and effectiveness of the models. A complete comparison with classical Machine Learning techniques was carried out, to contrast the results obtained using both techniques when working with small datasets. Thirteen algorithms were implemented and trained using three different small datasets (MNIST, Fashion MNIST, and CIFAR-10). Each experiment was evaluated using a well-established set of metrics (Accuracy, Precision, Recall, F1, and the Matthews correlation coefficient). The experimentation allowed concluding that it is possible to find a technique or combination of them to mitigate a lack of data, but this depends on the nature of the dataset, the amount of data, and the metrics used to evaluate them.

Deep learning classification is the state-of-the-art of machine learning approach. Earlier work proves that the deep convolutional neural network has successfully and brilliantly in different applications such as images or video data. Recognizing and clarifying the remote sensing aspect of the earth's surface and exploit land cover and land use (LCLU). First, this article summarized the remote sensing emerging application and challenges for deep learning methods. Second, we propose four approaches to learn efficient and effective CNNs to transfer image representation on the ImageNet dataset to recognize LCLU datasets. We use VGG16, Inception-ResNet-V2, Inception-V3, and DenseNet201 models to extract features from the EACC dataset. We use pre-trained CNNs on ImageNet to extract features. For feature selection we proposed principal component analysis (PCA) to improve accuracy and speed up the model. We train our model by multi-layer perceptron (MLP) as a classifier. Lastly, we apply the multi-granularity encoding ensemble model. We achieve an overall accuracy of 92.3% for the nine-class classification problem. This work will help remote sensing scientists understand deep learning tools and apply them in large-scale remote sensing challenges

Ship detection from synthetic aperture radar (SAR) imagery is crucial for various fields in real-world applications. Numerous deep learning-based detectors have been investigated for SAR ship detection, which require a substantial amount of labeled data for training. However, SAR data annotation is time-consuming and demands specialized expertise, resulting in that deep learning-based SAR ship detectors struggle due to a lack of annotations. With limited labeled data, semi-supervised learning is a popular approach for boosting detection performance by excavating valuable information from unlabeled data. In this paper, a semi-supervised SAR ship detection network is proposed, termed as Multi-Teacher Dempster-Shafer Evidence Fusion Net-work (MTDSEFN). The MTDSEFN is an enhanced framework based on the basic teacher-student skeleton frame, comprising two branches: the Teacher Group (TG) and the Agency Teacher (AT). The TG utilizes multiple teachers to generate pseudo-labels for different augmentation versions of unlabeled samples, which are then refined to obtain high-quality pseudo-labels by using Dempster-Shafer (D-S) fusion. The AT not only serves to deliver weights of its own teacher to the TG at the end of each epoch, but also updates its own weights after each iteration, enabling the model to effectively learn rich information from unlabeled data. The combination of TG and AT guarantees both reliable pseudo-label generation and learning comprehensive diversity information from numerous unlabeled samples. Extensive experiments were performed on two public SAR ship datasets, and the results demonstrated the effectiveness and superiority of the proposed approach.

This paper presents an on-device semi-supervised human activity detection system that can learn and predict human activity patterns in real time. The proposed semi-supervised learning (SSL) framework uses sparsely labelled user activity events acquired from Inertial Measurement Unit (IMU) sensors installed as wearable device. The objective is to learn classification of multiple classes of human activities in real-time. The proposed cluster-based learning model in this approach is trained with data from the same target user, thus preserving data privacy while providing personalized activity detection services. Two different cluster labelling strategies, namely, population-based, and distance-based, are employed to achieve the desired classification performance. A comparative study of these strategies has been presented in terms of classifiability and classification accuracies. The proposed system is shown to be computationally efficient, which is relevant in the context of limited computing resources on typical wearable devices. The trade-off between classification accuracy and computation complexity is analyzed for different algorithmic hyper-parameters and other system parameters. Extensive experimentation and simulation study have been conducted on a multi-user human activity data from the public do-main in order to validate the proposed learning paradigm.

Medical image segmentation is a critical step in various healthcare applications, aiding in diagnosis, treatment planning, and disease monitoring. In this study, we investigate the efficacy of a segmentation approach based on the 2D wavelet transform. Leveraging a dataset comprising 10 diverse medical images, we evaluate the performance of our segmentation method using three key metrics: accuracy, precision, and recall. Our findings demonstrate that the proposed approach enhances segmentation accuracy, offering promising results compared to existing methods. By harnessing the multi-resolution feature extraction capabilities of the 2D wavelet transform, our method achieves improved delineation of medical image structures, paving the way for more accurate and efficient healthcare interventions.

Accurate crop diseases classification is crucial for ensuring food security and enhancing agricultural productivity. However, existing crop disease classification algorithms primarily focus on a single image modality and typically require a large number of samples. Our research counters these issues by using pre-trained Vision-Language Models (VLMs), which enhances multimodal synergy for better crop disease classification than traditional unimodal approaches. Firstly, we apply the multimodal model Qwen-VL to generate meticulous textual descriptions for representative disease images selected through clustering from the training set, which will serve as prompt text for generating classifier weights. Compared to using solely the language model for prompt text generation, this approach better captures and conveys fine-grained and image-specific information, thereby enhancing prompt quality. Secondly, we integrate the cross-attention and the SE (Squeeze-and-Excitation) attention into the training-free mode VLCD (Vision-Language model for Crop Diseases classification) and the training-required mode VLCD-T (VLCD-Training) respectively for prompt text processing, enhancing classifier weights by emphasizing key text features. Experimental outcomes conclusively prove our method’s heightened classification effectiveness in few-shot crop disease scenarios, tackling data limitations and intricate disease recognition issues. It offers a pragmatic tool for agricultural pathology and reinforces the smart farming surveillance infrastructure.

Polarimetric features extracted from polarimetric synthetic aperture radar (PolSAR) images contain abundant scattering information of objects. Utilizing this information for PolSAR image classification can improve accuracy and enhance object monitoring. Although significant progress has been achieved in related research, there are still issues such as insufficient utilization of polarimetric information and features. In this paper, a deep learning classification method based on polarimetric channel power features for PolSAR is proposed. The distinctive characteristic of this method is that the polarimetric features inputted into the deep learning network are the power values of polarimetric channels, where these channels are treated equivalently and contain complete polarimetric information. In terms of experimental comparative study, besides polarimetric power combinations, three different data input schemes were employed as data for the input end of the convolutional neural network (CNN), taking into account existing research. The neural networks utilizes the extracted polarimetric features to classify images, and the classification accuracy analysis is employed to compare the strengths and weaknesses of different input schemes. It is worth mentioning that the polarized characteristics of the data input scheme mentioned in this article have been derived through rigorous mathematical deduction, and each polarimetric feature has clear physical meaning. By testing different data input schemes on the Gaofen-3 (GF-3) PolSAR image, the experimental results show that the method proposed in this article outperforms existing methods and can improve the accuracy of classification to a certain extent, validating the effectiveness of this method in large-scale area classification.

