MSD-Net: Multi-Scale Discriminative Network for COVID-19 Lung Infection Segmentation on CT

Abstract

Since the first patient reported in December 2019, 2019 novel coronavirus disease (COVID-19) has become global pandemic with more than 10 million total confirmed cases and 500 thousand related deaths. Using deep learning methods to quickly identify COVID-19 and accurately segment the infected area can help control the outbreak and assist in treatment. Computed tomography (CT) as a fast and easy clinical method, it is suitable for assisting in diagnosis and treatment of COVID-19. According to clinical manifestations, COVID-19 lung infection areas can be divided into three categories: ground-glass opacities, interstitial infiltrates and consolidation. We proposed a multi-scale discriminative network (MSD-Net) for multi-class segmentation of COVID-19 lung infection on CT. In the MSD-Net, we proposed pyramid convolution block (PCB), channel attention block (CAB) and residual refinement block (RRB). The PCB can increase the receptive field by using different numbers and different sizes of kernels, which strengthened the ability to segment the infected areas of different sizes. The CAB was used to fusion the input of the two stages and focus features on the area to be segmented. The role of RRB was to refine the feature maps. Experimental results showed that the dice similarity coefficient (DSC) of the three infection categories were 0.7422,0.7384,0.8769 respectively. For sensitivity and specificity, the results of three infection categories were (0.8593, 0.9742), (0.8268,0.9869) and (0.8645,0.9889) respectively. The experimental results demonstrated that the network proposed in this paper can effectively segment the COVID-19 infection on CT images. It can be adopted for assisting in diagnosis and treatment of COVID-19.

Keywords: COVID-19; CT; MSD segmentation network; deep learning.

FIGURE 1.

CT manifestations of different infection types. The samples in 1–3 column content single category on a CT slice. The samples in 4 and 5 columns include two categories. The last column shows the situation where the three types appear simultaneously. In the figure, green represents ground-glass opacities, yellow represents interstitial infiltrates and red represents consolidation.

FIGURE 2.

Overview of the proposed MSD-Net. PCB: Pyramid convolution block. CAB: Channel attention block. RRB: Residual refinement block. (Best viewed in color).

FIGURE 3.

Detailed structures of the (a) pyramid convolution block, (b) channel attention block, and (c) residual refinement block.

FIGURE 4.

Multi-class lung infection segmentation results obtained by the proposed model and other methods. The green, yellow, and red labels indicate the ground-glass opacities, interstitial infiltrates, and consolidation, respectively. (Best viewed in color).

FIGURE 5.

Examples of three categories feature maps generated by each decoder-layer, the hotter color represents the higher response value. RRB: residual refinement block. (Best viewed in color).

FIGURE 6.

Results with different noise. The first column is the input CT images. The second column is the ground truth, and the third column is the segmentation results. The first row is the original input, the second row is the input which is infected by gaussian noise with a variance of 0.05. The third row is infected with a variance of 0.10.