An Improved COVID-19 Lung X-Ray Image Classification Algorithm Based on ConvNeXt Network

Aiming at the new coronavirus that appeared in 2019, which has caused a large number of infected patients worldwide due to its high contagiousness, in order to detect the source of infection in time and cut off the chain of transmission, we developed a new Chest X-ray (CXR) image classification algorithm with high accuracy, simple operation and fast processing for COVID-19. The algorithm is based on ConvNeXt pure convolutional neural network, we adjusted the network structure and loss function, added some new Data Augmentation methods and introduced attention mechanism. Compared with other classical convolutional neural network classification algorithms such as AlexNet, ResNet-34, ResNet-50, ResNet-101, ConvNeXt-tiny, ConvNeXt-small and ConvNeXt-base, the improved algorithm has better performance on COVID dataset.

Sustainable strategy for online physical education teaching using ResNet34 and big data.

Since the global COVID-19 outbreak in the spring of 2020, online instruction has replaced traditional classroom instruction as the main method of educating students. Teaching physical education online can be challenging, as it may be difficult to teach students certain movements, accurate student mobility, and appropriate exercise assignments. This paper proposed an online teaching support system with sustainable development features that utilize several large data sets. The system is based on the deep learning image recognition algorithm ResNet34, which can analyze and correct student actions in real-time for gymnastics, dance, basketball, and other sports. By combining the attention mechanism module with the original ResNet34, the detection precision of the system can be enhanced. The sustainability of the system is evident from the fact that the data set can be expanded in response to the emergence of new sports categories and can be kept current in real-time. According to experiments, the target identification accuracy of the proposed system, which combines ResNet34 and the attention mechanism, is higher than that of several other methods currently in use. The proposed techniques outperform the original ResNet34 in terms of accuracy, precision, and recall by 4.1%, 2.8%, and 3.6%, respectively. The suggested approach significantly improves student action correction in virtual sports instruction.

A hybrid model for daily air quality index prediction and its performance in the face of impact effect of COVID-19 lockdown.

Accurate and dependable air quality forecasting is critical to environmental and human health. However, most methods usually aim to improve overall prediction accuracy but neglect the accuracy for unexpected incidents. In this study, a hybrid model was developed for air quality index (AQI) forecasting, and its performance during COVID-19 lockdown was analyzed. Specifically, the variational mode decomposition (VMD) was employed to decompose the original AQI sequence into some subsequences with the parameters optimized by the Whale optimization algorithm (WOA), and the residual sequence was further decomposed by the complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN). On this basis, a deep learning method bidirectional long short-term memory coupled with added time filter layer and attention mechanism (TFA-BiLSTM) was employed to explore the latent dynamic characteristics of each subsequence. This WOA-VMD-CEEMDAN-TFA-BiLSTM hybrid model was used to forecast AQI values for four cities in China, and results verified that the accuracy of the hybrid model outperformed other proposed models, achieving R2 values of 0.96-0.97. In addition, the improvement in MAE (34.71-49.65%) and RMSE (32.82-48.07%) were observed over single decomposition-based model. Notably, during the epidemic lockdown period, the hybrid model had significant superiority over other proposed models for AQI prediction.

ATFormer: Advanced transformer for medical image segmentation

Combining transformers and convolutional neural networks is considered one of the most important directions for tackling medical image segmentation problems. To learn the long-range dependencies and local contexts, previous approaches embedded a convolutional layer into feedforward neural network inside the transformer block. However, a common issue is the instability during training since large differences in amplitude across layers by pre-layer normalization. Furthermore, multi-scale features were directly fused using the transformer from the encoder to decoder, which could disrupt valuable information for segmentation. To address these concerns, we propose Advanced TransFormer (ATFormer), a novel hybrid architecture that combines convolutional neural networks and transformers for medical image segmentation. First, the traditional transformer block has been refined into an Advanced Transformer Block, which adopts post-layer normalization to obtain mild activation values and employs the scaled cosine attention with shifted window for accurate spatial information. Second, the Progressive Guided Fusion module is introduced to make multi-scale features more discriminative while reducing the computational complexity. Experimental results on the ACDC, COVID-19 CT-Seg, and Tumor datasets demonstrate the significant advantage of ATFormer over existing methods that rely solely on convolutional neural networks, transformers, or their combination.

