New AI Model Sharpens Blurry Faces Using Semantic Masks

Researchers introduced SMFD-UNet, a lightweight framework that restores high-quality facial images from blurry inputs by extracting semantic face masks (eyes, nose, mouth) directly from degraded photos. The model outperforms existing deblurring methods and could improve facial recognition, forensics, and medical imaging without requiring reference images.