Learnign a Discriminative Prior for Blind Image Deblurring

Lerenhan Li Jinshan Pan Wei-Sheng Lai Changxin Gao Nong Sang Ming-Hsuan Yang

Abstract

We present an effective blind image deblurring method based on a data-driven discriminative prior. Our work is motivated by the fact that a good image prior should favor clear images over blurred ones. In this work, we formulate the image prior as a binary classifier which can be achieved by a deep convolutional neural network (CNN). The learned prior is able to distinguish whether an input image is clear or not. Embedded into the maximum a posterior (MAP) framework, it helps blind deblurring in various scenarios, including natural, face, text, and low-illumination images. However, it is difficult to optimize the deblurring method with the learned image prior as it involves a non-linear CNN.Therefore, we develop an efficient numerical approach based on the half-quadratic splitting method and gradient descend algorithm to solve the proposed model. Furthermore, the proposed model can be easily extended to non-uniform deblurring. Both qualitative and quantitative experimental results show that our method performs favorably against state-of-the-art algorithms as well as domain-specific image deblurring approaches.

Bibtex

Conference version:

@InProceedings{li2018learning,
               title           = {Learning a Discriminative Prior for Blind Image Deblurring},
               author          = {Li, Lerenhan and Pan, Jinshan and Lai, Wei-Sheng and Gao, Changxin and Sang, Nong and                                Yang, Ming-Hsuan},
               booktitle       = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
               year            = {2018}
}

Journal version:

@article{li2019blind,
         title                = {Blind Image Deblurring via Deep Discriminative Priors},
         author               = {Li, Lerenhan and Pan, Jinshan and Lai, Wei-Sheng and Gao, Changxin and Sang, Nong and Yang, Ming-Hsuan},
         journal              = {International Journal of Computer Vision},
         year                 = {2019}
}

Experimental Results

Deblurred Results on the Köhler dataset

For Baidu Yun link, you might need to refresh again or directly copy the link to a new webpage.

More results are included in Supplemental materials.

Natural Image Deblurring

Blurred Image

Blurred Image

Xu et al. CVPR2013

Pan et al. CVPR2016

Pan et al. CVPR2016

Yan et al. CVPR2017

Ours

Ours

Text Image Deblurring

Blurred Image

Pan et al. CVPR2016

Pan et al. CVPR2014

Ours

Face Image Deblurring

Blurred Image

Xu et al. CVPR2013

Yan et al. CVPR2017

Ours

Quantitative Evaluations on Natural Image Deblurring Datasets

References

[1] L. Xu, S. Zheng, and J. Jia. “Unnatural L0 sparse representation for natural image deblurring”, CVPR 2013.

[2] J. Pan, Z. Hu, Z. Su, and M.-H. Yang. “Deblurring text images via L0-regularized intensity and gradient prior”, CVPR 2014.

[3] J. Pan, D. Sun, H. Pfister, and M.-H. Yang, "Blind Image Deblurring Using Dark Channel Prior", CVPR2016.

[4] Y. Yan, W. Ren, Y. Guo, R. Wang, and X. Cao, "Image Deblurring via Extreme Channels Prior ", CVPR2017.

[5] A. Levin, Y. Weiss, F. Durand, and W. T. Freeman, “Understanding and evaluating blind deconvolution algorithms,” CVPR 2009.

[6] R. Kohler, M. Hirsch, B. Mohler and B. Scholkopf. “Recording and playback of camera shake: benchmarking blind deconvolution with a real-world database”, ECCV 2012.

[7] L. Sun, S. Cho, J.Wang, and J. Hays. "Edge-based blur kernel estimation using patch priors", ICCP 2013.