English
Abstract
Chronic wound diagnosis using machine learning is hindered by the lighting and shadows in the captured images. Therefore, this study evaluates the impact of different color spaces on wound segmentation using a U-Net. Results show that the YDbDr color space outperforms RGB.
Fields of Study & Subjects
Medical Image Segmentation Techniques
Advanced Neural Network Applications
Smart Agriculture and AI
Publication details
- Journal / source
- RECIBE REVISTA ELECTRÓNICA DE COMPUTACIÓN INFORMÁTICA BIOMÉDICA Y ELECTRÓNICA
- Volume
- 14
- Issue
- 3
- Publication date
- 13 January 2026