Flow-based Kernel Prior with Application to Blind Super-Resolution

by   Jingyun Liang, et al.
ETH Zurich

Kernel estimation is generally one of the key problems for blind image super-resolution (SR). Recently, Double-DIP proposes to model the kernel via a network architecture prior, while KernelGAN employs the deep linear network and several regularization losses to constrain the kernel space. However, they fail to fully exploit the general SR kernel assumption that anisotropic Gaussian kernels are sufficient for image SR. To address this issue, this paper proposes a normalizing flow-based kernel prior (FKP) for kernel modeling. By learning an invertible mapping between the anisotropic Gaussian kernel distribution and a tractable latent distribution, FKP can be easily used to replace the kernel modeling modules of Double-DIP and KernelGAN. Specifically, FKP optimizes the kernel in the latent space rather than the network parameter space, which allows it to generate reasonable kernel initialization, traverse the learned kernel manifold and improve the optimization stability. Extensive experiments on synthetic and real-world images demonstrate that the proposed FKP can significantly improve the kernel estimation accuracy with less parameters, runtime and memory usage, leading to state-of-the-art blind SR results.


page 3

page 6

page 8


Spectrum-to-Kernel Translation for Accurate Blind Image Super-Resolution

Deep-learning based Super-Resolution (SR) methods have exhibited promisi...

Deep Constrained Least Squares for Blind Image Super-Resolution

In this paper, we tackle the problem of blind image super-resolution(SR)...

Mutual Affine Network for Spatially Variant Kernel Estimation in Blind Image Super-Resolution

Existing blind image super-resolution (SR) methods mostly assume blur ke...

Meta-Learned Kernel For Blind Super-Resolution Kernel Estimation

Recent image degradation estimation methods have enabled single-image su...

Learning Generalizable Latent Representations for Novel Degradations in Super Resolution

Typical methods for blind image super-resolution (SR) focus on dealing w...

Temporal Kernel Consistency for Blind Video Super-Resolution

Deep learning-based blind super-resolution (SR) methods have recently ac...

Kernel-aware Raw Burst Blind Super-Resolution

Burst super-resolution (SR) provides a possibility of restoring rich det...

Code Repositories


Official PyTorch code for Flow-based Kernel Prior with Application to Blind Super-Resolution (FKP), CVPR2021

view repo

Please sign up or login with your details

Forgot password? Click here to reset