Attentional Modulation Research Articles

TransRAUNet: A Deep Neural Network with Reverse Attention Module Using HU Windowing Augmentation for Robust Liver Vessel Segmentation in Full Resolution of CT Images.

Background: Liver cancer has a high mortality rate worldwide, and clinicians segment liver vessels in CT images before surgical procedures. However, liver vessels have a complex structure, and the segmentation process is conducted manually, so it is time-consuming and labor-intensive. Consequently, it would be extremely useful to develop a deep learning-based automatic liver vessel segmentation method. Method: As a segmentation method, UNet is widely used as a baseline, and a multi-scale block or attention module has been introduced to extract context information. In recent machine learning efforts, not only has the global context extraction been improved by introducing Transformer, but a method to reinforce the edge area has been proposed. However, the data preprocessing step still commonly uses general augmentation methods, such as flip, rotation, and mirroring, so it does not perform robustly on images of varying brightness or contrast levels. We propose a method of applying image augmentation with different HU windowing values. In addition, to minimize the false negative area, we propose TransRAUNet, which introduces a reverse attention module (RAM) that can focus edge information to the baseline TransUNet. The proposed architecture solves context loss for small vessels by applying edge module (RAM) in the upsampling phase. It can also generate semantic feature maps that allows it to learn edge, global context, and detail location by combining high-level edge and low-level context features. Results: In the 3Dricadb dataset, the proposed model achieved a DSC of 0.948 and a sensitivity of 0.944 in liver vessel segmentation. This study demonstrated that the proposed augmentation method is effective and robust by comparisons with the model without augmentation and with the general augmentation method. Additionally, an ablation study showed that RAM has improved segmentation performance compared to TransUNet. Compared to prevailing state-of-the-art methods, the proposed model showed the best performance for liver vessel segmentation. Conclusions: TransRAUnet is expected to serve as a navigation aid for liver resection surgery through accurate liver vessel and tumor segmentation.

Iterative optimization annotation pipeline and ALSS-YOLO-Seg for efficient banana plantation segmentation in UAV imagery.

Precise segmentation of unmanned aerial vehicle (UAV)-captured images plays a vital role in tasks such as crop yield estimation and plant health assessment in banana plantations. By identifying and classifying planted areas, crop areas can be calculated, which is indispensable for accurate yield predictions. However, segmenting banana plantation scenes requires a substantial amount of annotated data, and manual labeling of these images is both timeconsuming and labor-intensive, limiting the development of large-scale datasets. Furthermore, challenges such as changing target sizes, complex ground backgrounds, limited computational resources, and correct identification of crop categories make segmentation even more difficult. To address these issues, we propose a comprehensive solution. First, we designed an iterative optimization annotation pipeline leveraging SAM2's zero-shot capabilities to generate high-quality segmentation annotations, thereby reducing the cost and time associated with data annotation significantly. Second, we developed ALSS-YOLO-Seg, an efficient lightweight segmentation model optimized for UAV imagery. The model's backbone includes an Adaptive Lightweight Channel Splitting and Shuffling (ALSS) module to improve information exchange between channels and optimize feature extraction, aiding accurate crop identification. Additionally, a Multi-Scale Channel Attention (MSCA) module combines multi-scale feature extraction with channel attention to tackle challenges of varying target sizes and complex ground backgrounds. We evaluated the zero-shot segmentation performance of SAM2 on the ADE20K and Javeri datasets. Our iterative optimization annotation pipeline demonstrated a significant reduction in manual annotation effort while achieving high-quality segmentation labeling. Extensive experiments on our custom Banana Plantation segmentation dataset show that ALSS-YOLO-Seg achieves state-of-the-art performance. Our code is openly available at https://github.com/helloworlder8/computer_vision.

Application of MRI image segmentation algorithm for brain tumors based on improved YOLO.

