ADVANCING COLON CANCER DIAGNOSIS: A NOVEL APPROACH TO CLASSIFICATION THROUGH EXPLAINABLE AI METHODOLOGIES

Abstract

Deep learning algorithms are implemented to accurately find and diagnose colon cancer, specifically focusing on adenocarcinomas and benign colonic tissues. The EfficientNet model is a part of the convolutional neural networks (CNNs) with high accuracy. The priority is interpretability by integrating Explainable AI (XAI) and an XAI model that uses Grad-CAM to highlight and focus the regions implanted in the images used, enabling better comprehension of the diagnostic process. Healthcare can benefit a lot with the help of deep learning models and transparent AI approaches. The proposed technique provides doctors with insights into the model’s decision-making, which creates the ability to accelerate clinical decision-making, improve diagnostic accuracy, and ultimately improve their outcomes.