Abstract: Deep learning–based medical image segmentation models achieve high accuracy but are often impractical for clinical deployment due to their computational and memory demands. This study ...