Global Stereo Reconstruction under Second-Order Smoothness Priors. Woodford, O., Torr, P., Reid, I., & Fitzgibbon, A. Pattern Analysis and Machine Intelligence, IEEE Transactions on, 31(12):2115 –2128, December, 2009.
doi  abstract   bibtex   
Second-order priors on the smoothness of 3D surfaces are a better model of typical scenes than first-order priors. However, stereo reconstruction using global inference algorithms, such as graph cuts, has not been able to incorporate second-order priors because the triple cliques needed to express them yield intractable (nonsubmodular) optimization problems. This paper shows that inference with triple cliques can be effectively performed. Our optimization strategy is a development of recent extensions to alpha-expansion, based on the ldquo QPBOrdquo algorithm. The strategy is to repeatedly merge proposal depth maps using a novel extension of QPBO. Proposal depth maps can come from any source, for example, frontoparallel planes as in alpha-expansion, or indeed any existing stereo algorithm, with arbitrary parameter settings.
@article{woodford_global_2009,
	title = {Global {Stereo} {Reconstruction} under {Second}-{Order} {Smoothness} {Priors}},
	volume = {31},
	issn = {0162-8828},
	doi = {10.1109/TPAMI.2009.131},
	abstract = {Second-order priors on the smoothness of 3D surfaces are a better model of typical scenes than first-order priors. However, stereo reconstruction using global inference algorithms, such as graph cuts, has not been able to incorporate second-order priors because the triple cliques needed to express them yield intractable (nonsubmodular) optimization problems. This paper shows that inference with triple cliques can be effectively performed. Our optimization strategy is a development of recent extensions to alpha-expansion, based on the ldquo QPBOrdquo algorithm. The strategy is to repeatedly merge proposal depth maps using a novel extension of QPBO. Proposal depth maps can come from any source, for example, frontoparallel planes as in alpha-expansion, or indeed any existing stereo algorithm, with arbitrary parameter settings.},
	number = {12},
	journal = {Pattern Analysis and Machine Intelligence, IEEE Transactions on},
	author = {Woodford, O. and Torr, P. and Reid, I. and Fitzgibbon, A.},
	month = dec,
	year = {2009},
	keywords = {arbitrary parameter settings, global inference algorithms, global stereo reconstruction, graph cuts, graph theory, image reconstruction, intractable optimization problems, optimisation, optimization strategy, second-order smoothness priors, stereo image processing},
	pages = {2115 --2128},
}

Downloads: 0