AI & Machine Learning Projects

A web-based deep learning system that classifies MRI brain scans into four tumor types. Includes confidence scoring and automated PDF medical report generation.

An efficient CNN-Transformer for single-channel EEG sleep-stage classification with a Transition-Aware Smoothing layer, accepted at IEEE BECITHCON 2026. See full results on the Publications page.

A wavelet-gated CNN-Transformer for detecting Parkinson's freezing-of-gait episodes from wrist-worn accelerometer data, accepted at IEEE BECITHCON 2026 as first author. Full results on the Publications page.