This proposed research explores a novel approach to image classification by deploying a complex-valued neural network (CVNN) on a field-programmable gate array (FPGA), specifically for classifying 2D images transformed into polar form. The aim of this research is to address the limitations of existing neural network models in terms of energy and resource efficiency, by exploring the potential of FPGA-based hardware acceleration in conjunction with advanced neural network architectures like CVNNs. The methodological innovation of this research lies in the Cartesian to polar transformation of 2D images, effectively reducing the input data volume required for neural network processing. Subsequent efforts focused on constructing a CVNN model optimized for FPGA implementation, emphasizing the enhancement of computational efficiency and overall performance. The experimental findings provide empirical evidence supporting the efficacy of the image classification system developed in this study. One of the developed models, CVNN_128, achieves an accuracy of 88.3% with an inference time of just 1.6ms and a power consumption of 4.66mW for the classification of the MNIST test dataset consists of 10,000 frames. While there is a slight concession in accuracy compared to recent FPGA implementations that achieve 94.43%, our model significantly excels in classification speed and power efficiency—surpassing existing models by more than a factor of 100. In conclusion, the paper demonstrates the substantial advantages of FPGA-implementation of CVNNs for image classification tasks, particularly in scenarios where speed, resource, and power consumption are critical. The study’s reproducible results and corresponding code are available on GitHub at the following link: https://github.com/mahmad2005/CVNNonFPGA

Although neural network architectures are critical for their performance, how the structural characteristics of a neural network affect its performance has still not been fully explored. We here map architectures of neural network to directed acyclic graphs, and find that incoherence, a structural characteristic to measure the order of directed acyclic graphs, is a good indicator for the performance of corresponding neural networks. Therefore we propose a deep isotropic neural network architecture by folding a chain of same blocks then connecting the blocks with skip connections at different distances. Our models, named FoldNet, have two distinguishing features compared with traditional residual neural netowrks. First, the distances between block pairs connected by skip connections increase from always equal to one to specially selected different values, which lead to more incoherent graphs and let the neural network explore larger receptive fields and thus enhance its multi-scale representation ability. Second, the number of direct paths increases from one to multiple, which leads to a larger proportion of shorter paths and thus improve the direct propagation of information throughout the entire network. Image classification results on CIFAR-10 and Tiny ImageNet benchmarks suggested that our new network architecture performs better than traditional residual neural networks.

Recent neuroscience research results show that the nerve information in the brain is not only encoded by the spatial information. Spiking neural network based on pulse frequency coding plays a very important role in dealing with the problem of brain signal, especially complicated space-time information. In this paper, an unsupervised learning algorithm for bilayer feedforward spiking neural networks based on spike-timing dependent plasticity (STDP) competitiveness is proposed and applied to SAR image classification on MSTAR for the first time. The SNN learns autonomously from the input value without any labeled signal and the overall classification accuracy of SAR targets reached 80.8%. The experimental results show that the algorithm adopts the synaptic neurons and network structure with stronger biological rationality, and has the ability to classify targets on SAR image. Meanwhile, the feature map extraction ability of neurons is visualized by the generative property of SNN, which is a beneficial attempt to apply the brain-like neural network into SAR image interpretation.

In recent years, researchers have paid increasing attention on hyperspectral image (HSI) classification using deep learning methods. To improve the accuracy and reduce the training samples, we propose a double-branch dual-attention mechanism network (DBDA) for HSI classification in this paper. Two branches are designed in DBDA to capture plenty of spectral and spatial features contained in HSI. Furthermore, a channel attention block and a spatial attention block are applied to these two branches respectively, which enables DBDA to refine and optimize the extracted feature maps. A series of experiments on four hyperspectral datasets show that the proposed framework has superior performance to the state-of-the-art algorithm, especially when the training samples are signally lacking.

Enclosed public spaces are the hotspots for airborne disease transmission. To measure and maintain indoor air quality in terms of airborne transmission, an open source, low cost and distributed array of Particulate Matter Sensors has been developed and named as Dynamic Aerosol Transport for Indoor Ventilation or DATIV system. This system can use multiple Particulate Matter Sensors (PMS) simultaneously and can be remotely controlled using a Raspberry Pie based operating system. The data acquisition system can be easily operated using the GUI within any common browser installed on a remote device such as a PC or Smartphone with corresponding IP address. The software architecture and validation measurements are presented together with possible future developments.

It is important for food recognition to separate each ingredient within a food image at the pixel level. In this paper, we propose a new approach to segment ingredients by utilizing a CNN-based Single-Ingredient Classification Model. In detail, we firstly introduce a standardized biological-based hierarchical ingredient structure and construct a single-ingredient image dataset based on this structure. Then, we build a single-ingredient classification model based on a novel convolutional neural network (CNN) architecture that utilizes an attention mechanism. Afterwards, we propose a new framework for segmentation using the above single-ingredient classification model as the backbone. In this framework, two methods are involved in segmenting ingredients in the food images. We introduce five evaluation metrics (IoU, Dice, Purity, Entirety, Loss of GTs) to assess the performance of ingredient segmentation in terms of ingredient classification. Extensive experiments demonstrate the effectiveness of the proposed method, achieving an maximal mIoU of 0.65, mDice of 0.77, mPurity of 0.83, mEntirety of 0.80, and mLoGTs of 0.06 on the FoodSeg103 dataset. The results confirm that our CNN-based architecture achieves higher segmentation performance compared to ResNet18 and EfficientNet-B0 when used as the backbone for ingredient segmentation. We believe that our ingredient segmentation approach lays the foundation for subsequent ingredient recognition.

In this paper, we examine two strategies for boosting the performance of ensembles of Siamese networks (SNNs) for image classification using two loss functions (Triplet and Binary Cross Entropy) and two methods for building the dissimilarity spaces (FULLY and DEEPER). With FULLY, the distance between a pattern and a prototype is calculated by comparing two images using the fully connected layer of the Siamese network. With DEEPER, each pattern is described using a deeper layer combined with dimensionality reduction. The basic design of the SNNs takes advantage of supervised k-means clustering for building the dissimilarity spaces that train a set of support vector machines, which are then combined by sum rule for a final decision. The robustness and versatility of this approach are demonstrated on several cross-domain image data sets, including a portrait data set, two bioimage and two animal vocalization data sets. Results show that the strategies employed in this work to increase the performance of dissimilarity image classification using SNN is closing the gap with standalone CNNs. Moreover, when our best system is combined with an ensemble of CNNs, the resulting performance is superior to an ensemble of CNNs, demonstrating that our new strategy is extracting additional information.

Human beatboxing is a vocal art making use of speech organs to produce vocal drum sounds and imitate musical instruments. Beatbox sound classification is a current challenge that can be used for automatic database annotation and music-information retrieval. In this study, a large-vocabulary humanbeatbox sound recognition system was developed with an adaptation of Kaldi toolbox, a widely-used tool for automatic speech recognition. The corpus consisted of eighty boxemes, which were recorded repeatedly by two beatboxers. The sounds were annotated and transcribed to the system by means of a beatbox specific morphographic writing system (Vocal Grammatics). The image processing techniques plays vital role on image Acquisition, image pre-processing, Clustering, Segmentation and Classification techniques with different kind of images such as Fruits, Medical, Vehicle and Digital text images etc. In this study the various images to remove unwanted noise and performs enhancement techniques such as contrast limited adaptive histogram equalization, Laplacian and Harr filtering, unsharp masking, sharpening, high boost filtering and color models then the Clustering algorithms are useful for data logically and extract pattern analysis, grouping, decision-making, and machine-learning techniques and Segment the regions using binary, K-means and OTSU segmentation algorithm. It Classifying the images with the help of SVM and K-Nearest Neighbour(KNN) Classifier to produce good results for those images.

Breast cancer is the second most common cancer worldwide, primarily affecting women, while histopathological image analysis is one of the possibilities used to determine tumor malignancy. Regarding image analysis, the application of deep learning has become increasingly prevalent in recent years. However, a significant issue is the unbalanced nature of available datasets, with some classes having more images than others, which may impact the performance of the models due to poorer generalisability. A possible strategy to avoid this problem is downsampling the class with the most images to create a balanced dataset. Nevertheless, this approach is not recommended for small datasets as it can lead to poor model performance. Instead, techniques such as data augmentation are traditionally used to address this issue. These techniques apply simple transformations, such as translation or rotation, to the images to increase variability in the dataset. Recently, Generative Adversarial Networks (GANs) have emerged, which can generate images from a relatively small training set. This work aims to enhance the model’s performance in classifying histopathological images by applying data augmentation using GANs instead of traditional techniques.

