Partial enumeration and curvature regularization

Olsson, Carl and Ulen, Johannes and Boykov, Yuri and Kolmogorov, Vladimir N (2013) Partial enumeration and curvature regularization. In: ICCV: International Conference on Computer Vision ICCV, December 1-8, 2013, Sydney, Australia.

[img] Text
iccv13_part_enumeration.pdf - Accepted Version
Available under License All rights reserved.
[IST-2016-566-v1+1]
Download (369Kb)
Official URL: http://dx.doi.org/10.1109/ICCV.2013.365

Abstract

Energies with high-order non-submodular interactions have been shown to be very useful in vision due to their high modeling power. Optimization of such energies, however, is generally NP-hard. A naive approach that works for small problem instances is exhaustive search, that is, enumeration of all possible labelings of the underlying graph. We propose a general minimization approach for large graphs based on enumeration of labelings of certain small patches. This partial enumeration technique reduces complex high-order energy formulations to pairwise Constraint Satisfaction Problems with unary costs (uCSP), which can be efficiently solved using standard methods like TRW-S. Our approach outperforms a number of existing state-of-the-art algorithms on well known difficult problems (e.g. curvature regularization, stereo, deconvolution); it gives near global minimum and better speed. Our main application of interest is curvature regularization. In the context of segmentation, our partial enumeration technique allows to evaluate curvature directly on small patches using a novel integral geometry approach.

Item Type: Conference or Workshop Item (Paper)
Subjects: 000 Computer science, knowledge & general works > 000 Computer science, knowledge & systems
Research Group: Kolmogorov Group
SWORD Depositor: Sword Import User
Depositing User: Sword Import User
Date Deposited: 12 May 2016 14:10
Last Modified: 05 Sep 2017 14:28
URI: https://repository.ist.ac.at/id/eprint/566

Actions (login required)

View Item View Item