To assist in the rapid clinical identification of brain tumor types while achieving segmentation detection, this study investigates the feasibility of applying the deep learning YOLOv5s algorithm model to the segmentation of brain tumor magnetic resonance images and optimizes and upgrades it on this basis. The research institute utilized two public datasets of meningioma and glioma magnetic resonance imaging from Kaggle. Dataset 1 contains a total of 3,223 images, and Dataset 2 contains 216 images. From Dataset 1, we randomly selected 3,000 images and used the Labelimg tool to annotate the cancerous regions within the images. These images were then divided into training and validation sets in a 7:3 ratio. The remaining 223 images, along with Dataset 2, were ultimately used as the internal test set and external test set, respectively, to evaluate the model's segmentation effect. A series of optimizations were made to the original YOLOv5 algorithm, introducing the Atrous Spatial Pyramid Pooling (ASPP), Convolutional Block Attention Module (CBAM), Coordinate Attention (CA) for structural improvement, resulting in several optimized versions, namely YOLOv5s-ASPP, YOLOv5s-CBAM, YOLOv5s-CA, YOLOv5s-ASPP-CBAM, and YOLOv5s-ASPP-CA. The training and validation sets were input into the original YOLOv5s model, five optimized models, and the YOLOv8s model for 100 rounds of iterative training. The best weight file of the model with the best evaluation index in the six trained models was used for the final test of the test set. After iterative training, the seven models can segment and recognize brain tumor magnetic resonance images. Their precision rates on the validation set are 92.5, 93.5, 91.2, 91.8, 89.6, 90.8, and 93.1%, respectively. The corresponding recall rates are 84, 85.3, 85.4, 84.7, 87.3, 85.4, and 91.9%. The best weight file of the model with the best evaluation index among the six trained models was tested on the test set, and the improved model significantly enhanced the image segmentation ability compared to the original model. Compared with the original YOLOv5s model, among the five improved models, the improved YOLOv5s-ASPP model significantly enhanced the segmentation ability of brain tumor magnetic resonance images, which is helpful in assisting clinical diagnosis and treatment planning.

Distinguishing Multisystem Inflammatory Syndrome in Children From Typhus Using Artificial Intelligence: MIS-C Versus Endemic Typhus (AI-MET)

Abstract Background The pandemic emergent disease multisystem inflammatory syndrome in children (MIS-C) following coronavirus disease-19 infection can mimic endemic typhus. We aimed to use artificial intelligence (AI) to develop a clinical decision support system that accurately distinguishes MIS-C versus endemic typhus (MET). Methods Demographic, clinical, and laboratory features rapidly available following presentation were extracted for 133 patients with MIS-C and 87 patients hospitalized due to typhus. An attention module assigned importance to inputs used to create the 2-phase AI-MET. Phase 1 uses 17 features to arrive at a classification manually (MET-17). If the confidence level is not surpassed, 13 additional features are added to calculate MET-30 using a recurrent neural network. Results While 24 of 30 features differed statistically, the values overlapped sufficiently that the features were clinically irrelevant distinguishers as individual parameters. However, AI-MET successfully classified typhus and MIS-C with 100% accuracy. A validation cohort of 111 additional patients with MIS-C was classified with 99% accuracy. Conclusions Artificial intelligence can successfully distinguish MIS-C from typhus using rapidly available features. This decision support system will be a valuable tool for front-line providers facing the difficulty of diagnosing a febrile child in endemic areas.

Lesion classification and diabetic retinopathy grading by integrating softmax and pooling operators into vision transformer.

Diabetic retinopathy grading plays a vital role in the diagnosis and treatment of patients. In practice, this task mainly relies on manual inspection using human visual system. However, the human visual system-based screening process is labor-intensive, time-consuming, and error-prone. Therefore, plenty of automated screening technique have been developed to address this task. Among these techniques, the deep learning models have demonstrated promising outcomes in various types of machine vision tasks. However, most of the medical image analysis-oriented deep learning approaches are built upon the convolutional operations, which might neglect the global dependencies between long-range pixels in the medical images. Therefore, the vision transformer models, which can unveil the associations between global pixels, have been gradually employed in medical image analysis. However, the quadratic computation complexity of attention mechanism has hindered the deployment of vision transformer in clinical practices. Bearing the analysis above in mind, this study introduces an integrated self-attention mechanism with both softmax and linear modules to guarantee efficiency and expressiveness, simultaneously. To be specific, a portion of query and key tokens, which are much less than the original query and key tokens, are adopted in the attention module by adding a set of proxy tokens. Note that the proxy tokens can fully utilize both the advantages of softmax and linear attention. To evaluate the performance of the presented approach, the comparison experiments between state-of-the-art algorithms and the proposed approach are conducted. Experimental results demonstrate that the proposed approach achieves superior outcome over the state-of-the-art algorithms on the publicly available datasets. Accordingly, the proposed approach can be taken as a potentially valuable instrument in clinical practices.

