TR2016-068

Accelerated graph-based nonlinear denoising filters

    •  Knyazev, A., Malyshev, A., "Accelerated Graph-based Nonlinear Denoising Filters", International Conference on Computational Science (ICCS), DOI: 10.1016/​j.procs.2016.05.348, June 2016, vol. 80, pp. 607-616.
      BibTeX TR2016-068 PDF
      • @inproceedings{Knyazev2016jun,
      • author = {Knyazev, Andrew and Malyshev, Alexander},
      • title = {Accelerated Graph-based Nonlinear Denoising Filters},
      • booktitle = {International Conference on Computational Science (ICCS)},
      • year = 2016,
      • volume = 80,
      • pages = {607--616},
      • month = jun,
      • doi = {10.1016/j.procs.2016.05.348},
      • url = {https://www.merl.com/publications/TR2016-068}
      • }
  • Research Area:

    Digital Video

Abstract:

Denoising filters, such as bilateral, guided, and total variation filters, applied to images on general graphs may require repeated application if noise is not small enough. We formulate two acceleration techniques of the resulted iterations: conjugate gradient method and Nesterov's acceleration. We numerically show efficiency of the accelerated nonlinear filters for image denoising and demonstrate 2-12 times speed-up, i.e., the acceleration techniques reduce the number of iterations required to reach a given peak signal-to-noise ratio (PSNR) by the above indicated factor of 2-12.