A modified DeepLabV3+ based semantic segmentation of chest computed tomography images for COVID‐19 lung infections

Abstract

Coronavirus disease (COVID‐19) affects the lives of billions of people worldwide and has destructive impacts on daily life routines, the global economy, and public health. Early diagnosis and quantification of COVID‐19 infection have a vital role in improving treatment outcomes and interrupting transmission. For this purpose, advances in medical imaging techniques like computed tomography (CT) scans offer great potential as an alternative to RT‐PCR assay. CT scans enable a better understanding of infection morphology and tracking of lesion boundaries. Since manual analysis of CT can be extremely tedious and time‐consuming, robust automated image segmentation is necessary for clinical diagnosis and decision support. This paper proposes an efficient segmentation framework based on the modified DeepLabV3+ using lower atrous rates in the Atrous Spatial Pyramid Pooling (ASPP) module. The lower atrous rates make receptive small to capture intricate morphological details. The encoder part of the framework utilizes a pre‐trained residual network based on dilated convolutions for optimum resolution of feature maps. In order to evaluate the robustness of the modified model, a comprehensive comparison with other state‐of‐the‐art segmentation methods was also performed. The experiments were carried out using a fivefold cross‐validation technique on a publicly available database containing 100 single‐slice CT scans from >40 patients with COVID‐19. The modified DeepLabV3+ achieved good segmentation performance using around 43.9 M parameters. The lower atrous rates in the ASPP module improved segmentation performance. After fivefold cross‐validation, the framework achieved an overall Dice similarity coefficient score of 0.881. The results demonstrate that several minor modifications to the DeepLabV3+ pipeline can provide robust solutions for improving segmentation performance and hardware implementation.

Automatic Lenke classification of adolescent idiopathic scoliosis with deep learning

“…DeepLabV3+ was a novel visual segmentation architecture via the atrous convolution to extract multiple features and ASSP structure to capture information at different scales. [27][28][29][30] And it had been used in medical image region and achieved a better performance in segmentation of radiologic images than other frameworks like FCN, SegNet, and U-Net. [31][32][33][34] Therefore, we tried this model in pedicle detection task and ultimate high-quality consequences for the first time revealed the potential of DeepLabV3+ to be useful for image segmentation in x-ray films.…”

Fighting against COVID‐19: Innovations and applications

“…For segmenting COVID‐19 lesions from CT slices, Ma et al 50 propose the content‐aware pre‐activated residual UNet, CAPA‐ResUNet. Polat 51 presents a useful segmentation framework based on the modified DeepLabV3+ using lower atrous rates in the atrous spatial pyramid pooling module.…”

DeepLab V3+ Based Semantic Segmentation of COVID -19 Lesions in Computed Tomography Images

“…Although, in the case of segmentation problem, it requires increased feature maps resolution and receptive field of neurons. So, an Atrous convolution (Dilated convolution) is used to insert zero values between the weights of the filters for increasing the resolution of the image without the need to increase the number of parameters [5]. Fig.…”

“…One of these models is using semantic segmentation which is based on pixel-wise image classification, where each pixel in the image is classified into one of the predetermined classes. Its architecture is divided into two main parts, namely the encoder for features extraction, and the decoder part for generating the predicted segmented image from the extracted information [5].…”

“…Several researches applied deep learning networks for segmentation using DeepLabV3+, for example in [5] it was designed based on Res-Net 50, and Atrous Spatial Pyramid Pooling (ASPP) module with minor modification in its dilated rate. Small dilated rates, including 4, 8, and 12, were used instead of 6, 12, and 18, to capture small morphological details.…”