Online detection of potato drying stages based on improved YOLOv7-tiny model

To realize accurate online identification of different stages of the agricultural product drying process and overcome the limitations of empirical models, this study proposes a method for online identification of agricultural product drying stages based on machine vision, which enhances the YOLOv7-tiny model by adding an attention mechanism module to the feature layer and the up-adoption process. The recognition results were compared and evaluated with those of other versions of YOLO, Faster R-CNN, SSD, EfficientDet, and an unimproved YOLOv7-tiny network. The results showed that the average recognition accuracy of this method for the constant drying stage, first drying stage deceleration and second drying stage deceleration of potato slices reached 98.8%, which was superior to that of the model without the attentional mechanism module. This lays the foundation for the establishment of an on-line adaptive drying model for agricultural products.

Lightweight Deep Learning Model, ConvNeXt-U: An Improved U-Net Network for Extracting Cropland in Complex Landscapes from Gaofen-2 Images.

Extracting fragmented cropland is essential for effective cropland management and sustainable agricultural development. However, extracting fragmented cropland presents significant challenges due to its irregular and blurred boundaries, as well as the diversity in crop types and distribution. Deep learning methods are widely used for land cover classification. This paper proposes ConvNeXt-U, a lightweight deep learning network that efficiently extracts fragmented cropland while reducing computational requirements and saving costs. ConvNeXt-U retains the U-shaped structure of U-Net but replaces the encoder with a simplified ConvNeXt architecture. The decoder remains unchanged from U-Net, and the lightweight CBAM (Convolutional Block Attention Module) is integrated. This module adaptively adjusts the channel and spatial dimensions of feature maps, emphasizing key features and suppressing redundant information, which enhances the capture of edge features and improves extraction accuracy. The case study area is Hengyang County, Hunan Province, China, using GF-2 remote sensing imagery. The results show that ConvNeXt-U outperforms existing methods, such as Swin Transformer (Acc = 85.1%, IoU = 79.1%), MobileNetV3 (Acc = 83.4%, IoU = 77.6%), VGG16 (Acc = 80.5%, IoU = 74.6%), and ResUnet (Acc = 81.8%, IoU = 76.1%), achieving an IoU of 79.5% and Acc of 85.2%. Under the same conditions, ConvNeXt-U has a faster inference speed of 37 images/s, compared to 28 images/s for Swin Transformer, 35 images/s for MobileNetV3, and 0.43 and 0.44 images/s for VGG16 and ResUnet, respectively. Moreover, ConvNeXt-U outperforms other methods in processing the boundaries of fragmented cropland, producing clearer and more complete boundaries. The results indicate that the ConvNeXt and CBAM modules significantly enhance the accuracy of fragmented cropland extraction. ConvNeXt-U is also an effective method for extracting fragmented cropland from remote sensing imagery.

Ship target detection method based on improved YOLOv8 for SAR images

ABSTRACT Due to the complex background and less effective information in low-resolution and noisy images, the detection of small ships in SAR images suffers from low detection rate and high false alarm rate. In order to solve the above problems, this paper proposes a method to detect the small ship targets in SAR images based on improved YOLOv8. Firstly, the deformable convolutional networks are added to the leading network to improve feature extraction. Then, the efficient multi-scale attention module is fused with the backbone network to enhance the detection effect of small targets. The weighted intersection over union loss function is used to optimize the regression process of the prediction frame and the detection frame to enhance the localization capacity of small targets. Finally, there is an addition of a specialized small target detection layer, and a reconstruction of the feature extraction and fusion network to enhance the detection performance of small ship targets in SAR images. To verify the effectiveness and robustness of the method, we conduct experiments on SSDD and HRSID. The proposed method achieves high detection accuracy while also offering a more compact model size and less computation time compared to other prevalent methods.