In recent years, researches on combining wavelet decomposition and convolutional neural network (CNN) together to classify hyperspectral images (HSI) have emerged, and some effective classification models have been proposed and achieved good classification results. However, there are two problems for most of the proposed models. One is the heavy training parameters and the other is that there is no distinction between the classification effectiveness of low frequency and high frequency features after wavelet decomposition. In this paper, a new light-weighted HSI classification model (LLFWCNN) is proposed, which performs multi-layer wavelet decomposition for HSI after dimensionality reduction, and only arranges the low frequency features in a specific stack mode, then classifies them through a well-designed convolutional neural network. Compared with other classification models, the number of parameters is only 2.5% of that of other models. And only the low frequency features after wavelet decomposition are used, while the high frequency features are abandoned. The results showed that compared with the state-of-the-art classification models, LLFWCNN could obtain the same or even better classification results with fewer network parameters and proved that the low frequency features after wavelet decomposition provide LLFWCNN with more favorable information for HSI classification as well.

Few-shot fine-grained image classification (FSFGIC) methods refer to machine learning methods which aim to classify images (e.g., bird species, flowers, and airplanes) belonging to subordinate object categories of the same entry-level category with only a few samples. It is worth to note that feature representation learning is used not only to represent training samples, but also to construct classifiers for performing various FSFGIC tasks. In this paper, starting from the definition of FSFGIC, a taxonomy of feature representation learning for FSFGIC is proposed. According to this taxonomy, we discuss key issues on FSFGIC (including data augmentation, local or/and global deep feature representation learning, class representation learning, and task specific feature representation learning). The existing popular datasets and evaluation standards are introduced. Furthermore, a novel classification performance evaluation mechanism is designed with a 0.95 confidence interval for judging whether the classification accuracy obtained by a certain specified method is good or bad. Moreover, current challenges and future trends of feature representation learning on FSFGIC are elaborated.

The use of convolutional neural networks (CNNs) in image classification has become the standard method of approaching computer vision problems. Here we apply pre-trained networks to classify images of non-breaking, plunging and spilling breaking waves. The CNNs are used as basic feature extractors and a classifier is then trained on top of these networks. The dynamic nature of breaking waves is exploited by using image sequences to gain extra information and improve the classification results. We also see improved classification performance in using pre-computed image features such as the optical flow between image pairs. The inclusion of the dynamic information improves the classification between breaking wave classes. We also provide corrections to the methodology from the article from which the data originates to achieve a more accurate assessment of performance.

Aiming at the problems that large amount of video monitoring image data in underground coal mines leads to difficulties in transmission and storage, compressed sensing theory is introduced to encode and decode video images, and a new distributed video coding scheme is proposed. In order to obtain more sparse representation and more general applicability, a block-based adaptive sparse base scheme is proposed. For the acquisition of side information, fixed weight is usually used to synthesize side information and the correlation between different image blocks is neglected, a block-based classification weighted side information generation scheme is proposed. Experimental results show that the block-based classification codec scheme can make full use of inter-frame correlation. Under the appropriate sampling rate, the PSNR value of video reconstruction increases, which effectively improves the quality of video frame reconstruction.

This study compared two object-oriented land use change detection methods—detection after classification (DAC) and classification after detection (CAD) —based on a digital elevation model, slope data, and multi-temporal Landsat images (TM image for 2000 and ETM image for 2010). We noted that the overall accuracy of the DAC (86.42%) was much higher than that of the CAD (71.71%). However, a slight difference between the accuracies of the two methods exists for deciduous broadleaf forest, evergreen coniferous forest, mixed wood, upland, paddy, reserved land, and settlement. Owing to substantial spectrum differences, these land use types can be extracted using spectral indexes. The accuracy of DAC was much higher than that of CAD for industrial land, traffic land, green shrub, reservoir, lake, river, and channel, all of which share similar spectrums. The discrepancy was mainly because DAC can completely utilize various forms of information apart from spectrum information during a two-stage classification. In addition, the change-area boundary was not limited at first, but was adjustable in the process of classification. DAC can overcome smoothing effects to a great extent using multi-scale segmentations and multi-characters in detection. Although DAC yielded better results, it was more time-consuming (28 days) because it uses a two-stage classification approach. Conversely, CAD consumed less time (15 days). Thus, a hybrid of the two methods is recommended for application in land use change detection.

Deep learning methods have been widely used in PolSAR image classification. To learn the polarimetric information, many deep learning methods expect to learn high-level semantic features from original PolSAR data. However, only original data cannot learn multiple scattering features and complex structures for extremely heterogeneous terrain objects. In addition, deep learning methods always cause edge confusion due to the high-level features. To overcome these shortages, we propose a double-channel CNN network combined with an edge-preserving MRF model(DCCNN-MRF) for PolSAR image classification. Firstly, to combine complex matrix data and multiple scattering features together, a double-channel convolution network(DCCNN) is developed, which consists of a Wishart-based complex matrix and multi-feature subnetworks. The Wishart-based complex matrix network can learn the statistical characteristics and channel correlation well, and the multi-feature network can learn high-level semantic features well. Then, a unified network framework is designed to fuse two kinds of features to enhance advantageous features and reduce redundant ones. Finally, an edge preserving MRF model is designed to combine with the DCCNN network. In the MRF model, a sketch map-based edge energy function is designed by defining adaptive weighted neighborhood for edge pixels. Experiments are conducted on four real PolSAR data sets with different sensors and bands. Experimental results demonstrate the effectiveness of the proposed DCCNN-MRF method.

Underwater imaging presents unique challenges, notably color distortions and reduced contrast due to light attenuation and scattering. Most underwater image enhancement methods, first use linear transformations for color compensation and then enhance the image. We observed that linear transformation for color compensation is not suitable for certain images. For such images, a non-linear mapping is a better choice. Our paper introduces a unique underwater image restoration approach leveraging a streamlined convolutional neural network (CNN) for dynamic weight learning for linear and non-linear mapping. In the first phase, a classifier is applied that classifies the input images into Type-I or Type-II. In the second phase, depending on the classification, the Deep Line Model (DLM) for Type-I images or the Deep Curve Model (DCM) for Type-II images. For mapping an input image to an output image, DML creatively combines color compensation and contrast adjustment in a single step and uses deep lines for transformation, whereas DCM employs higher-order curves. Both models utilize lightweight neural networks that learn per-pixel dynamic weights based on the input image’s characteristics. Comprehensive evaluations on benchmark datasets using metrics like peak signal-to-noise ratio(PSNR) and root mean squares error (RMSE) affirm our method’s effectiveness in accurately restoring underwater images, outperforming existing techniques.

The key links of face recognition are digital image preprocessing, facial feature extraction and pattern recognition, this article aimed at the current problem of slow speed and low recognition accuracy of face recognition , from the above three key links, on the basic of analyzing the therories of Fractional Differential Masks Operator (FDMO), Principal Component Analysis (PCA) and Support Vector Machine (SVM), design a kind of FDMO+PVA+SVM coupling algorithm that applies to face recognition to improve the speed and accuracy of it. To realize FDMO+PCA+SVM coupling algorithm, first, we should apply FDMO to face image processing binary marginalization, the purpose is getting face contour; Then, we apply PCA to get the feature of face image which is disposed by binary marainalization. At last, we can apply One-Against All of the SVM classifier and LibSVM software package to realize face recognition. Beside, the article with nine different coupling algorithm design four groups of experimental reaults on the ORL face database verified by comparative analysic FDMO+PCA+SVM coupling algorithm in the superiority of face recognition accuracy and speed.

Computer vision is a powerful technology that has enabled solutions in various fields by analyzing visual attributes in images. One field that has taken advantage of computer vision is agricultural automation, which promotes high-quality crop production. The nutritional status of a crop is a crucial factor in determining its productivity. This status is mediated by approximately 14 chemical elements acquired by the plant, and their determination plays a pivotal role in farm management. To address the timely identification of nutritional disorders, this study focuses on the classification of three levels of phosphorus deficiencies through individual leaf analysis. The methodological steps include: (1) generating a database with laboratory-grown maize plants that were induced to total phosphorus deficiency, medium deficiency, and total nutrition, using different capture devices; (2) processing the images with state-of-the-art transfer learning architectures (i.e. VGG16, ResNet50, GoogLeNet, DenseNet201, and MobileNetV2); and (3) evaluating the classification performance of the models using the created database. The results show that the VGG16 model achieves superior performance, with 98% classification accuracy. However, the other studied architectures also demonstrate competitive performance and are considered state-of-the-art automatic leaf deficiency detection tools. The proposed method can be a starting point to fine-tune machine vision-based solutions tailored for real-time monitoring of crop nutritional status.

