Multilayer Stereo Image Matching Based upon Phase-Magnitude and Mean Field Approximation
Hong Jin Jeong, Jung‐Gu Kim, Myoung-Sik Chae · Open Access System for Information Sharing (Pohang University of Science and Technology) · 1997
This paper introduces a new energy function, as a maximum a posteriori(MAP) estimate of binocular dispariry, thar can deal with both random dot stereo-gram(RDS) and natural scenes. The energy function uses phase-magnitude as features to detect only the shift for a pair of corrupted conjugate images. Also we adopted Fleet singularity that effectively detects unstable areas of image plane and thus eliminates in advance error-prone stereo matching. The multi-scale concept is applied to the multi layer architecture that can search the solutions systematically from coarse to fine details and thereby avoids drastically the local minima. Using mean field approximation, we obtained a compact representation that is suitable for fast computation. In this manner, the energy function satisfies major natural constraints and requirements for implementing parallel relaxation. As an experiment, the proposed algorithm is applied to RDS and natural stereo images. As a result we will see that it reveals good performance in terms of recognition errors, parallel implementation, and noise characteristics.