A Similarity Measure for Stereo Feature Matching.pdfVIP

1460 IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 6, NO. 10, OCTOBER 1997 [14] A. Rosenfeld and A. C. Kak, Digital Image Processing , vol. 2, 2nd ed. taking into account such things as end points coordinates, segment New York: Academic, 1982. orientation, and average contrast. The disparity maps of this approach [15] D. Behar, J. Cheung, and L. Kurz, “Contrast techniques for line detection are somewhat sparser than in studies [2], [3], but were obtained with in a correlated noise environment,” IEEE Trans. Image Processing , vol. 6, pp. 625–641, May 1997. a faster execution time. Other related studies [5], [6] show also some interesting use of matching attributes. Weng, Ahuja, and Huang [5] use intensity, edge, and corner attributes for each image point to establish correspondence. A multiresolution (coarse to fine) strategy is used for establishing stereo matches. Nasrabadi [6] applies a curve- segment based matching algorithm. Curve segments are the feature A Similarity Measure for Stereo Feature Matching primitives used in the matching process; thus, curve segments must be extracted from the edge points detected