The coronavirus disease 2019 (COVID-19) epidemic has spread worldwide and the healthcare system is in crisis. Accurate, automated and rapid segmentation of COVID-19 lesion in computed tomography (CT) images can help doctors diagnose and provide prognostic information. However, the variety of lesions and small regions of early lesion complicate their segmentation.To solve these problems, we propose a new SAUNet++ model with squeeze excitation residual (SER) module and atrous spatial pyramid pooling (ASPP) module. The SER module can assign more weights to more important channels and mitigate the problem of gradient disappearance; the ASPP module can obtain context information by atrous convolution using various sampling rates. In addition, the generalized dice loss (GDL) can reduce the correlation between lesion size and dice loss, and is introduced to solve the problem of small regions segmentation of COVID-19 lesion. We collected multinational CT scan data from China, Italy and Russia and conducted extensive comparative and ablation studies. The experimental results demonstrated that our method outperforms state-of-the-art models and can effectively improve the accuracy of COVID-19 lesion segmentation on the dice similarity coefficient (our: 87.38% vs. U-Net++: 84.25%), sensitivity (our: 93.28% vs. U-Net++: 89.85%) and Hausdorff distance (our: 19.99 mm vs. U-Net++: 26.79 mm), respectively.
KeywordsCoronavirus disease 2019 (COVID-19) • Image segmentation • Computed tomography (CT) • Squeeze excitation residual (SER) • Atrous spatial pyramid pooling (ASPP) • Generalized dice loss (GDL) B Zhiqiang Ran
The corresponding arrhythmia often occurs before the onset of cardiovascular disease (CVD), electrocardiogram (ECG) can more intuitively detect any abnormality in the heartbeat as a sign of arrhythmia. There are many traditional ECG classification methods, but these methods are constrained human costs and inaccuracy since they rely on manually extracting features, and cannot fully mine the deep pathological information hidden in the data. Consequently, a novel deep learning model based on single-lead ECG signals and inter-patient paradigm is developed to improve the shortcomings of traditional ECG heartbeat classification. The residual connection is adopted in the proposed method to improve classification accuracy and alleviate the gradient disappearance issue. Heartbeat complexes that included a targeted heartbeat and an adjacent heartbeat is selected as the input of the model. The MIT-BIH arrhythmia database is employed to valid proposed method. Besides, a focal loss function is used to address the classes imbalance of the database. The experimental results show that the positive predictive values of the proposed classification method for N, S, V, and F are 99.10%, 96.80%, 61.32%, and 95.36%, respectively. In addition, the sensitivity values are 95.83%, 95.54%, 90.43%, and 84.79%; the specificity values are 92.96%, 99.88%, 96.05%, and 99.97%, respectively. Compared with the art-of-state inter-patient ECG heartbeat classification approaches, our proposed approach achieved better results. Therefore, the proposed deep learning model of heartbeat classification is effective and feasible for the single-lead ECG signals and interpatient paradigm.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.