Deep learning is improving by leaps and bounds in remote sensing images (RSIs) analysis, pre-trained convolutional neural networks (CNNs) have shown remarkable performance in remote sensing image scene classification (RSISC). Nonetheless, pre-trained CNNs require massive annotated data as samples for data training. When labeled samples are not sufficient, the most common solution is to the pre-trained CNNs using a great deal of natural image dataset (e.g. ImageNet). However, these pre-trained CNNs require a large quantity of labelled data for training, which is often not feasible in RSISC, especially when the target RSIs have different imaging mechanisms from RGB natural images. In this paper, we proposed an improved hybrid classical-quantum transfer learning CNNs composed of classical and quantum elements to classify open-source RSI dataset. The classical part of the model is made up of a ResNet network which extracts useful features from RSI datasets. To further refine the network performance, a tensor quantum circuit is subsequently employed by tuning parameters on near-term quantum processors. We tested our models on open-source RSI dataset. In our comparative study, we have concluded that the hybrid classical-quantum transferring CNN has achieved better performance than other pre-trained CNNs based RSISC methods with small training samples. Moreover, it has been proved that the proposed algorithm improves the classification accuracy while greatly decreasing sum of model parameters and sum of training data.

Deep neural networks have been extensively researched in the field of document image classification to improve classification performance and have shown excellent results. However, there is little research in this area that addresses the question of how well these models would perform in a real-world environment, where the data the models are confronted with often exhibits various types of noise or distortion. In this work, we present two separate benchmark datasets, namely RVL-CDIP-D and Tobacco3482-D, to evaluate the robustness of existing state-of-the-art document image classifiers to different types of data distortions that are commonly encountered in the real world. The proposed benchmarks are generated by inserting 21 different types of data distortions with varying severity levels into the well-known document datasets RVL-CDIP and Tobacco3482, respectively, which are then used to quantitatively evaluate the impact of the different distortion types on the performance of latest document image classifiers. In doing so, we show that while the higher accuracy models also exhibit relatively higher robustness, they still severely underperform on some specific distortions, with their classification accuracies dropping from ~90% to as low as ~40% in some cases. We also show that some of these high accuracy models perform even worse than the baseline AlexNet model in the presence of distortions, with the relative decline in their accuracy sometimes reaching as high as 300-450% that of AlexNet. The proposed robustness benchmarks are made available to the community and may aid future research in this area.

Open-vocabulary learning has recently gained prominence as a means to enable image segmentation for arbitrary categories based on textual descriptions. This advancement has extended the applicability of segmentation systems to a broader range of generally purpose scenarios. However, current methods often revolve around specialized architectures and parameters tailored to specific segmentation tasks, resulting in a fragmented landscape of segmentation models. In response to these challenges, we introduce OVAMTSeg, a versatile framework designed for Open-Vocabulary and Multitask Image Segmentation. OVAMTSeg harnesses adaptive prompt learning to empower the model to capture category-sensitive concepts, enhancing its robustness across diverse multi-task and scenario contexts. Text prompts are employed to effectively capture semantic and contextual features of the text, while cross-attention and cross-modal interactions enable the fusion of image and text features. Furthermore, a transformer-based decoder is incorporated for dense prediction. Extensive experimental results underscore the effectiveness of OVAMTSeg, showcasing its state-of-the-art performance and superior generalization capabilities across three segmentation tasks. Notable achievements include a 47.5 mIoU in referring expression segmentation, 51.6 mIoU on Pascal-VOC with four unseen classes, 46.6 mIoU on Pascal-Context in zero-shot segmentation, 65.9 mIoU on Pascal-5i, and 35.7 mIoU on COCO-20i datasets for one-shot segmentation.

Reducing the costs of repairing concrete structures damaged due to the appearance of cracks and reducing the number of people involved in the process of their repair is the subject of a multitude of experimental studies. Special emphasis should be placed on research involving industrial by-products, the disposal of which has a negative environmental impact, as is the case in the research presented in this paper. The basic idea was to prepare a mortar with added granulated blast furnace slag from Smederevo Steel Mill and then treat artificialy produced cracks with a Sporosarcina pasteurii DSM 33 suspension under the conditions of sterile demineralized and water from Danube river in order to simulate natural conditions. The results show a bio-stimulated healing efficiency of 32.02% in sterile demineralized water and 42.74% in Danube water already after 14 days. The SEM images clearly show calcium carbonate crystals as the main compound that has started to fill the crack, and the crystals are much more developed under the Danube water conditions. As a special type of research, microscopic images of cracks were classified into those with and without the presence of bacterial culture. By applying convolutional neural networks (ResNet 50), the classification success rate was 91.55%.

The paper describes the new method for image compression based on coordinated group signal transformation. The algorithm is a type of difference coding. Coordinated processing significantly simplifies the difference signal conversion scheme using a single group codec for all signals. The described method considers color channels as correlated signals of a multi-channel communication system. The specifics of the processed data required modification of the coordinated group processing algorithm. First, we changed how to calculate the difference signals to prevent data loss. Secondly, the group codec was supplemented with a neural network to improve the quality of reconstructed images. We considered the following types of neural networks: fully connected, recurrent, convolution, and convolution in the Fourier space. Based on the simulation results, it is best to use fully connected neural networks if the goal is to minimize processing delay time. These networks have a response time of 13 ms. On the other hand, if the priority is to improve quality in cases where delays are not critical, then convolution neural networks in the Fourier space should be used. These networks have a response time of 140 ms and are of significant interest.

Multimodal remote sensing data classification can enhance the model’s ability to distinguish land features through multimodal data fusion. In this context, how to help models understand the relationship between multimodal data and target tasks is the focus of researchers. Inspired by human feedback learning mechanisms, causal reasoning mechanisms, and knowledge induction mechanisms, this paper integrates causal learning, reinforcement learning, and meta learning into a unified remote sensing data classification framework and proposed the causal meta-reinforcement learning (CMRL). First, based on feedback learning mechanisms, we have overcome the limitations of traditional implicit optimization of fusion features and customized a reinforcement learning environment for multimodal remote sensing data classification tasks. Through feedback interactive learning between agents and the environment, we help them understand the complex relationships between multimodal data and labels, thereby achieving full mining of multimodal complementary information. Second, based on the causal inference mechanism we designed causal distribution prediction action, classification reward, and causal intervention reward, capturing pure causal factors in multimodal data and cutting off false statistical associations between non-causal factors and class labels. Finally, based on the knowledge induction mechanism, we designed a bi-layer optimization mechanism based on meta-learning. By constructing a meta training task and meta validation task simulation model in the generalization scenario of unseen data, we helped the model induce cross-task shared knowledge, thereby improving its generalization ability for unseen multimodal data. The experimental results on multiple sets of multimodal datasets show that the proposed method achieved state-of-the-art performance in multimodal remote sensing data classification tasks.

Color image binarization plays a pivotal role in image preprocessing work, significantly impacting subsequent tasks, particularly in text recognition. This paper concentrates on Document Image Binarization (DIB), aiming to separate a image into foreground (text) and background (non-text content). Through a thorough analysis of conventional and deep learning-based approaches, we conclude that prevailing DIB methods leverage deep learning technology. Furthermore, we explore the receptive fields pre- and post-network training to underscore the Transformer model's advantages. Subsequently, we introduce a lightweight model based on the U-Net structure, enhanced with the Mobile ViT module to better capture global information features in document images. Given its adeptness at learning both local and global features, our proposed model exhibits superior performance on two standard datasets (DIBCO2012 and DIBCO2017) compared to state-of-the-art methods. Notably, our proposed DIB method presents a straightforward end-to-end model devoid of additional image preprocessing or post-processing. Moreover, its parameter count is less than a quarter of the HIP'23 model, which achives best results on three datasets(DIBCO2012, DIBCO2017 and DIBCO2018). Finally, two sets of ablation experiments were conducted to verify the effectiveness of the proposed binarization model.