Multi-scale sequential feature selection for disease classification using Raman spectroscopy data.

Raman spectroscopy (RS) optical technology promises non-destructive and fast application in medical disease diagnosis in a single step. However, achieving clinically relevant performance levels remains challenging due to the inability to search for significant Raman signals at different scales. Here we propose a multi-scale sequential feature selection method that can capture global sequential features and local peak features for disease classification using RS data. Specifically, we utilize the Long short-term memory network (LSTM) module to extract global sequential features in the Raman spectra, as it can capture long-term dependencies present in the Raman spectral sequences. Meanwhile, the attention mechanism is employed to select local peak features that were ignored before and are the key to distinguishing different diseases. Experimental results on three public and in-house datasets demonstrate the superiority of our model compared with state-of-the-art methods for RS classification. In particular, our model achieves an accuracy of 97.9 ± 0.2% on the COVID-19 dataset, 76.3 ± 0.4% on the H-IV dataset, and 96.8 ± 1.9% on the H-V dataset.

Face mask detection based on improved YOLOv3 algorithm

In the current situation of COVID-19 prevention and control, wearing masks remains an important way to prevent the transmission of the Novel Coronavirus. Aiming at the problem that the detection accuracy of the traditional YOLOv3 algorithm can still be improved, this paper proposes an improved yolov3 algorithm and applies it to the practical problem of detecting whether to wear a mask. Firstly, the algorithm introduces the residual structure of structural reparameterization in the feature extraction network named Darknet53 of YOLOv3 to obtain the input features;Secondly, the SimSPPF (Simplified Spatial Pyramid Pooling-Fast) is introduced to enhance feature extraction;Finally, an improved attention mechanism is introduced to make the model focus on regions with more prominent features. Besides, in order to ensure the accuracy of target detection, CIoU and Focal loss function was used in the training process. The results show that compared with the traditional YOLOv3, the detection accuracy of the improved algorithm for normal face and mask face is improved by 16.98% and 7.30% respectively, and the mAP is improved by 12.14%, which can meet the requirements of daily use and lay a foundation for rapid face recognition when wearing mask. () © 2023 IEEE.

Mask Wear Detection Algorithm Based on YOLOv5m

Wearing a mask during the COVID-19 epidemic can effectively prevent the spread of the virus. In view of the problems of small target size, crowd blocking each other and dense arrangement of targets in crowded places, a target detection algorithm based on the improved YOLOv5m model is proposed to achieve efficient detection of whether a mask is worn or not. This paper introduces four attention mechanisms in the feature extraction network based on the YOLOv5m model to suppress irrelevant information, enhance the information representation of the feature map, and improve the detection capability of the model for small-scale targets. The experimental results showed that the introduction of the SE module increased the mAP value of the original network by 9.3 percentage points, the most significant increase among the four attention mechanisms. And then a dual-scale feature fusion network is used in the Neck layer, giving different weights to the feature layers to convey more effective feature information. In the image pre-processing, the Mosaic method was used for data enhancement, and the CIoU loss function was used for coordinate frame positioning in the prediction layer. Experiments on the improved YOLOv5m algorithm demonstrate that the mean recognition accuracy of the method improves by 10.7 percentage points over the original method while maintaining the original model size and detection speed, and better solves the problems of small scale, dense arrangement and mutual occlusion of targets in mask wearing detection tasks in crowded places. © 2023 IEEE.

Attention-augmented U-Net (AA-U-Net) for semantic segmentation.