EMNet: A Novel Few-Shot Image Classification Model with Enhanced Self-Correlation Attention and Multi-Branch Joint Module.

In this research, inspired by the principles of biological visual attention mechanisms and swarm intelligence found in nature, we present an Enhanced Self-Correlation Attention and Multi-Branch Joint Module Network (EMNet), a novel model for few-shot image classification. Few-shot image classification aims to address the problem of image classification when data are limited. Traditional models require a large amount of labeled data for training, while few-shot learning trains models using only a small number of samples (just a few samples per class) to recognize new categories. EMNet shows its potential for bio-inspired algorithms in optimizing feature extraction and enhancing generalization capabilities. It features two key modules: Enhanced Self-Correlated Attention (ESCA) and Multi-Branch Joint Module (MBJ Module). EMNet tackles two main challenges in few-shot learning: how to make an effective important feature extraction and enhancement in images, and improving generalization to new categories. The ESCA module boosts the precision in extracting crucial local features, enhancing classification accuracy. The MBJ module focuses on shared features across images, emphasizing similarities within classes and subtle differences between them. This enhances model adaptability and generalization to new categories. Experimental results show that our model performs better than existing models in one-shot and five-shot tasks on mini-ImageNet, CUB-200, and CIFAR-FS datasets, which proves the proposed model to be an efficient end-to-end solution for few-shot image classification. In the five-way one-shot and five-way five-shot experiments on the CUB-200-2011 dataset, EMNet achieved classification accuracies that were 1.27 and 0.54 percentage points higher than those of RENet, respectively. In the five-way one-shot and five-way five-shot experiments on the miniImageNet dataset, EMNet's classification accuracies were 0.02 and 0.48 percentage points higher than those of RENet, respectively. In the five-way one-shot and five-way five-shot experiments on the CIFAR-FS dataset, EMNet's classification accuracies were 0.19 and 0.18 percentage points higher than those of RENet.

MT‐BAAN: Multi‐View Topological Bilinear Aggregation Attention Network Model for Alzheimer's Disease Diagnosis

ABSTRACTAlzheimer's disease (AD) and mild cognitive impairment (MCI) are common cognitive disorders. Research has shown that cognitive decline is closely related to abnormal connections between different functional areas of the brain. However, research on brain functional network (BFN) has mainly focused on individual topological structures, seldom considering the sparsity of the BFNs and the complexity of multi‐level interactions among brain regions. To tackle this problem, in this article, we propose a multi‐view topological bilinear aggregation attention network model (MT‐BAAN) for disease diagnosis and brain network analysis. Based on rs‐fMRI data, the model mainly includes a multi‐view graph construction module (MVGC), a feature enhancement module (FEM), a dual‐level attention module (DLAM), and a graph relation convolution network module (GRCN). MVGC module uses two sparse methods to construct high‐view and low‐view graphs and retains fully connected BFN topology as the full‐view, aiming at capturing multi‐scale topological features. FEM and DLAM utilize bilinear aggregation and attention mechanisms, respectively, to learn topological features and obtain weight coefficients that reflect the importance of different network views. The GRCN module employs two convolutional operators to learn the BFN topology information at the node and network levels and completes the classification. The experimental results indicate that the complementary learning of multi‐view topologies can effectively improve model performance. Across binary classification tasks and ternary classification tasks, MT‐BAAN shows superior performance compared to other experimental methods, which is valuable for research and clinical diagnosis of attention deficit disorder AD and MCI.

Multi-stream and multi-scale fusion rib fracture segmentation network based on UXNet.

