Other Free Encyclopedias » Online Encyclopedia » Encyclopedia - Featured Articles » Contributed Topics from K-O

Occlusions

matching disparity constraints dynamic

Definition: Occlusion regions mark disparity discontinuity jumps, which can be used to improve stereo image encoding and transmission, segmentation, motion analysis, and object identification processes which must preserve object boundaries.

In early stereo image research, the segmentation into visible and occlusion regions is treated as a secondary process, postponed until matching is completed and smoothing is underway. Techniques are also proposed that indirectly address the occlusion problem by minimizing spurious mismatches resulting from occlusion regions and discontinuities. It has been pointed out that occlusion areas must be identified and incorporated into matching process. By using Bayesian reasoning, they derive an energy function based on pixel intensity as the matching feature. Dynamic programming is then used to find a minimal-energy solution.

Techniques based on dynamic programming have been used for the purpose of disparity estimation and simultaneous occlusion detection. A significant advantage of these techniques is that they can provide a global solution for the disparity estimation/occlusion detection problem under local constraints such as constraints related to correlation, smoothness, or disparity gradient limit.

Other methods directly address occlusion regions by defining an a priori probability for the disparity field based upon a smoothness function and an occlusion constraint. For matching two shifted windows are used to avoid errors over discontinuity jumps.

Assuming monotonicity, the matching problem is solved using dynamic programming. The stereo occlusion problem is formulated as a path-finding problem in a left-scanline to right-scanline matching process. It has been also proposed a dynamic programming solution to stereo matching based on matching edge-delimited intervals between corresponding scan lines that does not require the smoothing term. It has been pointed out that several equally good paths can be found through matching space when using only the occlusion and ordering constraints. To provide enough constraints forcing their system to produce a single solution, they optimize a Bayesian maximum likelihood cost function minimizing inter- and intra-scanline disparity discontinuities. A similar strategy uses a dynamic programming algorithm for the detection of significant disparity changes and large occlusion areas in a stereoscopic image pair. In this approach no smoothness or interscan-line compatibility constraints are used. Two additional constraints, namely the extended continuity constraint and the disparity gradient limit, can also be considered. Finally, an algorithm using forms of Bayes decision criteria can be employed for segmenting a stereo pair into occlusion and visible background and foreground regions.

Ochoa, Lorena - Professional golfer, Career, Sidelights [next]

User Comments

Your email address will be altered so spam harvesting bots can't read it easily.
Hide my email completely instead?

Cancel or