Detection and Classification of Covid-19 from X-Ray Images using SVM-DCNN

Authors

  • K. Ravishankar, C. Jothi Kumar

Abstract

There is lack of attention on pre-processing the datasets to remove the diaphragm regions, normalize image contrast and reduce image noises. Apart from Covid-19 Perception and classification, the severity of the COVID-19 infection has to be determined by applying deep learning (DL) based segmentation techniques which localizes the infection region. Hence in order to overcome these issues, a detection and classification technique for Covid-19 from X-ray images using Support Vector Machine (SVM) and Deep Convolutional Neural Network (DCNN) is proposed. In the pre-processing stage, a Modified Anisotropic Diffusion Filtering (MADF) method was propelled to remove the noises from the images.  After the pre-processing step, the features Histogram-oriented gradient (HOG) and Image profile (IP) are extracted and fused. The fused feature is used then trainedand classified using hybrid SVM-CNN algorithm. It classifies the chest X-ray images into three categories that includes Covid-19, Pneumonia and normal.  Experimental results have shown that the proposed SVM-DCNN algorithm attains highest accuracy of 91.8, Precision of 91.9, Recall of 88.5 and F1-score of 94.2 when compared to DCNN and SVM.

Downloads

Published

2022-07-21