Learning Generic Diffusion Processes for Image Restoration
This work addresses a domain-specific problem in computer vision by enabling more efficient and transferable regularization for image restoration, though it is incremental as it builds on existing TNRD models.
The authors tackled the trade-off between transferability and performance in image restoration by proposing a generic diffusion process (genericDP) that shares a diffusion term across multiple Gaussian denoising problems, achieving promising denoising performance and high training efficiency, with competitive results when transferred to unseen non-blind deconvolution tasks.
Image restoration problems are typical ill-posed problems where the regularization term plays an important role. The regularization term learned via generative approaches is easy to transfer to various image restoration, but offers inferior restoration quality compared with that learned via discriminative approaches. On the contrary, the regularization term learned via discriminative approaches are usually trained for a specific image restoration problem, and fail in the problem for which it is not trained. To address this issue, we propose a generic diffusion process (genericDP) to handle multiple Gaussian denoising problems based on the Trainable Non-linear Reaction Diffusion (TNRD) models. Instead of one model, which consists of a diffusion and a reaction term, for one Gaussian denoising problem in TNRD, we enforce multiple TNRD models to share one diffusion term. The trained genericDP model can provide both promising denoising performance and high training efficiency compared with the original TNRD models. We also transfer the trained diffusion term to non-blind deconvolution which is unseen in the training phase. Experiment results show that the trained diffusion term for multiple Gaussian denoising can be transferred to image non-blind deconvolution as an image prior and provide competitive performance.