Deep learning-based image segmentation models rely strongly on capturing sufficient spatial context without requiring complex models that are hard to train with limited labeled data. For COVID-19 infection segmentation on CT images, training data are currently scarce. Attention models, in particular the most recent self-attention methods, have shown to help gather contextual information within deep networks and benefit semantic segmentation tasks. The recent attention-augmented convolution model aims to capture long range interactions by concatenating self-attention and convolution feature maps. This work proposes a novel attention-augmented convolution U-Net (AA-U-Net) that enables a more accurate spatial aggregation of contextual information by integrating attention-augmented convolution in the bottleneck of an encoder-decoder segmentation architecture. A deep segmentation network (U-Net) with this attention mechanism significantly improves the performance of semantic segmentation tasks on challenging COVID-19 lesion segmentation. The validation experiments show that the performance gain of the attention-augmented U-Net comes from their ability to capture dynamic and precise (wider) attention context. The AA-U-Net achieves Dice scores of 72.3% and 61.4% for ground-glass opacity and consolidation lesions for COVID-19 segmentation and improves the accuracy by 4.2% points against a baseline U-Net and 3.09% points compared to a baseline U-Net with matched parameters. Supplementary Information: The online version contains supplementary material available at 10.1007/s11760-022-02302-3.

Prediction of atmospheric pollutants in urban environment based on coupled deep learning model and sensitivity analysis.

Accurate and efficient predictions of pollutants in the atmosphere provide a reliable basis for the scientific management of atmospheric pollution. This study develops a model that combines an attention mechanism, convolutional neural network (CNN), and long short-term memory (LSTM) unit to predict the O3 and PM2.5 levels in the atmosphere, as well as an air quality index (AQI). The prediction results given by the proposed model are compared with those from CNN-LSTM and LSTM models as well as random forest and support vector regression models. The proposed model achieves a correlation coefficient between the predicted and observed values of more than 0.90, outperforming the other four models. The model errors are also consistently lower when using the proposed approach. Sobol-based sensitivity analysis is applied to identify the variables that make the greatest contribution to the model prediction results. Taking the COVID-19 outbreak as the time boundary, we find some homology in the interactions among the pollutants and meteorological factors in the atmosphere during different periods. Solar irradiance is the most important factor for O3, CO is the most important factor for PM2.5, and particulate matter has the most significant effect on AQI. The key influencing factors are the same over the whole phase and before the COVID-19 outbreak, indicating that the impact of COVID-19 restrictions on AQI gradually stabilized. Removing variables that contribute the least to the prediction results without affecting the model prediction performance improves the modeling efficiency and reduces the computational costs.

NAG-Net: Nested attention-guided learning for segmentation of carotid lumen-intima interface and media-adventitia interface.

Cardiovascular diseases (CVD), as the leading cause of death in the world, poses a serious threat to human health. The segmentation of carotid Lumen-intima interface (LII) and Media-adventitia interface (MAI) is a prerequisite for measuring intima-media thickness (IMT), which is of great significance for early screening and prevention of CVD. Despite recent advances, existing methods still fail to incorporate task-related clinical domain knowledge and require complex post-processing steps to obtain fine contours of LII and MAI. In this paper, a nested attention-guided deep learning model (named NAG-Net) is proposed for accurate segmentation of LII and MAI. The NAG-Net consists of two nested sub-networks, the Intima-Media Region Segmentation Network (IMRSN) and the LII and MAI Segmentation Network (LII-MAISN). It innovatively incorporates task-related clinical domain knowledge through the visual attention map generated by IMRSN, enabling LII-MAISN to focus more on the clinician's visual focus region under the same task during segmentation. Moreover, the segmentation results can directly obtain fine contours of LII and MAI through simple refinement without complicated post-processing steps. To further improve the feature extraction ability of the model and reduce the impact of data scarcity, the strategy of transfer learning is also adopted to apply the pretrained weights of VGG-16. In addition, a channel attention-based encoder feature fusion block (EFFB-ATT) is specially designed to achieve efficient representation of useful features extracted by two parallel encoders in LII-MAISN. Extensive experimental results have demonstrated that our proposed NAG-Net outperformed other state-of-the-art methods and achieved the highest performance on all evaluation metrics.

