Crafting Training Degradation Distribution for the Accuracy-Generalization Trade-off in Real-World Super-Resolution
This addresses the accuracy-generalization trade-off for real-world super-resolution applications, but it is incremental as it builds on existing degradation modeling approaches.
The paper tackles the trade-off between generalization and accuracy in real-world super-resolution by crafting training degradation distributions using reference images, resulting in significant performance improvements on test images while maintaining generalization.
Super-resolution (SR) techniques designed for real-world applications commonly encounter two primary challenges: generalization performance and restoration accuracy. We demonstrate that when methods are trained using complex, large-range degradations to enhance generalization, a decline in accuracy is inevitable. However, since the degradation in a certain real-world applications typically exhibits a limited variation range, it becomes feasible to strike a trade-off between generalization performance and testing accuracy within this scope. In this work, we introduce a novel approach to craft training degradation distributions using a small set of reference images. Our strategy is founded upon the binned representation of the degradation space and the Fréchet distance between degradation distributions. Our results indicate that the proposed technique significantly improves the performance of test images while preserving generalization capabilities in real-world applications.