Good Image Priors for Non-blind Deconvolution: Generic vs Specific

Libin Sun1      Sunghyun Cho2     Jue Wang2     James Hays1     
1Brown University            2Adobe Research

Abstract

Most image restoration techniques build "universal" image priors, trained on a variety of scenes, which can guide the restoration of any image. But what if we have more specific training examples, e.g. sharp images of similar scenes? Surprisingly, state-of-the-art image priors don't seem to benefit from from context-specific training examples. Re-training generic image priors using ideal sharp example images provides minimal improvement in non-blind deconvolution. To help understand this phenomenon we explore non-blind deblurring performance over a broad spectrum of training image scenarios. We discover two strategies that become beneficial as example images become more context-appropriate: (1) locally adapted priors trained from region level correspondence significantly outperform globally trained priors, and (2) a novel multi-scale patch-pyramid formulation is more successful at transferring mid and high frequency details from example scenes. Combining these two key strategies we can qualitatively and quantitatively outperform leading generic non-blind deconvolution methods when context-appropriate example images are available. We also compare to recent work which, like ours, tries to make use of context-specific examples.

paper thumbnail

Paper

goodpriors_eccv2014.pdf, 20MB

Supplementary Materials

Full Results, 53MB

Poster

goodpriors_eccv2014_poster.pdf, 12MB

Citation

Libin Sun, Sunghyun Cho, Jue Wang, James Hays. Good Image Priors for Non-blind Deconvolution: Generic vs Specific. Proceedings of the European Conference on Computer Vision (ECCV), 2014.

Bibtex

@inproceedings{goodpriors_eccv2014,
  Author    = {Libin Sun and Sunghyun Cho and Jue Wang and James Hays},
  Title     = {Good Image Priors for Non-blind Deconvolution: Generic vs Specific},
  Booktitle = {Proceedings of the European Conference on Computer Vision (ECCV)},
  Year      = {2014}}

Example Results

Comparison of our non-blind deconvolution results against numerous well-known methods in generic non-blind deconvolution (2~5) as well as leading method in example-based deconvolution (8), please choose a test image and click through the slider buttons. Please note that methods 2 through 5 do not make use of specific example images, whereas method 8 and ours do. For full detailed results, please refer to our supplementary material. From left to right: