Multi-Wound Classification: Exploring Image Enhancement and Deep Learning Techniques

Wounds account for a significant share of hospital visits and deaths across Africa, yet they remain widely underreported, particularly in rural communities with limited access to healthcare. This research explores AI-driven approaches for the online diagnosis of disease-related wounds, supporting telemedicine services for traditional health practitioners and village health workers. Focusing on diabetic, pressure, surgical, and venous ulcers, the study evaluates several machine and deep learning models for wound classification. Among them, the FixCaps model demonstrated the strongest performance, achieving over 93% accuracy while using significantly fewer computational resources. The findings highlight the potential of lightweight, AI-powered diagnostic tools to strengthen wound care and expand access to healthcare in underserved settings.