Diagnosis of COVID-19 from chest X-ray images using wavelets-based depthwise convolution network

Abstract

Coronavirus disease 2019 also known as COVID-19 has become a pandemic. The disease is caused by a beta coronavirus called Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2). The severity of the disease can be understood by the massive number of deaths and affected patients globally. If the diagnosis is fast-paced, the disease can be controlled in a better manner. Laboratory tests are available for diagnosis, but they are bounded by available testing kits and time. The use of radiological examinations that comprise Computed Tomography (CT) can be used for the diagnosis of the disease. Specifically, chest X-Ray images can be analysed to identify the presence of COVID-19 in a patient. In this paper, an automated method for the diagnosis of COVID-19 from the chest X-Ray images is proposed. The method presents an improved depthwise convolution neural network for analysing the chest X-Ray images. Wavelet decomposition is applied to integrate multiresolution analysis in the network. The frequency sub-bands obtained from the input images are fed in the network for identifying the disease.The network is designed to predict the class of the input image as normal, viral pneumonia, and COVID-19. The predicted output from the model is combined with Grad-CAM visualization for diagnosis. A comparative study with the existing methods is also performed. The metrics like accuracy, sensitivity, and F1-measure are calculated for performance evaluation. The performance of the proposed method is better than the existing methodologies and thus can be used for the effective diagnosis of the disease.

Objective evaluation of deep uncertainty predictions for COVID-19 detection

“…The input modalities are often chest X-ray and CT scan images which are processed using a wide variety of DNNs including pretrained networks for feature extraction, segmentation, and generative adversarial networks (GANs) 20 . For example, Singh and Singh 21 proposed an automated COVID-19 diagnosis method based on DNNs and Wavelet decomposition. Their model could classify input sample (i.e.…”

Heterogeneous Systolic Array Architecture for Compact CNNs Hardware Accelerators

“…Compared to SConv, DWConv saves lots of parameters and Multiply-Accumulate Operations (MACs) while maintaining high accuracy. Due to the advantages of DWConv, the compact CNNs using DWConv have become a hot research topic and a series of compact CNNs models have also been widely used [4], [5], [6], [7], [8], [9].…”

AIoT Used for COVID-19 Pandemic Prevention and Control

“…The result of the algorithm reaches 95.79%, which illustrates that the model can be used for early diagnosis. Singh and Singh [ 28 ] put an improved deep convolutional neural network for automatic diagnosis of COVID-19. The advantages which are brought by the combination of Wavelet Transform and Deep Network can be widely used to diagnose COVID-19 from chest X-ray images.…”