Detection of Pneumonia Based on ResNet Improved by Attention Mechanism

In recent years, Covid-19 is one of the major health challenges facing the human population. Due to the highly infectious nature of Covid-19 and the difficulty of detecting symptoms in the early stages, it is definitely necessary to combine X-ray for the diagnosis of pneumonia. Using traditional neural networks such as VGG, ResNet, and DenseNet to diagnose pneumonia based on X-ray images faces a number of difficulties. These models have insufficient spatial information extraction capability and are prone to overfitting on the training set. The attention mechanism is a means to improve model performance by helping the model better extract channel and spatial features from the feature maps. To identify pneumonia more accurately, we combined the ResNet network and CBAM attention mechanism to design the ResNet101-cbam model with a series of data augmentation methods as well as training strategies. We used the same approach to add attention mechanisms to ResNet50, ResNet101 and ResNet152 and tested their performance. The results show that ResNet101-cbam is the best performing model overall. It achieved a recall of 0.8205, a precision of 0.822, and an accuracy of 0.8285 on the test set, while the original pretrained ResNet101 had a precision of 0.7280 and an accuracy of 0.7644. Its performance were better than the more complex model: ResNet152-cbam, a little bit, but the training speed is improved by more than 25%. More importantly, the model with the added attention mechanism effectively overcomes the effects of positive and negative sample imbalance. The ResNet101-cbam model can be used as a medical aid, which can improve diagnostic efficiency and help us better deal with large-scale pneumonia epidemics. © 2023 IEEE.

Fusing hybrid attentive network with self-supervised dual-channel heterogeneous graph for knowledge tracing

Recently the large-scale influence of Covid-19 promoted the fast development of intelligent tutoring systems (ITS). As a major task of ITS, Knowledge Tracing (KT) aims to capture a student's dynamic knowledge state based on his historical response sequences and provide personalized learning assistance to him. However, most existing KT methods have encountered the data sparsity problem. In real scenarios, an online tutoring system usually has an extensive collection of questions while each student can only interact with a limited number of questions. As a result, the records of some questions could be extremely sparse, which degrades the performance of traditional KT models. To resolve this issue, we propose a Dual-channel Heterogeneous Graph Network (DHGN) to learn informative representations of questions from students' records by capturing both the high-order heterogeneous and local relations. As the supervised learning manner applied in previous methods is incapable of exploiting unobserved relations between questions, we innovatively integrate a self-supervised framework into the KT task and employ contrastive learning via the two channels of DHGN, supplementing as an auxiliary task to improve the KT performance. Moreover, we adopt the attention mechanism, which has achieved impressive performance in natural language processing tasks, to effectively capture students' knowledge state. But the standard attention network is inapplicable to the KT task because the current knowledge state of a student usually shows strong dependency on his recently interactive questions, unlike the situation of language processing tasks, which focus more on the long-term dependency. To avoid the inefficiency of standard attention networks in the KT task, we further devise a novel Hybrid Attentive Network (HAN), which produces both the global attention and the hierarchical local attention to model the long-term and short-term intents, respectively. Then, by the gating network, a student's long-term and short-term intents are combined for efficient prediction. We conduct extensive experiments on several real-world datasets. Experimental results demonstrate that our proposed methods achieve significant performance improvement compared to existing state-of-the-art baselines, which validates the effectiveness of the proposed dual-channel heterogeneous graph framework and hybrid attentive network. © 2023 Elsevier Ltd

Research and Optimization of a Lightweight Refined Mask-Wearing Detection Algorithm Based on an Attention Mechanism

To address the current problems of the incomplete classification of mask-wearing detection data, small-target miss detection, and the insufficient feature extraction capabilities of lightweight networks dealing with complex faces, a lightweight method with an attention mechanism for detecting mask wearing is presented in this paper. This study incorporated an "incorrect_mask” category into the dataset to address incomplete classification. Additionally, the YOLOv4-tiny model was enhanced with a prediction feature layer and feature fusion execution, expanding the detection scale range and improving the performance on small targets. A CBAM attention module was then introduced into the feature enhancement network, which re-screened the feature information of the region of interest to retain important feature information and improve the feature extraction capabilities. Finally, a focal loss function and an improved mosaic data enhancement strategy were used to enhance the target classification performance. The experimental results of classifying three objects demonstrate that the lightweight model's detection speed was not compromised while achieving a 2.08% increase in the average classification precision, which was only 0.69% lower than that of the YOLOv4 network. Therefore, this approach effectively improves the detection effect of the lightweight network for mask-wearing.

