Publications
Selected research work connecting applied deep learning with medical imaging, biomedical signal processing, and efficient AI systems.
Robust Breast Cancer Classification Across Magnifications Using Fine-Tuned ResNeXt50 with Hybrid Loss
A medical AI study for benign-versus-malignant breast tumor classification from histopathology images, using ResNeXt50_32x4d, attention-aware modeling, patient-wise validation, and hybrid loss design for robust multi-magnification performance.
SpiTranNet-LIF: A Hybrid Spiking–Transformer Framework for Efficient Motor Imagery Decoding
A BCI-focused EEG decoding study that combines spiking neural dynamics with transformer attention for efficient motor imagery classification, emphasizing temporal modeling, compact inference, and neuromorphic design principles.