Classification of COVID-19 on Chest X-Ray Images Using Deep Learning Model with Histogram Equalization and Lung Segmentation

Artificial intelligence techniques coupled with biomedical analysis have been play a critical role during COVID-19 pandemics as it helps to release the overwhelming pressure from healthcare systems and physicians. As the ongoing COVID-19 crisis worsens in countries having dense populations and inadequate testing kits like Brazil, India, and United States of America, the radiological imaging can act as an important diagnostic tool to accurately classify COVID-19 patients and prescribe the necessary treatment in due time. With this motivation, a deep learning-based architecture is proposed in this study for detecting COVID-19 infected lungs using chest X-ray (CXR) images. A total of 2470 CXR images are collected for three different class labels namely healthy lung, ordinary pneumonia, and COVID-19 infected pneumonia, out of which 470 CXR images belong to the COVID-19 category. Primarily, all CXR images are preprocessed by using histogram equalization technique over median filter and segment them through U-Net architecture. Afterward, a VGG-16 model is utilized for feature extraction from the preprocessed CXR images, which is further sampled through synthetic minority oversampling technique to realize a balanced dataset. Finally, the class-balanced features are classified using a support vector machine classifier with 10-fold cross-validation, and the performance metrics are evaluated. The proposed novel approach combining well-known preprocessing techniques, feature extraction methods, and dataset balancing method, leads us to an outstanding rate of recognition of 98