A New Pneumonia Detection Model Based on Transformer with Improved Self-Attention Mechanism

Clinical diagnosis based on computed tomography (CT) could be used, as part of diagnosis standard of COVID-19 pneumonia. Addressing the problem that accuracy of CT-based traditional pneumonia classification diagnosis models is relatively low when employed for classification of community-acquired pneumonia (CP), COVID-19 pneumonia (NCP) and normal cases, a new network model is proposed which combines application of Swin Transformer and multi-head axial self-attention (MASA) mechanism, to analyze CT images and make intelligence-assisted diagnosis. The method in detail is to partially replace traditional multi-head self-attention (MSA) mechanism in encoders of Swin Transformer by MASA. The improved model is applied to train and test on commonly used pneumonia CT dataset CC-CCII. The results show that the proposed network outperforms traditional networks ResNet50 and Vision Transformer in indicators of accuracy, sensitivity and F1-measure. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

A Predictive Model of COVID-19 mRNA Vaccine Reactivity based on Dual Attention for Federated Learning Scenarios

The base reactivity of the mRNA sequence has a significant impact on the effectiveness of the mRNA vaccine in fighting against the pandemic of COVID-19. The annotation of mRNA sequence reactivity value is a time-consuming and labor-intensive work, which belongs to the private digital assets of each medical institution. It is not practical to train a predictive model by pooling private data from various parties. Fortunately, federated learning techniques can serve to collaboratively train a predictive model among medical institutions while preserving respective digital assets. However, due to the scarcity of data from each participant, conventional sequential prediction mod-els often fail to perform well. To overcome such a challenge, we propose a reactivity value prediction model based on both the self-attention and the convolutional attention mechanisms only requiring a small dataset of labeled samples. Inspired by BERT, we first train a self-attention feature extraction model through self-supervision using both labeled and unlabeled mRNA samples. In this way, the information of mRNA in the semantic space is deeply mined. Then, a convolutional attention block follows the self-attention block, to extract the attention matrix from the base-pair probability matrix and adjacency matrix. By doing so, the attention matrix can compensate for the insensitivity of the self-attention mechanism to the spatial information of mRNA. By using the Open Vaccine RNA database, experiments show that our prediction model for unseen mRNA has a better performance than other state-of-the-art deep learning models that are used to process gene sequences. Further ablation experiments demonstrate that the existence of the dual attention mechanism reduces the risk of overfitting, resulting in an excellent generalization capability of our model. © 2022 IEEE.

Specific mining of COVID-19 reviews using deep learning methods

Sentiment classification is a valid measure to monitor public opinion on the COVID-19 epidemic. This study provides a significant basis for preventing the spread of adverse public opinion. Firstly, in epidemic texts, we use a convolutional neural network and bidirectional long short-term memory neural network BiLSTM model to classify and analyze the sentiment of the comment texts about the epidemic situation on Weibo. Secondly, embedded in the model layer to generate adversarial samples and extract semantics. Then, semantic information is weighted using the attention mechanism. Finally, the RMS optimizer is used to update the neural network weights iteratively. According to comparative experiments, the experimental results show that such four evaluation metrics as accuracy, precision, recall, and f1-score with our proposed model have obtained better classification performance. © 2022 IEEE.

Deep-Learning-Based COVID-19 Diagnosis and Implementation in Embedded Edge-Computing Device.

The rapid spread of coronavirus disease 2019 (COVID-19) has posed enormous challenges to the global public health system. To deal with the COVID-19 pandemic crisis, the more accurate and convenient diagnosis of patients needs to be developed. This paper proposes a deep-learning-based COVID-19 detection method and evaluates its performance on embedded edge-computing devices. By adding an attention module and mixed loss into the original VGG19 model, the method can effectively reduce the parameters of the model and increase the classification accuracy. The improved model was first trained and tested on the PC X86 GPU platform using a large dataset (COVIDx CT-2A) and a medium dataset (integrated CT scan); the weight parameters of the model were reduced by around six times compared to the original model, but it still approximately achieved 98.80%and 97.84% accuracy, outperforming most existing methods. The trained model was subsequently transferred to embedded NVIDIA Jetson devices (TX2, Nano), where it achieved 97% accuracy at a 0.6-1 FPS inference speed using the NVIDIA TensorRT engine. The experimental results demonstrate that the proposed method is practicable and convenient; it can be used on a low-cost medical edge-computing terminal. The source code is available on GitHub for researchers.

