Authors:
S. Rubin Bose, J. Angelin Jeba, N. Christy Evangeline, R. Regin, S. Suman Rajest, Dilli Kasi Rao Kotha
Addresses:
School of Computer Science and Engineering, SRM Institute of Science and Technology, Ramapuram, Chennai, Tamil Nadu, India. Department of Electronics and Communication Engineering, S. A. Engineering College, Chennai, Tamil Nadu, India. Department of Electronics and Instrumentation Engineering, Madras Institute of Technology, Chennai, Tamil Nadu, India. Department of Research and Development, Dhaanish Ahmed College of Engineering, Chennai, Tamil Nadu, India. Faculty of Engineering, Environment and Computing, Coventry University, Coventry, England, United Kingdom.
Abstract:
The early and accurate diagnosis of brain tumours represents a critical challenge in modern neuro-oncology, directly influencing treatment efficacy and patient survival rates. While Magnetic Resonance Imaging (MRI) provides exceptional soft-tissue contrast for visualisation, manual interpretation remains labour-intensive, subjective, and prone to diagnostic delays. This research presents a comprehensive deep learning-based system for automated brain tumour classification from MRI scans, designed as a robust clinical decision-support tool. The system classifies brain MRI images into four clinically relevant categories: Glioma, Meningioma, Pituitary tumour, and No Tumour. Employing transfer learning methodology centred on the Xception Convolutional Neural Network architecture, the model achieves classification accuracy exceeding 95% on held-out test data. A pivotal innovation is the integration of Explainable AI through Gradient-weighted Class Activation Mapping (Grad-CAM), which generates visual heatmaps highlighting regions most influential in classification decisions. The full-stack web application features a Python Flask backend for REST API services and TensorFlow/Keras for image processing and model inference, paired with a React frontend styled with Tailwind CSS. This integrated approach addresses the critical need for both high accuracy and model interpretability in medical AI applications, demonstrating significant potential as a reliable assistive tool for radiologists and oncologists in brain tumour diagnosis workflows.
Keywords: Brain Tumour Detection; Deep Learning; Convolutional Neural Networks; Xception Model; Transfer Learning; Explainable AI; Medical Image Analysis; MRI Classification; Computer-Aided Diagnosis.
Received: 29/12/2024, Revised: 17/04/2025, Accepted: 02/08/2025, Published: 09/12/2025
DOI: 10.64091/ATIHL.2025.000210
AVE Trends in Intelligent Health Letters, 2025 Vol. 2 No. 4 , Pages: 228-239