CHS-Net: A Deep Learning Approach for Hierarchical Segmentation of COVID-19 via CT Images

Abstract

The pandemic of novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) also known as COVID-19 has been spreading worldwide, causing rampant loss of lives. Medical imaging such as computed tomography (CT), X-ray, etc., plays a significant role in diagnosing the patients by presenting the visual representation of the functioning of the organs. However, for any radiologist analyzing such scans is a tedious and time-consuming task. The emerging deep learning technologies have displayed its strength in analyzing such scans to aid in the faster diagnosis of the diseases and viruses such as COVID-19. In the present article, an automated deep learning based model, COVID-19 hierarchical segmentation network (CHS-Net) is proposed that functions as a semantic hierarchical segmenter to identify the COVID-19 infected regions from lungs contour via CT medical imaging using two cascaded residual attention inception U-Net (RAIU-Net) models. RAIU-Net comprises of a residual inception U-Net model with spectral spatial and depth attention network (SSD) that is developed with the contraction and expansion phases of depthwise separable convolutions and hybrid pooling (max and spectral pooling) to efficiently encode and decode the semantic and varying resolution information. The CHS-Net is trained with the segmentation loss function that is the defined as the average of binary cross entropy loss and dice loss to penalize false negative and false positive predictions. The approach is compared with the recently proposed approaches and evaluated using the standard metrics like accuracy, precision, specificity, recall, dice coefficient and Jaccard similarity along with the visualized interpretation of the model prediction with GradCam++ and uncertainty maps. With extensive trials, it is observed that the proposed approach outperformed the recently proposed approaches and effectively segments the COVID-19 infected regions in the lungs.

Weakly supervised segmentation of COVID-19 infection with local lesion coherence on CT images

“…In particular [21] , a novel deep learning framework structure with multi-view slice decomposition is proposed for fine-grained CT lesion area segmentation evaluation, and prior knowledge is integrated into the model training to effectively improve the model performance. In [22] , a multilayer segmentation network learning method, CHS-Net, was designed capable of automatic deep learning of COVID-19 from CT images, with shrinking and expanding phases of depth-separable convolution and hybrid pooling to efficiently encode and decode semantic and focal region feature information. In [23] , Lung Ultrasound data and deep learning were proposed as a deep fusion artificial neural network framework to assess COVID-19 severity using Lung Ultrasound data.…”

A teacher–student framework with Fourier Transform augmentation for COVID-19 infection segmentation in CT images

“…Deep-learning-based segmentation methods have played an essential role in diagnosing COVID-19 [25] , [26] , [27] , [28] , [29] . Some studies explored to construct new networks suitable for COVID-19 CT segmentation task [30] , [31] , [32] . For example, Ouyang et al.…”

LWU-Net approach for Efficient Gastro-Intestinal Tract Image Segmentation in Resource-Constrained Environments

“…U-Net Ronneberger et al (2015), the foundation of our proposed architecture, has been widely adopted in medical imaging tasks due to its ability to capture detailed anatomical structures. Many studies have explored different variants of U-Net for specific medical segmentation tasks Punn and Agarwal (2022c), such as brain tumor segmentation Punn and Agarwal (2021), retinal vessel segmentation Yue et al (2019), anatomical brain segmentation using DenseNetGottapu and Dagli (2018), Tran et al (2021) employed a Triple-unet with multi-scale input features and dense skip connection and Huang et al (2020) proposed a UNet3+ a full scale connected U-Net, Xiao et al (2018) proposed a weighted ResUNet for high quality retina vessel segmentation and Punn and Agarwal (2022a) proposed a self-supervised Unet framework for biomedical image segmentation applications, Jafari et al (2020) employed an efficient deep convolutional neural network, while Punn and Agarwal (2022b) proposed attention based U-Net and Li et al (2020) employed a nested attention aware U-Net for liver CT image segmentation Accurate segmentation of the GI tract from medical images is essential for various applications, including polyp detection, pathology assessment, and surgical planning. Several studies have focused on GI tract segmentation, with approaches ranging from traditional techniques to deep learning-based methods.…”