Penerapan Algoritma Convolutional Neural Network Arsitektur ResNet50V2 Untuk Mengidentifikasi Penyakit Pneumonia

Abstract

Pneumonia is a disease that infects the respiratory tract, disrupting the normal function of the human body. Viruses and bacteria are known as common causes of pneumonia. Identification of Pneumonia can use Convolutional Neural Network (CNN). CNN is an effective artificial neural network architecture for image analysis, inspired by how the human brain processes visual information. CNNs are capable of understanding the hierarchical features in images, from lines and angles to complex shapes and objects. This research aims to use ResNet50V2, a popular CNN architecture, to classify X-ray images as either normal or indicative of pneumonia, with the goal of creating an accurate and efficient diagnostic tool. The research method involves using X-ray image datasets for training, validation, and testing, using the ResNet50V2 CNN architecture. The test results show that ResNet50V2 achieves a pneumonia classification accuracy of 93.26%. This study innovatively explores alternative CNN architectures for pneumonia classification, focusing on ResNet50V2.