A Multi-resolution Cooperative Algorithm for Stereo Disparity Analysis

• Main Authors: Wang, Huang-Yu, 王皇又
• Other Authors: Din-Chang Tseng
• Format: Others
• Language: zh-TW
• Published: 1997
• Online Access: http://ndltd.ncl.edu.tw/handle/40724680898697522529

碩士 === 國立中央大學 === 資訊工程研究所 === 85 === This paper presents a model for stereo disparity analysis. The model refers to several psychological and physiological evidences and the model is extended from Marr-Poggio cooperative algorithm. We add multiple- resolution channels into the proposed model. Coarse-to-fine control strategy is used in the model by the inter-channel connections between different resolution channels. In feature-extracting stage, zero crossings are extracted from left and right image pair in all different resolution channels as the matching primitives. In feature-matching stage, the cooperative algorithm is applied in all different resolution channels and is simulated as concurrently working. Coarse-to-fine strategy in multi- resolution channels reduces the iteration steps to obtain the disparities from an image pair. The model is illustrated the performance base on several laboratory scene images.