LUNGS-CT-SCAN CANCER PREDICTION USING CONVOLUTIONAL NEURAL NETWORKS (CNN) AND VGG16 METHODS

Deep learning plays a vital role in the field of medical science, solving health problems and diagnosing various heart diseases. In this article, we consider CT scanned images of lungs from the Kaggle dataset and classify them into three different Models (M1,M2,M3) generated by the CNN and VGG16 methods. M1 is a convolutional neural network (CNN) based on a deep learning algorithm. The M2 architecture is made of VGG16. Here 16 means it has 16 layers, which is its weight. M3 has augmented data as input, unlike M1. The proposed models have been tested with validation and test sets. In testing, we achieved an accuracy of 78% on the validation set and 50% on the test set for the M1. For M2, we achieved 83.3% accuracy in the validation set and 76.5% in the test set. For M3, we achieved 65% accuracy in the validation set and 57.46% in the test set.