The present study comprises the second part of our previous work that dealt mainly with the phytochemical and physicochemical characterization of green coffee, clove, cinnamon/clove and nutmeg ethanolic extracts of grape origin. In the present study, we focused on producing a discriminating model concerning green coffee, clove, cinnamon, cinnamon and clove, and nutmeg fine powders based on multivariate analysis of variance and supervised learning on data of volatile compounds analysis, carried out with solid phase dynamic extraction in combination with gas chromatography/mass spectrometry. Results showed that 7 volatile compounds namely ethylene, methanol, 3-methylpentane, ethyl acetate, 9-hexadecen-1-ol, toluene, and methyl acetate could differentiate the investigated samples resulting in a 100% classification rate using the cross-validation method of linear discriminant analysis. Results were further confirmed using partial least squares regression analysis. The study contributes to the typification of green coffee, cinnamon, clove, cinnamon and clove and nutmeg, based on selected volatile compounds, and provides further support to the literature by means of possible adulteration given the statistical models developed.

This paper presents a two-stage hierarchical neural network using image classification and object detection algorithms as key building blocks for a system that automatically detects a potential design right infringement. This neural network is trained to return the Top-N original design right records that highly resemble the input image of a counterfeit. Design rights specify the unique aesthetic characteristics of a product. Due to the rapid change of trends, new design rights are continuously generated. This work proposes an Ensemble Neural Network (ENN), an artificial neural network model that aims to deal with a large amount of counterfeit data and design right records that are frequently added and deleted. At first, we performed image classification and objection detection learning per design right using the existing models with a proven track record of high accuracy. The distributed models form the backbone of the ENN and yield intermediate results aggregated at a master neural network. This master neural network is a deep residual network paired with a fully connected network. This ensemble layer is trained to determine the sub-models that return the best result for a given input image of a product. In the final stage, the ENN model multiples the inferred similarity coefficients to the weighted input vectors produced by the individual sub-models to assess the similarity between the test input image and the existing product design rights to see any sign of violation. Given 84 design rights and the sample product images taken meticulously under various conditions, our ENN model achieved average Top-1 and Top-3 accuracies of 98.409% and 99.460%, respectively. Upon introducing new design rights data, a partial update of the inference model was done an order of magnitude faster than the single model. ENN maintained a high level of accuracy as it scaled out to handle more design rights. Therefore, the ENN model is expected to offer practical help to the inspectors in the field, such as the customs at the border that deal with a swarm of products.

Motion related human activity recognition using wearable sensors can potentially enable various useful daily applications. So far, most studies view it as a stand-alone mathematical classification problem without considering the physical nature of human motions. Consequently, they suffer from data dependencies and encounter the dimension disaster problem and the over-fitting issue, and their models are never human-readable. In this study, we start from a deep analysis on natural physical properties of human motions, and then propose a useful feature selection method to quantify each feature's classification contribution capability. On one hand, the "dimension disaster" problem can be avoid to some extent, due to the affined dimension of key features; On the other hand, over-fitting issue can be depressed since the knowledge implied in human motions are nearly invariant, which compensates the possible data inadequacy. The experiment results indicate that the proposed method performs superior to those adopted in related works, such as decision tree, k-NN, SVM, neural networks.

To address three important issues related to extraction of water features from Landsat imagery, i.e., selection of water indexes and classification algorithms for image classification, collection of ground truth data for accuracy assessment, this study applied four sets (ultra-blue, blue, green, and red light based) of water indexes (NWDI, MNDWI, MNDWI2, AWEI_ns, and AWEI_s) combined with three types of image classification methods (zero-water index threshold, Otsu, and kNN) to 24 selected lakes across the globe to extract water features from Landsat-8 OLI imagery. 1440 (4x5x3x24) image classification results were compared with the extracted water features from high resolution Google Earth images with the same (or ±1 day) acquisition dates through computing the Kappa coefficients. Results show the kNN method is better than the Otsu method, and the Otsu method is better than the zero-water index threshold method. If the computational cost is not an issue, the kNN method combined with the ultra-blue light based AWEI_ns is the best method for extracting water features from Landsat imagery because it produced the highest Kappa coefficients. If the computational cost is taken into account, the Otsu method is a good choice. AWEI_ns and AWEI_s are better than NDWI, MNDWI and MNDWI2. AWEI_ns works better than AWEI_s under the Otsu method, and the average rank of the image classification accuracy from high to low is the ultra-blue, blue, green, and red light-based AWEI_ns.

This study introduces SkinLiTE, a lightweight supervised contrastive learning model, tailored to enhance the detection and typification of skin lesions in dermoscopic images. The core of SkinLiTE lies in its unique integration of supervised and contrastive learning approaches, which leverages labeled data to learn generalizable representations. This approach is particularly adept at handling the complexities and imbalances inherent in skin lesions datasets. The methodology encompasses a two-phase learning process. In the first phase, SkinLiTE utilizes an encoder network and a projection head to transform and project dermoscopic images into a feature space where contrastive loss is applied, focusing on minimizing intra-class variations while maximizing inter-class differences. The second phase freezes the encoder's weights, leveraging the learned representations for classification through a series of dense and dropout layers. The model was evaluated using three datasets from Skin Cancer ISIC 2019-2020, covering a wide range of skin conditions. SkinLiTE demonstrated superior performance across various metrics—accuracy, AUC , and F1 score—particularly when compared with traditional supervised learning models. Notably, SkinLiTE achieved an accuracy of 0.9087 using AugMix augmentation for binary classification of skin lesions. It also showed comparable results with the state-of-the-art approaches of ISIC challenge without relying on external data, underscoring its efficacy and efficiency. The results highlight the potential of SkinLiTE as a significant step forward in the field of dermatological AI, offering a robust, efficient, and accurate tool for skin lesion detection and classification. Its lightweight architecture and ability to handle imbalanced datasets make it particularly suited for integration into Internet of Medical Things environments, paving the way for enhanced remote patient monitoring and diagnostic capabilities. This research contributes to the evolving landscape of AI in healthcare, demonstrating the impact of innovative learning methodologies in medical image analysis.

Precision agriculture is a farm management technology that involves sensing and then responding to the observed variability in the field. Remote sensing is one of the tools of precision agriculture. The emergence of small unmanned aerial vehicles (sUAV) have paved the way to accessible remote sensing tools for farmers. This paper describes the comparison of two popular off-the-shelf sUAVs: 3DR Iris and DJI Phantom 2. Both units are equipped with a camera gimbal attached with a GoPro camera. The comparison of the two sUAV involves a hovering test and a rectilinear motion test. In the hovering test, the sUAV was allowed to hover over a known object and images were taken every second for two minutes. The position of the object in the images was measured and this was used to assess the stability of the sUAV while hovering. In the rectilinear test, the sUAV was allowed to follow a straight path and images of a lined track were acquired. The lines on the images were then measured on how accurate the sUAV followed the path. Results showed that both sUAV performed well in both the hovering test and the rectilinear motion test. This demonstrates that both sUAVs can be used for agricultural monitoring.

