CARes‐UNet: Content‐aware residual UNet for lesion segmentation of COVID‐19 from chest CT images

Abstract

Purpose: Coronavirus disease 2019 (COVID-19) has caused a serious global health crisis. It has been proven that the deep learning method has great potential to assist doctors in diagnosing COVID-19 by automatically segmenting the lesions in computed tomography (CT) slices. However, there are still several challenges restricting the application of these methods, including high variation in lesion characteristics and low contrast between lesion areas and healthy tissues. Moreover, the lack of high-quality labeled samples and large number of patients lead to the urgency to develop a high accuracy model, which performs well not only under supervision but also with semi-supervised methods. Methods: We propose a content-aware lung infection segmentation deep residual network (content-aware residual UNet (CARes-UNet)) to segment the lesion areas of COVID-19 from the chest CT slices. In our CARes-UNet, the residual connection was used in the convolutional block,which alleviated the degradation problem during the training. Then, the content-aware upsampling modules were introduced to improve the performance of the model while reducing the computation cost. Moreover, to achieve faster convergence, an advanced optimizer named Ranger was utilized to update the model's parameters during training. Finally, we employed a semi-supervised segmentation framework to deal with the problem of lacking pixel-level labeled data. Results: We evaluated our approach using three public datasets with multiple metrics and compared its performance to several models. Our method outperforms other models in multiple indicators, for instance in terms of Dice coefficient on COVID-SemiSeg Dataset, CARes-UNet got the score 0.731, and semi-CARes-UNet further boosted it to 0.776. More ablation studies were done and validated the effectiveness of each key component of our proposed model. Conclusions: Compared with the existing neural network methods applied to the COVID-19 lesion segmentation tasks, our CARes-UNet can gain more accurate segmentation results, and semi-CARes-UNet can further improve it using semi-supervised learning methods while presenting a possible way to solve the problem of lack of high-quality annotated samples. Our CARes-UNet and semi-CARes-UNet can be used in artificial intelligence-empowered computer-aided diagnosis system to improve diagnostic accuracy in this ongoing COVID-19 pandemic. K E Y W O R D S computed tomography (CT) image, content-aware residual UNet, coronavirus disease 2019 (COVID-19), deep learning, segmentation

COVID‐19 lung infection segmentation from chest CT images based on CAPA‐ResUNet

“…To evaluate the effect of our model, some comparative experiments were carried out. Concretely, CAPA‐ResUNet was compared with various advanced deep learning models, including UNet, 9 SegNet, 11 CopleNet, 24 DDANet, 25 and CARes‐UNet 27 . All models were trained from the open‐source codes in the article.…”

“…Concretely, CAPA‐ResUNet was compared with various advanced deep learning models, including UNet, 9 SegNet, 11 CopleNet, 24 DDANet, 25 and CARes‐UNet. 27 All models were trained from the open‐source codes in the article. In our experiments, all networks shared the same dataset to ensure fairness.…”

“…Stefano et al 26 proposed a customized ENET for real‐time and automatic recognition and segmentation of computer tomography images infected with COVID‐19 disease and achieved good results. Xu et al 27 proposed a content‐aware residual network (CARes‐UNet) to segment COVID‐19 lesions from chest CT images. In the network, the content‐aware upsampling module was introduced to enhance model performance.…”

Interactive framework for Covid-19 detection and segmentation with feedback facility for dynamically improved accuracy and trust

“…The proposed TV-Unet model was compared with U-Net+, Inf-Net, Semi-Inf-Net on Covid chest CT images; it outperformed them with 0.801 dice score. Another variation of U-Net named ‘CARes‐UNet (content-aware residual UNet)’ was proposed in [ 41 ] for lesion segmentation from Covid chest CT images. The residual network was used for improving the segmentation performance.…”

Research on Intelligent Fault Diagnosis of Rolling Bearing Based on Adaptive Resource Allocation Deep Neural Network

“…The LR optimizer can dynamically turn on or off the adaptive learning rate according to the variance dispersion of the current training data set, ensuring the stability of the training stage Lookahead-Radam (LR) optimizer has been successfully applied in the research of deep learning. Pan Zhang et al [20] applied LR optimizer to efficientNet and established a new method for identifying cucumber diseases in greenhouse under natural complex environment, and obtained 96% accuracy in the classification of similar cucumber diseases; Xinhua Xu et al [21]applied LR optimizer to the neural network for segmenting covid-19 lesions from chest CT images, so as to help doctors diagnose covid-19 virus. This…”