Accurate segmentation of rib fractures represents a pivotal procedure within surgical interventions. This meticulous process not only mitigates the likelihood of postoperative complications but also facilitates expedited patient recuperation. However, rib fractures in computed tomography (CT) images exhibit an uneven morphology and are not fixed in position, posing difficulties in segmenting fractures. This study aims to enhance the accuracy of elongated rib fracture segmentation, ultimately improving the efficiency of clinical diagnosis. In this study, we propose multi-stream and multi-scale fusion network based on efficient attention UXNet (M2SUXNet). It aims to enhance the segmentation accuracy of elongated rib fractures through multi-scale fusion attention enhancement. Firstly, we propose the multi-stream and multi-scale fusion (M2SF) module in the feature extraction stage. The module is designed with two parallel paths. Each path analyzes the image content using a different feature level. Then, the module effectively distinguishes the more critical feature information in the channel according to the feature weight ratio. The M2SF module integrates information from different scales to obtain comprehensive information on global and local features, achieving a more diverse feature representation. Secondly, the efficient attention (EA) module combines different channel information of input features to integrate channel and spatial features of different channels. The module better combines the context information, establishes the dependency between the space and the channel, enhances the focusing ability of the network on the fractures of different shapes, and improves the segmentation accuracy. Thirdly, the joint loss function of BCE with Logits Loss and Dice Loss is used to solve the sample imbalance problem. We verified the effectiveness of the proposed model on the public RibFrac dataset. The experimental results demonstrated that the model achieved a Dice coefficient of 75.34%, a joint intersection over union (IoU) of 60.44%, and a precision of 93.79%. The proposed model for rib fracture segmentation has higher accuracy and feasibility than other existing models. Besides, the M2SUXNet can effectively improve the segmentation performance of elongated rib fractures.

MacNet: a mobile attention classification network combining convolutional neural network and transformer for the differentiation of cervical cancer.

Cervical cancer remains a critical global health issue, responsible for over 600,000 new cases and 300,000 deaths annually. Pathological imaging of cervical cancer is a crucial diagnostic tool. However, distinguishing specific areas of cellular differentiation remains challenging because of the lack of clear boundaries between cells at various stages of differentiation. To address the limitations of conventional clinical and deep learning (DL) methods, we developed a mobile attention classification network (MacNet) with multiscale features, aiming to increase the accuracy of differentiation classification and quantitatively analyze cervical cancer cell differentiation. We investigated the application of MacNet for classifying non-background images into 3 stages of cervical cancer differentiation. The feature maps are processed through the Mobile Convolution Neural Network with Mobile Attention (MCMA) module, which integrates mobile convolutional blocks and mobile attention blocks. MacNet harnesses the benefits of the image pyramid structure and self-attention mechanism, enabling multiscale feature extraction and emulation of clinical pathologist analysis. The final prediction is generated by the adaptive fusion module, which aggregates features into a unified output. Comparative evaluations demonstrated that MacNet outperforms existing models. The proposed method achieved the best classification accuracy of 92.34% among all 7 DL-based models. Specifically, the result achieved by MacNet was 2.62% greater than that of Inception Version 3, 7.9% greater than that of vision transformer, 8.08% greater than that of the visual geometry group network, 3.21% greater than that of Densely Connected Convolutional Network, 2.85% greater than that of shifted window transformer (Swin transformer), 5.4% greater than that of Cross Stage Partial DarkNet, and 5.41% greater than that of Residual Neural Network. At the same time, MacNet also achieved superior results in recall, precision, and F1 score. We have proposed a lightweight neural network method that innovatively combines attention mechanisms with convolutional neural networks (CNNs) to efficiently utilize multiscale information from histopathological images. This integration enables the precise quantitative display of different stages of differentiation in cervical cancer. By doing so, our method not only enhances diagnostic accuracy but also provides clinicians with a more effective tool for faster and more reliable diagnosis, representing a significant advancement in the field of pathological imaging.

MobileNetV2-Based deep learning architecture with progressive transfer learning for accurate monkeypox detection

