Denoising Diffusion Probabilistic Models for Robust Image Super-Resolution in the Wild
This addresses robust image enhancement for real-world applications where input quality varies, representing a strong but incremental advance over prior diffusion and GAN methods.
The paper tackles blind super-resolution for out-of-distribution images with unknown degradations by introducing SR3+, a diffusion-based model that establishes a new state-of-the-art, outperforming Real-ESRGAN with FID scores of 36.82 vs. 37.22 and improving to 32.37 with larger models.
Diffusion models have shown promising results on single-image super-resolution and other image- to-image translation tasks. Despite this success, they have not outperformed state-of-the-art GAN models on the more challenging blind super-resolution task, where the input images are out of distribution, with unknown degradations. This paper introduces SR3+, a diffusion-based model for blind super-resolution, establishing a new state-of-the-art. To this end, we advocate self-supervised training with a combination of composite, parameterized degradations for self-supervised training, and noise-conditioing augmentation during training and testing. With these innovations, a large-scale convolutional architecture, and large-scale datasets, SR3+ greatly outperforms SR3. It outperforms Real-ESRGAN when trained on the same data, with a DRealSR FID score of 36.82 vs. 37.22, which further improves to FID of 32.37 with larger models, and further still with larger training sets.