Lung Cancer Classification in X-Ray Images Using Probabilistic Neural Network

Most hospitals and clinics use x-ray for diagnosis lung disease because the price is relatively cheaper than other lung diagnostic tools. In this research, the chest x-ray image is used as input to the program that consist of two categories such as lung cancer and healthy lung. Categorization images are done by a doctor. The research aimed to classify the X-ray image of the lungs between lung cancer and healthy lung. There are two main stages in this research, namely image processing and classification using a probabilistic neural network. The first step of image processing is preprocessing such as cropping, resizing, thresholding, and median filter. The next step is feature extraction using Haar wavelet transform. The feature of energy and coefficients of each subband produced by Haar wavelet transform is used as input in the classification process. The classification process used a Probabilistic Neural Network (PNN) method to distinguish between lung cancer and healthy lung. The training data used PNN show that all x-ray images could be correctly classified between lung cancer and healthy lung. While test results from PNN using new data obtained at 80 % accuracy rate in detecting abnormalities of the X-ray image of the lungs.