Mar 29, 2021 · Extensive experiments on synthetic and real-world images demonstrate that the proposed FKP can significantly improve the kernel estimation ...

In this paper, we propose a flow-based kernel prior (FKP) for kernel distribution modeling and incorporate it into ex- isting blind SR models. Based on ...

This repository is the official PyTorch implementation of Flow-based Kernel Prior with Application to Blind Super-Resolution (arxiv, supp). News:.

Extensive experiments on synthetic and real- world images demonstrate that the proposed FKP can sig- nificantly improve the kernel estimation accuracy with less.

To address this issue, this paper proposes a normalizing flow-based kernel prior (FKP) for kernel modeling. By learning an invertible mapping between the ...

To evaluate the effects of the total number of flow blocks and fully-connected network (FCN) depths in FKP, we gen- erate a testing set by sampling kernels ...

This paper proposes a normalizing flow-based kernel prior (FKP) that optimizes the kernel in the latent space rather than the network parameter space, ...

More recently, Liang et al. [13] establish a flow-based kernel prior (FKP) model that realizes an improved performance through pre-training a kernel estimator ...

Flow-based kernel prior with application to blind super-resolution. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition ...

The goal of blind image super-resolution (BISR) is to recover the corresponding high-resolution image from a given low-resolution image with unknown degradation ...