Computer-Aided Diagnostic System for Brain Tumor Classification using Explainable AI
S. Padmapriya, M.S. Gayathri Devi
Abstract
The field of computer-aided diagnosis (CAD) of brain tumors has been transformed by developments in medical imaging and artificial intelligence. The accuracy and interpretability of brain tumor classification are improved in this research using Explainable AI (XAI) techniques. A timely and correct diagnosis of brain tumors is essential for the best possible care and treatment of the patient. In contrast, traditional machine learning models often lack transparency and interpretability, making it difficult for clinicians to rely on their judgment. This study uses a Grad-Cam algorithm to create an understandable and interpretable CAD system for classifying brain tumors. Our approach not only achieves high classification accuracy, but also provides physicians with insights into the decision-making process, improving their understanding and confidence in the system’s recommendations. We evaluate our model using a large and diverse dataset and compare it to modern deep learning models and traditional CAD systems. The results show that our CAD system with XAI extensions not only achieves improved accuracy but also provides useful insights into the decision-making process. Using this method, doctors may be able to diagnose patients more accurately.