Lung and Colon Cancer Detection using Deep Learning Techniques

Lung and colon cancer are among the most frequent cancers that claim lives around the globe. For people to heal, early detection and assistance are crucial. Physicians diagnose patients using histopathological images of tissue biopsies from potentially infected areas of the colon and lung. Most types of lung and colon cancer are challenging to diagnose and often require numerous tests. The article presents a novel approach to identify pulmonary and colorectal tumors using deep learning algorithms on histopathology images. Our objective is to develop a dependable and efficient system that uses pre-trained CNNs with LC25000 histopathological image datasets to accurately identify malignant tissues. This work promotes medical examination and artificial intelligence by resolving significant challenges in cancer detection and improving the accuracy and effectiveness of pathological assessment. The implemented deep learning algorithms are: Vgg16 with random forest classifier, Vgg16 with Xgboost classifier, and GLCM with LightGBM. The results of Vgg16 with the Xgboost classifier technique showed better performance, with 97.46% accuracy. In the future, we can use image segmentation to conduct research on nuclei detection and cell segmentation. We can utilize GAN, which significantly enhances the accuracy of medical image segmentation due to its strong generation capabilities and ability to capture data distribution.