Progressive attention integration-based multi-scale efficient network for medical imaging analysis with application to COVID-19 diagnosis.

In this paper, a novel deep learning-based medical imaging analysis framework is developed, which aims to deal with the insufficient feature learning caused by the imperfect property of imaging data. Named as multi-scale efficient network (MEN), the proposed method integrates different attention mechanisms to realize sufficient extraction of both detailed features and semantic information in a progressive learning manner. In particular, a fused-attention block is designed to extract fine-grained details from the input, where the squeeze-excitation (SE) attention mechanism is applied to make the model focus on potential lesion areas. A multi-scale low information loss (MSLIL)-attention block is proposed to compensate for potential global information loss and enhance the semantic correlations among features, where the efficient channel attention (ECA) mechanism is adopted. The proposed MEN is comprehensively evaluated on two COVID-19 diagnostic tasks, and the results show that as compared with some other advanced deep learning models, the proposed method is competitive in accurate COVID-19 recognition, which yields the best accuracy of 98.68% and 98.85%, respectively, and exhibits satisfactory generalization ability as well.

Mask detection algorithm based on YOLOv5 integrating attention mechanism

Wearing masks correctly during the COVID-19 pandemic can effectively prevent the spread of the virus. In response to the detection challenge posed by dense crowds and small detection targets in public places, a mask wearing detection algorithm based on the YOLOv5s model and integrating an attention mechanism was proposed. Four attention mechanisms were introduced into the backbone network of the YOLOv5s model to respectively suppress irrelevant information, enhance the ability of the feature map to express information, and improve the modelʹs detection ability for small-scale targets. Experimental results show that the introduction of the convolutional block attention module could increase the mAP value by 6.9 percentage points compared with the original network, with the greatest improvement among the four attention mechanisms. The normalization-based attention module also showed excellent performance, with the least quantity of parameters while losing a small amount of mAP. Through comparative experiments, the GIoU loss function was selected to calculate the bounding box regression loss, resulting in further improvements to positioning accuracy, resulting in an mAP value that was improved by 8.5 percentage points compared to the original network. The detection results of the improved model in different scenarios prove the accuracy and practicability of the algorithm for small target detection. © 2023, Editorial of Board of Journal of Graphics. All rights reserved.

An Effective Segmentation Method for COVID-19 CT Image Based on Attention Mechanism and Inf-Net

The corona virus disease 2019 (COVID-19) pandemic has recently ravaged the world, seriously affecting the life and health of human society. Computerized tomography (CT) imaging technology is an important diagnostic method for detecting COVID-19. Automatic and accurate segmentation of the lesion is of great significance for diagnosis, treatment and prognosis. Aiming at the segmentation of new coronary pneumonia lesions, an improved automatic segmentation method based on the Inf-Net algorithm is proposed, which introduces the channel attention module to improve feature representation and attention gate model to better fuse edge information. The experimental results on COVID-19 CT Segmentation dataset show that the Dice similarity coefficient, Sensitivity and Specificity of the proposed method are 75.1%, 75.4% and 95.4%. The segmentation performance of it is superior to that of other state-of-the-art ones. (English) [ABSTRACT FROM AUTHOR] 新型冠状病毒肺炎肆虐全球, 严重影响了人类社会的生活和健康。CT影像技术是检测新冠肺炎的重要诊断方式, 从CT图像中自动准确分割出新冠肺炎病灶区域, 对于诊断、治疗和预后都有重要意义。针对新冠肺炎病灶的自动分割, 文中提出基于Inf-Net算法改进的自动分割方法, 通过引入通道注意力机制加强特征表示, 并运用注意力门模块来更好地融合边缘信息。在COVID-19 CT分割数据集上的实验结果表明, 文中所提出新冠肺炎图像分割方法的Dice系数、灵敏度、特异率分别为75.1%、75.4%和95.4%, 算法性能也优于部分主流方法 (Chinese) [ABSTRACT FROM AUTHOR] Copyright of Electronic Science & Technology is the property of Electronic Science & Technology Editorial Office and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)