The recent monkeypox disease outbreak and its serious implications emphasize the urgent necessity for advanced diagnostic tools capable of rapid and accurate disease detection. In response, this study introduces a novel architecture for monkeypox disease identification, leveraging the capabilities of MobileNetV2 combined with progressive transfer learning and innovative model enhancements. The research began by analyzing six pre-trained models, with MobileNetV2 emerging as the most promising due to its balance of efficiency and performance. Using a self-assembled dataset, the model initially achieved an accuracy of 93.79%. Through strategic data augmentation, this performance increased to 97.85%. Further enhancements incorporated a Residual Dilated Spatial Pyramid Integration (ResDSPI) block, boosting feature extraction capabilities and elevating accuracy to 98.83%. The integration of an Efficient Channel Attention (ECA) module propelled the accuracy to an exceptional 99.34%. Each development stage was meticulously designed to ensure systematic improvements in the model’s ability to diagnose monkeypox effectively. The proposed model was rigorously tested on two datasets: the Monkeypox Skin Images Dataset (MSID) and the Monkeypox Skin Lesion Dataset (MSLD). In the binary classification scenario of the MSLD (monkeypox vs. others), our model achieved superior performance, with an accuracy of 95.55%, precision of 95.45%, and sensitivity of 92%. For the multi-class classification task on the MSID, which includes four classes (Chickenpox, Measles, Monkeypox, and Normal), the model achieved an accuracy of 92.95%, precision of 93.09%, and sensitivity of 89.00%. The model’s superior sensitivity and specificity underscore its potential as a critical asset in the surveillance and control of monkeypox outbreaks. These significant advancements not only enhance predictive performance but also mark a major step forward in the field of monkeypox disease image analysis, setting the stage for future applications in combating infectious diseases.

UMS[formula omitted]-ODNet: Unified-scale domain adaptation mechanism driven object detection network with multi-scale attention

Unsupervised domain adaptation techniques improve the generalization capability and performance of detectors, especially when the source and target domains have different distributions. Compared with two-stage detectors, one-stage detectors (especially YOLO series) provide better real-time capabilities and become primary choices in industrial fields. In this paper, to improve cross-domain object detection performance, we propose a Unified-Scale Domain Adaptation Mechanism Driven Object Detection Network with Multi-Scale Attention (UMS2-ODNet). UMS2-ODNet chooses YOLOv6 as the basic framework in terms of its balance between efficiency and accuracy. UMS2-ODNet considers the adaptation consistency across different scale feature maps, which tends to be ignored by existing methods. A unified-scale domain adaptation mechanism is designed to fully utilize and unify the discriminative information from different scales. A multi-scale attention module is constructed to further improve the multi-scale representation ability of features. A novel loss function is created to maintain the consistency of multi-scale information by considering the homology of the descriptions from the same latent feature. Multiply experiments are conducted on four widely used datasets. Our proposed method outperforms other state-of-the-art techniques, illustrating the feasibility and effectiveness of the proposed UMS2-ODNet.

Attentional modulation of peripheral pointing hypometria in healthy participants: an insight into optic ataxia?

Damage to the superior parietal lobule and intraparietal sulcus (SPL-IPS) causes optic ataxia (OA), characterized by pathological gaze-centered hypometric pointing to targets in the affected peripheral visual field. The SPL-IPS is also involved in covert attention. Here, we investigated the possible link between attention and action. This study investigated the effect of attention on pointing performance in healthy participants and two OA patients. In invalid trials, targets appeared unpredictably across different visual fields and eccentricities. Valid trials involved cued targets at specific locations. The first experiment used a central cue with 75% validity, the second used a peripheral cue with 50% validity. The effect of attention on pointing variability (noise) or time was expected as a confirmation of cueing efficiency. Critically, if OA reflects an attentional deficit, then healthy participants, in the invalid condition (without attention), were expected to produce the gaze-centered hypometric pointing bias characteristic of OA. RESULTS: revealed main effects of validity on pointing biases in all participants with central predictive cueing, but not with peripheral low predictive cueing. This suggests that the typical underestimation of visual eccentricity in OA (visual field effect) at least partially results from impaired endogenous attention orientation toward the affected visual field.

V2X-ViTv2: Improved Vision Transformers for Vehicle-to-Everything Cooperative Perception.

