Covid-19 Classification Based on Gray-Level Co-occurrence Matrix and Support Vector Machine

Abstract

AbstractCovid-19 is a new epidemic recently. Early diagnosis of related diseases relies on the analysis of the patient’s clinical symptoms and kit testing. To identify this disease efficiently and automatically, we proposed an effective classification system by identifying CT images of chest based on Histogram Equalization (HE), Gray-Level Co-Occurrence Matrix (GLCM) and Support Vector Machine (SVM) algorithm. We collected 148 CT images of healthy people and 148 CT images of patients as our first-hand dataset, the size of which is 512*512*3. To enhance the features of the images, we center cropped the images to 400*400*3. GLCM is an efficient method to extract features focusing on the texture features and SVM can be accurately utilized to classify. In our experiment, we proposed a 10-fold Cross-Validation (CV) to ensure the reliability of experimental results. The results show that the average accuracy of our system is better than other common methods. The performance of our proposed method is effective for Covid-19 identification.KeywordsCovid-19Histogram equalizationGray-Level Co-occurrence matrix10-fold Cross-ValidationSupport vector machine