Numerous image processing techniques (IPTs) have been employed to detect crack defects, offering an alternative to human-conducted onsite inspections. These IPTs manipulate images to extract defect features, particularly cracks in surfaces produced through Additive Manufacturing (AM). This article presents a vision-based approach that utilizes deep convolutional neural networks (CNNs) for crack detection in AM surfaces. Traditional image processing techniques face challenges with diverse real-world scenarios and varying crack types. To overcome these challenges, our proposed method leverages CNNs, eliminating the need for extensive feature extraction. Annotation for CNN training is facilitated by LabelImg without the requirement for additional IPTs. The trained CNN, enhanced by OpenCV preprocessing techniques, achieves an outstanding 99.54% accuracy on a dataset of 14,982 annotated images with resolutions of 1536 × 1103 pixels. Evaluation metrics exceeding 96% precision, 98% recall, and a 97% F1-score highlight the precision and effectiveness of the entire process.

Cloud-induced atmospheric extinction and occlusion have a major effect on the effectiveness and quality of telescope observations. Real-time cloud cover distribution and long-term statistical data are essential for astronomical siting and telescope operations. At ground-based astronomical telescope sites, cloud cover distribution is currently analyzed using manual observation methods. However, the main disadvantages of manual observation methods are human subjective, heavy workloads and poor real-time performance. Therefore, a real-time automatic cloud images classification method is desperately needed. This paper presents a new cloud identification method, which is named PSO+XGBoost model by combining eXtreme Gradient Boosting (XGBoost) and Particle Swarm Optimization (PSO). The entire cloud image is divided into 37 sub-regions in order to more precisely identify the distribution of the clouds. 19 features are then extracted, including the sky background, star density, lighting conditions, and subregion grayscale values. The experimental results have shown that the overall classification accuracy of 96.91% and our model is able to outperform several state-of-the-art baseline methods. Our approach achieves high accuracy in comparison with the manual observation methods. Moreover, this method meets telescope real-time scheduling requirements.

We present a method to improve the reconstruction and generation performance of variational autoencoder (VAE) by injecting an adversarial learning. On the other hand, instead of comparing the reconstructed with the original data to calculate the reconstruction loss, we use a consistency principle for deep features. The training process of the VAE is then divided into two steps, training the encoder and then training the decoder. By using this two-step learning process, our method can be more widely used in applications other than image processing. While training the encoder, the label information is integrated to better structure the latent space in a supervised way. The adversarial constraints allow the decoder to generate data with better authenticity and more realistic than the conventional VAE. We present experimental results to show that our method gives better performance than the original VAE.

This paper proposes a new approach for multi-focus images fusion based on Region Mosaicing on Contrast Pyramids (REMCP). A density-based region growing method is developed to construct a focused region mask for multi-focus images. The segmented focused region mask is decomposed into a mask pyramid, which is then used for supervised region mosaicking on a contrast pyramid. In this way, the focus measurement and the continuity of focused regions are incorporated and the pixel level pyramid fusion is improved at the region level. Objective and subjective experiments show that the proposed REMCP is more robust to noise than compared algorithms and can fully preserves the focus information of the multi-focus images meanwhile reducing distortions of the fused images.

Breast cancer stands as a predominant health concern for women globally. As mammography is the primary screening tool for breast cancer detection, improving the detection of breast cancer at screening could save more lives. This mammography re-view paper provides a systematic review of computer-aided techniques during a specific time frame for the segmentation and classification of microcalcification, evaluating image processing, machine learning, and deep learning techniques. The systematic review is meticulously carried out, adhering closely to the preferred reporting items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines. This article focuses on mammographic breast cancer detection approaches based on automated systems, discussed chronologically from 1970 through 2023. This article encompasses the breadth of artificial intelligence-based methods from the most primitive to the most sophisticated models. Image processing and machine learning-based methods are compre-hensively reviewed. The evaluation of a deep learning architecture based on self-extracted features for classification tasks demonstrated outclass performance. Large-scale datasets required for a broader and in-depth analysis of novel methods for breast cancer detection are also discussed in this article. This research work is aligned with the United Nations' sustainability de-velopment goals.

Many types of banana (Musa spp.)—one of the most economically important horticultural crops—exist, and their ploidy (diploid, triploid, and tetraploid, usually) and genome types (containing A or/and B genome in most cases) differ. At present, observation and genome type detection are commonly used to identify banana germplasm resources. However, the former is tedious, while the latter cannot distinguish categories below genome types. It is therefore urgent that a simple and effective method for identifying banana germplasm resources is established. We sequenced and analyzed the ribosomal DNA internal transcribed spacer (ITS) sequences of 62 banana germplasm and found that the sequencing peaks, especially the 20 bp region near the 420 bp position (referred to as the “420 bp region”), exhibited relatively recognizable and repeatable polymorphism characteristics. Using the “420 bp region” as a marker, we were able to quickly distinguish bananas belonging to different genome type groups (comprising the AA, AB, AAA, ABB, AAB, ABBB, AAAB, and AAAA groups) or different subgroups in the same genome type group (for example, the Cavendish, Gros Michel, Red, Lakatan, and Ibota Bota subgroups in the AAA group; and the Pisang Raja, Plantain, Pome and Silk subgroups in the AAB group). Moreover, it appeared that Sanger sequencing of ITS could be used for identifying banana hybrid offspring. These results demonstrated that the banana ITS region contains rich genetic information that is useful for genotyping. In general, ITS sequencing simplifies the classification of banana germplasm resources and has potential application in several areas of Musa improvement.

This paper describes the first results of an Image Recognition Based Location (IRBL) for mobile application focusing on the procedure to generate a Database of range images (RGB-D). In an indoor environment, to estimate the camera position and orientation, a prior spatial knowledge of the surrounding is needed. In order to achieve this objective a complete 3D survey of two different environment (Bangbae metro station of Seoul and E.T.R.I. building in Daejeon – Republic of Korea) was performed using LiDAR (Light Detection And Ranging) instrument and the obtained scans were processed in order to obtain a spatial model of the environments. From this, two databases of reference images were generated using a specific software realized by the Geomatics group of Politecnico di Torino (ScanToRGBDImage). This tool allow to generate synthetically different RGB-D images) centered in the each scan position in the environment. Later, the external parameters (X, Y, Z, ω, φ, κ) and the range information extracted from the DB images retrieved, are used as reference information for pose estimation of a set of acquired mobile pictures in the IRBL procedure. In this paper the survey operations, the approach for generating the RGB-D images and the IRB strategy are reported. Finally the analysis of the results and the validation test are described.

With the development of deep learning, several Graph Neural Networks (GNN)-based approaches have been utilized for text classification. However, GNNs encounter challenges in capturing contextual text information within a document sequence. To address this, a novel text classification model RB-GAT is proposed by combining RoBERTa-BiGRU embedding and a multi-head Graph ATtention Network (GAT). First, the pre-trained RoBERTa model is exploited to learn word and text embeddings in different contexts. Second, the Bidirectional Gated Recurrent Unit (BiGRU) is employed to capture long-term dependencies and bidirectional sentence information from the text context. Next, the multi-head graph attention network is applied to analyze this information, which serves as a node feature for the document. Finally, the classification results are generated through a Softmax layer. Experimental results on three benchmark datasets demonstrate that our method can achieve an accuracy of 71.48%, 98.45%, and 80.32% on Ohsumed, R8, and MR, which is superior to the existing nine text classification approaches.

Rapid and accurate detection of road cracks is of great significance for road health monitoring, but currently this work is mainly completed through manual site surveys. Low-altitude UAV remote sensing can obtain images with centimeter or even subcentimeter ground resolution, which provides a new efficient and economical approach for rapid crack detection. Nevertheless, crack detection networks face challenges such as edge blurring and misidentification due to the heterogeneity of road cracks and the complexity of the background. To address these issues, we propose a real-time edge reconstruction crack detection network (ERNet) which adopts multi-level information aggregation to reconstruct crack edges and improve the accuracy of segmentation between the target and background. To capture global dependencies across spatial and channel levels, we propose an efficient bilateral decomposed convolutional attention module (BDAM) that combines depth separable convolution and dilated convolution to capture global dependencies across spatial and channel levels. To enhance the accuracy of crack detection, we use a coordinate-based fusion module that integrates spatial, semantic, and edge reconstruction information. In addition, we propose an automatic measurement of crack information for extracting the crack trunk and its corresponding length and width. Experimental results demonstrate that our network achieves the best balance between accuracy and inference speed.

