Virtual Booth

This project presents an automated machine learning–based system for brain tumour classification using Magnetic Resonance Imaging (MRI). Brain tumours are life-threatening conditions that require early and accurate diagnosis; however, manual interpretation of MRI scans is time-consuming, highly dependent on radiologist expertise, and can be affected by image quality variation, subtle tumour characteristics, unclear boundaries, and clinical workload. These challenges may lead to diagnostic inconsistency and delayed treatment decisions. The main objective of this study is to develop a computationally efficient and reliable brain tumour classification system that can support radiologists by improving diagnostic consistency and reducing analysis time. Unlike many existing approaches that rely on deep learning and require large datasets and high computational resources, this project introduces a novel feature-based machine learning approach that utilises basic statistical image features, including mean intensity, standard deviation, entropy, skewness, and kurtosis. These features are extracted from MRI images and classified using a Random Forest algorithm to distinguish between glioma, meningioma, pituitary tumours, and normal brain images. In addition, a graphical user interface (GUI) is developed to enable real-time image upload, preprocessing, and tumour prediction, making the system accessible to users with minimal technical background. The results demonstrate that high classification accuracy can be achieved without complex models, while maintaining low computational requirements. This system offers significant societal benefits, including faster diagnosis, reduced workload for medical professionals, improved accessibility in resource-limited settings, and enhanced support for early detection and clinical decision-making in healthcare environments.