CNN INCEPTIONRESNET-V2 MACHINE LEARNING ARCHITECTURE FOR PNEUMONIA CHEST X-RAY CLUSTERING

Abstract

In the last century, the use of machine learning, especially the Convolution Neural Network (CNN) has greatly helped the world of health (medicine). Through action research on image datasets, CNN succeeded and was able to show classification or grouping based on the same characteristics and properties on unlabeled images with higher accuracy and faster than other machine learning methods. This is very useful for the world of health, especially in the use of chest x-rays (chest x-rays) in the medical world. This study aims to optimize the CNN InceptionResNet-V2 architecture, for classifying Covid-19 disease, by training 4000 chest x-ray image datasets. The accuracy test results from InceptionResNet-V2 yielded 98%, with the precision of each CNN InceptionRestNet-V2 architecture class being Covid (99%), Lung_Opacity (97%), Normal (98%), Viral_Pneumonia (98%). The CNN InceptionRestNet architecture can help quickly and accurately produce chest x-rays.