As the fast growth of information technology (IT), the safety of image transmission and the storing of images are increasingly concerned. Traditional image encryption algorithms have certain limitations in transmission and security, so there is an urgent need for a secure and reliable image encryption algorithm. A universal compression sensing (CS) image encryption (IE) algorithm based on DNA-triploid mutation (DTM) is presented in this paper. Firstly, by using CS algorithm an image is compressed while obtaining a range of chaotic sequences by iteration of chaotic map. Then DNA sequences are generated by encoding the image, and based on the DTM, new mutant DNA sequences are generated according to specific rules. Next, the chaotic sequences are operated at the DNA level to do confusion and diffusion operations on the image to ensure the security of the data. At last, DNA decoding is carried out to obtain the compressed-encrypted image. Both simulation experiments and performance tests fully show that a high level of security and reliability of the proposed algorithm in protecting image privacy is achieved.

In order to improve contrast and restore color for underwater image captured by camera sensors without suffering from insufficient details and color cast, a fusion algorithm for image enhancement in different color spaces based on contrast limited adaptive histogram equalization (CLAHE) is proposed in this article. The original color image is first converted from RGB color space to two different special color spaces: YIQ and HSI. The color space conversion from RGB to YIQ is a linear transformation, while the RGB to HSI conversion is nonlinear. Then, the algorithm separately operates CLAHE in YIQ and HSI color spaces to obtain two different enhancement images. The luminance component (Y) in the YIQ color space and the intensity component (I) in the HSI color space are enhanced with CLAHE algorithm. The CLAHE has two key parameters: Block Size and Clip Limit, which mainly control the quality of CLAHE enhancement image. After that, the YIQ and HSI enhancement images are respectively converted backward to RGB color. When the three components of red, green, and blue are not coherent in the YIQ-RGB or HSI-RGB images, the three components will have to be harmonized with the CLAHE algorithm in RGB space. Finally, with 4 direction Sobel edge detector in the bounded general logarithm ratio operation, a self-adaptive weight selection nonlinear image enhancement is carried out to fuse YIQ-RGB and HSI-RGB images together to achieve the final fused image. The enhancement fusion algorithm has two key factors: average of Sobel edge detector and fusion coefficient, and these two factors determine the effects of enhancement fusion algorithm. A series of evaluate metrics such as mean, contrast, entropy, colorfulness metric (CM), mean square error (MSE) and peak signal to noise ratio (PSNR) are used to assess the proposed enhancement algorithm. The experiments results showed that the proposed algorithm provides more detail enhancement and higher values of colorfulness restoration as compared to other existing image enhancement algorithms. The proposed algorithm can suppress effectively noise interference, improve the image quality for underwater image availably.

In this paper, we propose an imagined speech-based brain wave pattern recognition using deep learning. Multiple features were extracted concurrently from eight-channel Electroencephalography (EEG) signals. To obtain classifiable EEG data with fewer number of sensors, we placed the EEG sensors on carefully selected spots on the scalp. To decrease the dimensions and complexity of the EEG dataset and to avoid overfitting during the deep learning algorithm, we utilized the wavelet scattering transformation. A low-cost 8-channel EEG headset was used with MATLAB 2023a to acquire the EEG data. The Long-Short Term Memory Recurrent Neural Network (LSTM-RNN) was used to decode the identified EEG signals into four audio commands: Up, Down, Left, and Right. Wavelet scattering transformation was applied to extract the most stable features by passing the EEG dataset through a series of filtration processes. Filtration has been implemented for each individual command in the EEG datasets. The proposed imagined speech-based brain wave pattern recognition approach achieved a 92.50% overall classification accuracy. This accuracy is promising for designing a trustworthy imagined speech-based Brain-Computer Interface (BCI) future real-time systems. For better evaluation of the classification performance, other metrics were considered, and we obtained 92.74%, 92.50% and 92.62% for precision, recall, and F1-score, respectively.

With the development of deep learning, the Super-Resolution (SR) reconstruction of microscopic images has improved significantly. However, the scarcity of microscopic images for training, the underutilization of hierarchical features in original Low Resolution (LR) images, and the high-frequency noise unrelated with the image structure generated during reconstruction process are still challenges in the Single Image Super-Resolution (SISR) field. Faced with these issues, we first collected sufficient microscopic images through Motic, a company engaged in the design and production of optical and digital microscopes, to establish a dataset. Secondly, we proposed a Residual Dense Attention Generative Adversarial Network (RDAGAN). The network comprises a generator, an image discriminator, and a feature discriminator. The generator includes a Residual Dense Block (RDB) and a Convolutional Block Attention Module (CBAM), focusing on extracting the hierarchical features of the original LR image. Simultaneously, the added feature discriminator enables the network to generate high-frequency features pertinent to the image’s structure. Finally, we conducted experimental analysis and compared our model with six classic models. Compared with the best model, our model improved PSNR and SSIM by about 1.5dB and 0.2, respectively.

The Range–Doppler (R-D) model is extensively employed for the geometric processing of Synthetic Aperture Radar (SAR) images. Refining the sensor motion state and imaging parameters is the most common method for achieving high-precision geometric processing using the R-D model, comprising a process that involves numerous parameters and complex computations. In this study, we introduce a novel approach to improve the positioning accuracy of the R-D model, implementing a low-order polynomial to compensate for the original imaging errors without altering the initial positioning parameters. We also design low-order polynomial compensation models with different parameters. The models are evaluated on various SAR images from different platforms and bands, including spaceborne TerraSAR-X and GF3-C images, manned airborne SAR-X images, and unmanned aerial vehicle-mounted miniSAR-Ku images. Furthermore, image positioning experiments involving the use of different polynomial compensation models and various numbers and distributions of Ground Control Points (GCPs) are conducted. The experimental results demonstrate that geometric processing accuracy comparable to that of the classical rigorous positioning method can be achieved even when applying only an affine transformation model to the images. Compared to classical refinement models, however, the proposed image-compensated R-D model is much simpler and easy to implement. Thus, this study provides a convenient, robust, and widely applicable method for the geometric positioning processing of SAR images, offering a potential approach for the joint-positioning processing of multi-source SAR images.

Significant progress has made in pattern recognition technology. However, one obstacle that has not yet overcome is the recognition of words in the Brahmi script, specifically the recognition of characters, compound characters, and word because of complex structure. For this kind of complex pattern recognition problem, it is always difficult to decide which feature extraction and classifier would be the best choice. Moreover, it is also true that different feature extraction and classifiers offer complementary information about the patterns to be classified. Therefore, combining feature extraction and classifiers, in an intelligent way, can be beneficial compared to using any single feature extraction. This study proposed the combination of HOG +zonal density with SVM to recognize the Brahmi words. Keeping these facts in mind, in this paper, information provided by structural and statistical based features are combined using SVM classifier for script recognition (word-level) purpose from the Brahmi words images. Brahmi word dataset contains 6,475 and 536 images of Brahmi words of 170 classes for the training and testing, respectively, and the database is made freely available. The word samples from the mentioned database are classified based on the confidence scores provided by support vector machine (SVM) classifier while HOG and zonal density use to extract the features of Brahmi words. Maximum accuracy suggested by system is 95.17% which is better than previously suggested studies.

