Fast MRF Optimization with Application to Depth Reconstruction

CVPR 2014

doi:10.1109/CVPR.2014.500 /cvpr/2014/chen2014cvpr-fast-a/

Abstract

We describe a simple and fast algorithm for optimizing Markov random fields over images. The algorithm performs block coordinate descent by optimally updating a horizontal or vertical line in each step. While the algorithm is not as accurate as state-of-the-art MRF solvers on traditional benchmark problems, it is trivially parallelizable and produces competitive results in a fraction of a second. As an application, we develop an approach to increasing the accuracy of consumer depth cameras. The presented algorithm enables high-resolution MRF optimization at multiple frames per second and substantially increases the accuracy of the produced range images.

PDF CVPR Semantic Scholar

Cite

Text

Chen and Koltun. "Fast MRF Optimization with Application to Depth Reconstruction." Conference on Computer Vision and Pattern Recognition, 2014. doi:10.1109/CVPR.2014.500

Markdown

[Chen and Koltun. "Fast MRF Optimization with Application to Depth Reconstruction." Conference on Computer Vision and Pattern Recognition, 2014.](https://mlanthology.org/cvpr/2014/chen2014cvpr-fast-a/) doi:10.1109/CVPR.2014.500

BibTeX

@inproceedings{chen2014cvpr-fast-a,
  title     = {{Fast MRF Optimization with Application to Depth Reconstruction}},
  author    = {Chen, Qifeng and Koltun, Vladlen},
  booktitle = {Conference on Computer Vision and Pattern Recognition},
  year      = {2014},
  doi       = {10.1109/CVPR.2014.500},
  url       = {https://mlanthology.org/cvpr/2014/chen2014cvpr-fast-a/}
}