<![CDATA[This research aims to develop an automated diagnostic system for classifying lung diseases in X-ray images based on feature extraction using the Gray Level Co-occurrence Matrix (GLCM) with a Backpropagation Artificial Neural Network employing a Python GUI. In this study, 200 lung image data were utilized, divided into four classes with 50 data points each. The four categories of image classes are normal lungs, Pneumonia, Tuberculosis, and Covid-19. The training and testing data were split in a 92:8 ratio, resulting in 184 training data and 16 testing data. The parameters include four input layers, eight hidden layers, two output layers, alpha 0.8, 2000 iteration, and target error = 0.0001. Then, it continued with feature extraction using the GLCM to obtain texture characteristics in lung images. In the training stage, the best results were obtained in iteration 2000 with a Mean Squared Error of 0.005% and a calculated time of 167.319 seconds. At the testing stage, a reasonably high accuracy was obtained, 93.75%, with a calculated time of 0.014 seconds. This result indicates that the method can prove lung images.]]>
Copyrights © 2024