In this paper, we study the application of Vehicle-to-Everything (V2X) communication to improve the perception performance of autonomous vehicles. We present V2X-ViTs, a robust cooperative perception framework with V2X communication using novel vision Transformer models. First, we present V2X-ViTv1 containing holistic attention modules that can effectively fuse information across on-road agents (i.e., vehicles and infrastructure). Specifically, V2X-ViTv1 consists of alternating layers of heterogeneous multi-agent self-attention and multi-scale window self-attention, which captures inter-agent interaction and per-agent spatial relationships. These key modules are designed in a unified Transformer architecture to handle common V2X challenges, including asynchronous information sharing, pose errors, and heterogeneity of V2X components. Second, we propose an advanced architecture, V2X-ViTv2, that enjoys increased ability for multi-scale perception. We also propose advanced data augmentation techniques tailored for V2X applications to improve performance. We construct a large-scale V2X perception dataset using CARLA and OpenCDA to validate our approach. Extensive experimental results on both synthetic and real-world datasets show that V2X-ViTs achieve state-of-the-art performance for 3D object detection and are robust even under harsh, noisy environments. All the code and trained models will be available at https://github.com/DerrickXuNu/OpenCOOD.

Automated Chest X-Ray Diagnosis Report Generation with Cross-Attention Mechanism

In the medical field, it is extremely important to use deep learning technology to automatically generate diagnostic reports for chest X-ray images. This technology provides an effective solution to the challenges faced by the medical field in processing large numbers of chest X-ray images. Especially during large-scale outbreaks of epidemics such as the new COVID-19, rapid and accurate screening and diagnosis of cases become important tasks. This study uses deep learning technology to automatically generate diagnostic reports for chest X-ray images, which significantly reduces the workload of doctors, reduces the risk of misdiagnosis and missed diagnosis, and provides technical support for improving public health emergency response capabilities. In this study, we propose an innovative network architecture to address the limitations of traditional image description networks in generating chest X-ray diagnostic reports, especially the large area deviation between abnormal and normal areas, and the lack of effective alignment of the two modalities of image and text. The convolutional block attention module (CBAM) is adopted to effectively alleviate the data bias problem through a sophisticated feature attention mechanism and improve the model’s ability to recognize abnormal image areas. The cross-attention mechanism is adopted to optimize the alignment process between images and texts, ensuring the accuracy and reliability of the diagnosis report.

A Pre-Activation Residual Convolutional Network With Attention Modules for High-Resolution Segmented EEG Emotion Recognition

A Pre-Activation Residual Convolutional Network With Attention Modules for High-Resolution Segmented EEG Emotion Recognition

EMSPLA for accurate feature molecular extraction from protein-ligand interactions

Protein-ligand interactions are fundamental in various biological and medical fields, influencing drug discovery and therapeutic development. In recent years, deep learning (DL) has revolutionized the study of these interactions, but significant challenges remain in accurately representing molecular structures for DL models. Traditional featurization techniques often depend on handcrafted features, requiring expert knowledge and potentially missing crucial molecular aspects. This work addresses these challenges by developing and evaluating a novel protein-ligand feature extraction system using an enhanced molecular similarity protein-ligand aligner (EMSPLA). The primary objective is to leverage EMSPLA for similarity matching in protein-ligand interactions, improving predictive model accuracy. The methodology combines convolutional neural networks (CNN) for local feature extraction with an attention module to capture long-distance dependencies, enhancing binding site predictions. Using the PDBbind v.2020 dataset, the EMSPLA model demonstrated superior performance with a root mean square error (RMSE) of 0.67, surpassing current state-of-the-art models. These findings highlight the system’s potential for efficient deployment and scalability, positioning it as a powerful tool in computational biology and drug discovery, ultimately advancing our understanding of protein-ligand interactions.

SAASNets: Shared attention aggregation Siamese networks for building change detection in multispectral remote sensing.

Interfered by external factors, the receptive field limits the traditional CNN multispectral remote sensing building change detection method. It is difficult to obtain detailed building changes entirely, and redundant information is reused in the encoding stage, which reduces the feature representation and detection performance. To address these limitations, we design a Siamese network of shared attention aggregation to learn the detailed semantics of buildings in multispectral remote sensing images. On the one hand, a special attention embedding module is introduced into each subspace of the feature extractor to promote the interaction between multi-scale local features and enhance the representation of global features. On the other hand, a highly efficient channel and position multi-head attention module is added to the Siamese features to encode position details while sharing channel information. In addition, adopting a feature aggregation module with a residual strategy to fuse the features of different stages of the Siamese network is beneficial for detecting different scales and irregular object buildings. Finally, experimental results on LEVIR-CD and CDD datasets show that designed SAASNets have better accuracy and robustness.