MS-DCANet: A Novel Segmentation Network For Multi-Modality COVID-19 Medical Images
This work addresses the challenge of accurate and efficient segmentation of COVID-19 medical images for clinical applications, though it appears incremental as it builds on existing segmentation methods with novel components.
The paper tackles the problem of segmenting COVID-19 medical images with blurred boundaries and low contrast by proposing MS-DCANet, a symmetric segmentation framework that balances accuracy and complexity, achieving state-of-the-art performance on multi-modality tasks and demonstrating strong generalization on other datasets like ISIC 2018 and BAA.
The Coronavirus Disease 2019 (COVID-19) pandemic has increased the public health burden and brought profound disaster to humans. For the particularity of the COVID-19 medical images with blurred boundaries, low contrast and different infection sites, some researchers have improved the accuracy by adding more complexity. Also, they overlook the complexity of lesions, which hinder their ability to capture the relationship between segmentation sites and the background, as well as the edge contours and global context. However, increasing the computational complexity, parameters and inference speed is unfavorable for model transfer from laboratory to clinic. A perfect segmentation network needs to balance the above three factors completely. To solve the above issues, this paper propose a symmetric automatic segmentation framework named MS-DCANet. We introduce Tokenized MLP block, a novel attention scheme that use a shift-window mechanism to conditionally fuse local and global features to get more continuous boundaries and spatial positioning capabilities. It has greater understanding of irregular lesions contours. MS-DCANet also uses several Dual Channel blocks and a Res-ASPP block to improve the ability to recognize small targets. On multi-modality COVID-19 tasks, MS-DCANet achieved state-of-the-art performance compared with other baselines. It can well trade off the accuracy and complexity. To prove the strong generalization ability of our proposed model, we apply it to other tasks (ISIC 2018 and BAA) and achieve satisfactory results.