Fractional calculus gained a lot of attention in the last couple of years. Researchers discovered that processes in various fields follow rather fractional dynamics than ordinary integer-ordered dynamics, meaning the corresponding differential equations feature non-integer valued derivatives. There are several arguments for why this is the case, one of them being that fractional derivatives’ inherit spatiotemporal memory and/or the ability to express complex naturally occurring phenomena. Another popular topic nowadays is machine learning, i.e., learning behavior and patterns from historical data. In our ever-changing world with ever-increasing amounts of data, machine learning is a powerful tool for data analysis, problem-solving, modeling, and prediction. It further provides many insights and discoveries in various scientific disciplines. As these two modern-day topics provide a lot of potential for combined approaches to describe complex dynamics, this article reviews combined approaches of fractional derivatives and machine learning from the past, puts them into context, and thus provides a list of possible combined approaches and the corresponding techniques. Note, however, that this article does not deal with neural networks, as there already is profound literature on neural networks and fractional calculus. We sorted past combined approaches from the literature into three categories, i.e., preprocessing, machine learning & fractional dynamics, and optimization. The contributions of fractional derivatives to machine learning are manifold as they provide powerful preprocessing and feature augmentation techniques, can improve physically informed machine learning, and are capable of improving hyperparameter optimization. Thus, this article serves to motivate researchers dealing with data-based problems, to be specific machine learning practitioners, to adopt new tools and enhance their existing approaches.

Incorrect sitting posture, characterized by asymmetrical or uneven positioning of the body, often leads to spinal misalignment and muscle tone imbalance. Prolonged maintenance of such postures can adversely impact well-being and contribute to the development of spinal deformities and musculoskeletal disorders. In response, smart-sensing chairs equipped with cutting-edge sensor technologies have been introduced as a viable solution for real-time detection, classification, and monitoring of sitting postures, aiming to mitigate the risk of musculoskeletal disorders and promote overall health. This comprehensive literature review evaluates the current body of research on smart-sensing chairs, with a specific focus on the strategies used for posture detection and classification, as well as the effectiveness of different sensor technologies. A meticulous search across MDPI, IEEE, and Google Scholar databases yielded 34 pertinent studies that utilize non-invasive methods for posture monitoring. The analysis reveals that Force Sensing Resistors (FSR) are the predominant sensors utilized for posture detection, whereas Convolutional Neural Networks (CNN) and Artificial Neural Networks (ANN) are the leading machine learning models for posture classification. However, it was observed that CNNs and ANNs do not outperform traditional statistical models in terms of classification accuracy, due to constrained size and lack of diversity within training datasets. These datasets often fail to comprehensively represent the array of human body shapes and musculoskeletal configurations. Moreover, this review identifies a significant gap in the evaluation of user feedback mechanisms, essential for alerting users to their sitting posture and facilitating corrective adjustments.

In this paper, a seamless connection between sensor technology and machine learning (ML) has been presented which has wide applicability in rehabilitation and physical medicine. The intent is to identify an ML algorithm which has a better performance, easy to deploy, and has less training time and consumes less computational resources. In short, we propose an algorithm which maximizes the synergy between a sensor and ML by maximizing classification performance and easy to do implementation, which is robust in computational power and lower in resource consumption. The premise of the work is to show the important considerations for evaluating ML algorithms for Upper-limb activity recognition for Physical medicine and rehabilitation field.

The ability to detect the presence as well as classify the activities of individuals behind visually obscuring structures is of significant benefit to police, security and emergency services in many situations. This paper presents the analysis from a series of experimental results generated using a through-the-wall (TTW) Frequency Modulated Continuous Wave (FMCW) C-Band radar system named Soprano. The objective of this analysis was to classify whether an individual was carrying an item in both hands or not using micro-Doppler information from a FMCW sensor. The radar was deployed at a standoff distance, of approximately 0.5 m, outside a residential building and used to detect multiple people walking within a room. Through the application of digital filtering, it was shown that significant suppression of the primary wall reflection is possible, significantly enhancing the target signal to clutter ratio. Singular Value Decomposition (SVD) signal processing techniques were then applied to the micro-Doppler signatures from different individuals. Features from the SVD information have been used to classify whether the person was carrying an item or walking free handed. Excellent performance of the classifier was achieved in this challenging scenario with accuracies up to 94%, suggesting that future through wall radar sensors may have the ability to reliably recognize many different types of activities in TTW scenarios using these techniques.

Domain generalization has led to remarkable achievements in Hyperspectral Image (HSI) classification. Inspired by contrastive language-image pre-training (CLIP), the language-aware domain generalization method has been explored to learn cross-domain-invariant representation. However, existing methods face some challenges: 1) The weak capacity to extract long-range contextual information and inter-class correlation. 2) Due to the inadequacies of the large-scale pre-training for HSI data, the spatial-spectral features of HSI and linguistic features can not be straightforwardly alignment. To address the above problems, a novel network has been proposed with a CLIP framework, which consists of an image encoder, based on an encoder-only transformer to obtain the global contextual information and inter-class correlation, a frozen text encoder, and a cross-attention mechanism, named Linguistic-Interact-with-Visual Engager (LIVE), enhances the interaction between two modalities. Extensive experiments demonstrating superior performance over state-of-the-art methods in HSI Domain Generalization with a CLIP framework.

Accurate prostate volume estimation is crucial for effective prostate disease management. Ultrasound (US) imaging, particularly transrectal ultrasound, offers a cost-effective and rapid assessment. However, US images often suffer from artifacts and poor contrast, making prostate volume estimation challenging. This review explores recent advancements in deep learning (DL) techniques for automatic prostate segmentation in US images as a primary step for prostate volume estimation. We examine various DL architectures, including traditional U-Net modifications and innovative designs incorporating residual connections, multi-directional image data, and attention mechanisms. Additionally, we discuss pre-processing methods to enhance image quality, the integration of shape information, and strategies to improve the consistency and robustness of DL models. The effectiveness of these techniques is evaluated through metrics such as Dice Similarity Coefficient, Jaccard Index, and Hausdorff Distance. The review highlights the potential of DL in improving prostate volume estimation accuracy and reducing clinical workload, while also identifying areas for future research to enhance model performance and generalizability.

In this paper we propose a novel self-supervised approach of keywords and keyphrases retrieval and extraction by an end-to-end deep learning approach, which is trained by contextually self-labelled corpus. Our proposed approach is novel to use contextual and semantic features to extract the keywords and has outperformed the state of the art. Through the experiment the proposed approach has been proved to be better in both semantic meaning and quality than the existing popular algorithms of keyword extraction. In addition, we propose to use contextual features from bidirectional transformers to automatically label short-sentence corpus with keywords and keyphrases to build the ground truth. This process avoids the human time to label the keywords and do not need any prior knowledge. To the best of our knowledge, our published dataset in this paper is a fine domain-independent corpus of short sentences with labelled keywords and keyphrases in the NLP community.

Background: Microscopic examination of acid-fast stained sputum smears for detecting mycobacterial acid-fast bacilli (AFB) remains the most economical and readily available method for laboratory diagnosis of tuberculosis (TB). However, this conventional approach has limitations, including low sensitivity and labor-intensive procedures. Methods: An automated microscopy system incorporating artificial intelligence (AI) and machine learning for AFB identification was evaluated. The study was conducted at an Infectious Disease Hospital in Jiangsu Province, China, utilizing an intelligent microscope system (TB-Scan, Wellgen Medical, Kaohsiung). A total of 1,000 sputum smears were included in the analysis, with the system capturing digital microscopic images and employing an image recognition model to automatically identify and classify AFB. Referee technicians provided the gold standard for discrepant results. Results: The automated system demonstrated an overall accuracy of 95.00% (950/1,000), sensitivity of 91.24% (177/194), and specificity of 95.91% (773/806). Notably, the system identified 21 smears as positive that were previously reported as negative, with referee technicians confirming 17 of these as true positive and retracting the test results accordingly. Recalculating the performance, the accuracy increased to 96.70% (967/1,000), sensitivity to 91.94% (194/211), and specificity to 97.97% (773/789), with a false negative rate of 8.06% (17/211) and a false positive rate of 2.03% (16/789). Conclusions: The incorporation of AI and machine learning into an automated microscopy system demonstrated the potential to enhance the sensitivity and efficiency of AFB detection in sputum smears compared to conventional manual microscopy. This approach holds promise for widespread application in TB diagnostics and potentially other fields requiring labor-intensive microscopic examination.