Acceleration of RED via Vector Extrapolation

@article{Hong2018AccelerationOR,
  title={Acceleration of RED via Vector Extrapolation},
  author={Tao Hong and Yaniv Romano and Michael Elad},
  journal={CoRR},
  year={2018},
  volume={abs/1805.02158}
}
Models play an important role in inverse problems, serving as the prior for representing the original signal to be recovered. REgularization by Denoising (RED) is a recently introduced general framework for constructing such priors using state-of-the-art denoising algorithms. Using RED, solving inverse problems is shown to amount to an iterated denoising process. However, as the complexity of denoising algorithms is generally high, this might lead to an overall slow algorithm. In this paper, we… CONTINUE READING
Recent Discussions
This paper has been referenced on Twitter 11 times over the past 90 days. VIEW TWEETS
1 Citations
35 References
Similar Papers

Citations

Publications citing this paper.

References

Publications referenced by this paper.
Showing 1-10 of 35 references

Svd-mpe: An svd-based vector extrapolation method of polynomial type

  • A. Sidi
  • arXiv preprint arXiv:1505.00674, 2015.
  • 2015
Highly Influential
12 Excerpts

Similar Papers

Loading similar papers…