MMSR: Multiple-Model Learned Image Super-Resolution Benefiting from Class-Specific Image Priors
Authors: Cansu Korkmaz, A. Murat Tekalp, Zafer Dogan
Venue: IEEE International Conference on Image Processing (ICIP) 2022, Bordeaux, France, 16-19 October 2022
DOI: 10.1109/ICIP46576.2022.9897278
Overview
MMSR introduces a multiple-model approach to learned image super-resolution that leverages class-specific image priors to improve reconstruction quality across different image categories. By training specialized models for different image classes and intelligently combining their outputs, the method achieves superior performance compared to single-model approaches.
Key Contributions
- Multiple-model super-resolution framework
- Class-specific image prior utilization
- Improved reconstruction quality across image categories