Deep Learning for Medical Burn Detection

This research developed a deep learning system for automated burn wound classification, achieving 92% accuracy in distinguishing burn severity levels. The project involved close collaboration with medical professionals to ensure clinical relevance and accuracy. Key contributions: - Custom dataset of 1,000+ labeled burn wound images - YOLOv7-based detection model optimized for medical imaging - Flask-based deployment framework for clinical use - National recognition at ISEF competition (top-30) - Publication at International Conference on Computing & Data Science