Detection of COVID from Chest X-Ray Images using Pivot Distribution Count Method

Abstract

The Diagnosis of Corona Virus Disease (COVID) manually from a Chest X-Ray (CX-R) is time-consuming and may be inaccurate. In this paper, a new feature extraction method called the "Pivot Distribution Count (PDC)" method has been proposed, which finds the white spots in COVID infected lungs. The state of art method called "Gliding Box Method (GBM)" and a recently developed technique called Pixel Range Calculation (PRC) method have been applied for comparing the results obtained from texture features from the Chest X-Ray (CX-R) images with that of the proposed method. For carrying out the experiment Chest X-Ray dataset from the Kaggle database has been used. From the experimental result, it is observed that the PDC and PRC method has got the maximum detection rate of 100%, whereas, GBM detects COVID with a detection rate of 56%. For Non-COVID samples, the PDC method outperforms the other two methods with an accuracy of 96%.