Classification of Pneumonia Using a Deep Learning Convolutional Neural Network

Classification of Pneumonia Using a Deep Learning Convolutional Neural Network

연구 분야: Artificial Intelligence

학회: 2023 IEEE 9th Information Technology International Seminar (ITIS)

An acute lung disease diagnosed by examining white spots on X-ray results is called Pneumonia. However, based on several evaluation results by doctors, those white spots occasionally lead to other lung diseases such as Tuberculosis (TBC). The data-driven approach for computer-aided diagnosis can be one of the innovative solutions to help classify the types of pneumonia that attack the lungs. This research utilizing Deep Convolutional Neural Network (CNN) algorithm to classified Xray images with several models namely GoogLeNet, ShuffleNet V2, and a model designed by the author. From 5,856 data of Xray image results, GoogLeNet was able to detect the accuracy of 74.9%, ShuffleNet V2 was 76%, and the Convolutional Neural Network model that the author designed was 79.8% with the fastest training time than the other models.