Efficient belief propagation for early vision

DMCA

Efficient belief propagation for early vision (2004)

Cached

Download Links

Other Repositories/Bibliography

DBLP

by Pedro F. Felzenszwalb , Daniel P. Huttenlocher

Venue:	In CVPR
Citations:	515 - 8 self

BibTeX

@INPROCEEDINGS{Felzenszwalb04efficientbelief,
    author = {Pedro F. Felzenszwalb and Daniel P. Huttenlocher},
    title = {Efficient belief propagation for early vision},
    booktitle = {In CVPR},
    year = {2004},
    pages = {261--268}
}

OpenURL

Abstract

Markov random field models provide a robust and unified framework for early vision problems such as stereo, optical flow and image restoration. Inference algorithms based on graph cuts and belief propagation yield accurate results, but despite recent advances are often still too slow for practical use. In this paper we present new algorithmic techniques that substantially improve the running time of the belief propagation approach. One of our techniques reduces the complexity of the inference algorithm to be linear rather than quadratic in the number of possible labels for each pixel, which is important for problems such as optical flow or image restoration that have a large label set. A second technique makes it possible to obtain good results with a small fixed number of message passing iterations, independent of the size of the input images. Taken together these techniques speed up the standard algorithm by several orders of magnitude. In practice we obtain stereo, optical flow and image restoration algorithms that are as accurate as other global methods (e.g., using the Middlebury stereo benchmark) while being as fast as local techniques. 1

Keyphrases

early vision efficient belief propagation optical flow image restoration inference algorithm belief propagation approach middlebury stereo benchmark small fixed number markov random field model standard algorithm good result message passing iteration present new algorithmic technique early vision problem graph cut running time possible label global method several order practical use belief propagation yield accurate result input image second technique recent advance local technique large label unified framework

DMCA