Automated Methods To Detect and Classify Lung & Colon Cancer Using Image Processing and Deep Learning Algorithm

Abstract

One of the most hazardous and severe diseases that people experience globally is colon and lung cancer, which has spread to become a common medical issue. It is very important to make a reliable and early discovery to reduce the danger of mortality. The difficulty of the work ultimately depends on the histopathologists' experience. Underprepared histologists may potentially endanger the patient's life. Recent times have seen a rise in the popularity of deep learning, which is now appreciated in the interpretation of medical imaging. In order to diagnose lung and colon cancer utilizing histopathological image dataset and more effective augmentation approaches, this research aims to leverage and transform the present pre-trained CNN-based architecture. In this study, the LC25000 dataset is used to train three different pre-trained Convolutional Neural Network models: EffecientNetB7, ResNet50 and VGG16. The model performances are evaluated based on accuracy, precision, recall & f1-score. The findings illustrate that all three models produced impressive outcomes, ranging from 93% to 98% accuracy.