Examples of non-blind deconvolution obtained with our method. The blurred images on the left (lighthouse) and
bottom right (parrot) were created convolving two images from the Kodak Lossless True Color Image Suite with the 19x19
blur kernel of Krishnan and Fergus and adding 1% of Gaussian noise. The results recovered by our method are shown
on the center and on the top right, respectively. Note the proper reconstruction of fine image details.

Fast High-Quality non-Blind Deconvolution Using Sparse Adaptive Priors

Horacio E. Fortunato horacio_fortunato@uniritter.edu.br

Manuel M. Oliveira oliveira@inf.ufrgs.br

Instituto de Informática, UFRGS

The Visual Computer.
Volume 30, Numbers 6-8, June 2014, pp. 661-671. [DOI]
2nd Best Paper Award at Computer Graphics International 2014

Contents

We present an efficient approach for high-quality non-blind deconvolution based on the use of sparse adaptive priors. Our regularization term enforces preservation of strong edges while removing noise. We model the image-prior deconvolution problem as a linear system, which is solved in the frequency domain. Our approach's clean formulation lends to a simple and efficient implementation. We demonstrate its effectiveness by performing an extensive comparison with existing non-blind deconvolution methods, and by using it to deblur actual photographs degraded by camera shake or motion. Our experiments show that our solution is faster and its results tend to have higher peak signal-to-noise ratio (PSNR) than the state-of-the-art techniques. Thus, it provides an attractive alternative to perform high-quality non-blind deconvolution of large images, as well as to be used as the final step of blind-deconvolution algorithms.

Paper

Full Paper (pre-print)

The final publication is available at Springer

Suplemmentary Materials

Citation

Fortunato, Horacio E. and Manuel M. Oliveira. "Fast high-quality non-blind deconvolution using sparse adaptive priors".  The Visual Computer. Volume 30 (2014), Numbers 6-8,  pp. 661-671.

BibTeX

@article {FortunaoOliveira2014FD,
  author = {Horacio E. Fortunato and Manuel M. Oliveira},
  title = {Fast high-quality non-blind deconvolution using sparse adaptive priors},
  journal = {The Visual Computer},
  year = {2014},
  volume = {30},
  pages = {661-671},
  number = {6-8},
  doi = {10.1007/s00371-014-0966-x},
  issn = {0178-2789}
}

Keywords

Non-blind deconvolution, Adaptive priors, Deblurring, Computational photography.

CNPq-Brazil fellowships and grants #482271/2012-4 and 308936/2010-8.

We thank the authors of the compared techniques for making their code available, and Shan et al. for providing the blurry photographs and camera-motion kernels shown in Figs. 9 and 10. The images used for technique comparisons are from the Kodak Lossless True Color Image Suite (PhotoCD PCD0992).