Detection Of CT - Scan Lungs COVID-19 Image Using Convolutional Neural Network And CLAHE
Ronaldus Morgan James, Andi Sunyoto
Abstract
Detecting COVID-19 is a significant task for medical professionals today because of its rapid spread. To overcome this problem, medical professionals have used various techniques and methods to detect to inhibit the proliferation of COVID-19. CT (Computed Tomography) Scan is currently the best method for detecting COVID-19. This diagnostic method is very accurate because it can see organs in three dimensions. However, this method requires a radiologist to detect the disease and requires a long time, which means it will cut valuable time for medical practitioners if a patient is sick. Therefore it is necessary to implement a system to detect the coronavirus automatically as an alternative quickly. This study intends to help medical practitioners to detect computed tomography (CT) Scans of lungs infected with COVID-19. The methods to be used are Limited Adaptive histogram equalization (CLAHE) contrast to improve the quality of CT (Computed Tomography) Scan images of COVID-19 lungs and Convolutional Neural Network (CNN) for the image classification process. The dataset used is 698 RGB images. This study uses three convolutional layers, 3 Max-Pooling layers, and two fully connected layers, resulting in 83.28% accuracy.