×

Improving stereovision matching through supervised learning. (English) Zbl 0911.68210

Summary: Most classical local stereovision matching algorithms use features representing objects in both images and compute the minimum difference attribute values. We have verified that the differences in attributes for the true matches cluster in a cloud around a centre. The correspondence is established on the basis of the minimum squared Mahalanobis distance between the difference of the attributes for a current pair of features and the cluster centre (similarity constraint). We introduce a new supervised learning strategy derived from the learning vector quantization (LVQ) approach to get the best cluster centre. Additionally, we obtain the contribution or specific weight of each attribute for matching. We improve the learning law introducing a variable learning rate. The supervised learning and the improved learning law are the most important findings, which are justified by the computed better results compared with classical local stereovision matching methods without learning and with other learning strategies. The method is illustrated with 47 pairs of stereo images.

MSC:

68U10 Computing methodologies for image processing
PDFBibTeX XMLCite
Full Text: DOI

References:

[1] Lee SH, Leou JJ. A dynamic programming approach to line segment matching in stereo vision. Pattern Recognition 1994; 27: 961-986
[2] Marr D, Poggio T. A computational theory of human stereovision. Proc Roy Soc Lond B 1979; 207: 301-328
[3] Medioni G, Nevatia R. Segment based stereo matching. Computer Vision, Graphics and Image Processing 1985; 31: 2-18
[4] Pollard SB, Mayhew JEW, Frisby JP. PMF: A stereo correspondence algorithm using a disparity gradient limit. Perception 1981; 14: 449-470
[5] Kohonen T. Self-Organization and Associative Memory, Springer-Verlag, New York, 1989 · Zbl 0528.68062
[6] Kohonen T. Self-Organizing Maps, Springer-Verlag, Berlin, 1995 · Zbl 0957.68097
[7] Patterson DW. Artificial Neural Networks Prentice-Hall, Singapore, 1996
[8] Wu JK. Neural Networks and Simulation Methods, Marcel Dekker, New York, 1994
[9] Cruz JM, Pajares G, Aranda J. A neural network approach to the stereovision correspondence problem by unsupervised learning. Neural Networks 1995; 8(5): 805-813 · Zbl 05479111
[10] Pajares G. Estrategia de Solucion al Problema de la Correspondencia en Vision Estereoscópica por la Jerarquía Metodológica y la Integración de Criterios. PhD thesis Dpto. Informática y Automática, Facultad Ciencias UNED: Madrid, 1995
[11] Pajares G, Cruz JM, Aranda J. Stereo matching based on the self-organizing feature mapping algorithm. Pattern Recognition Letters 1998 (accepted) · Zbl 0905.68118
[12] Pajares G, Cruz JM, Aranda J. Relaxation by Hopfield network in stereo image matching. Pattern Recognition 1998; 31(5): 561-574
[13] Dhond AR, Aggarwal JK. Structure from stereo ? a review. IEEE Trans Syst Man Cybern 1989; 19: 1489-1510
[14] Ozanian T. Approaches for stereo matching ? a review. Modeling Identification Control 1995; 16(2): 65-94 · Zbl 0854.68108
[15] Fua P. A parallel algorithm that produces dense depth maps and preserves image features. Machine Vision and Applic 1993; 6: 35-49
[16] Kim YC, Aggarwal JK. Positioning three-dimensional objects using stereo images. IEEE J Robotics and Automation 1987; 3(4): 361-373
[17] Mousavi MS, Schalkoff RJ. ANN Implementation of stereo vision using a multi-layer feedback architecture. IEEE Trans Sys Man Cybern 1994; 24(8): 1220-1238
[18] Ayache N, Faverjon B. Efficient registration of stereo images by matching graph descriptions of edge segments. Int J Computer Vision 1987; 1: 107-131
[19] Ayache N. Artificial Vision for Mobile Robots: Stereo Vision and Multisensory Perception, MIT Press, Cambridge, MA, 1991
[20] Kim DH, Choi WY, Park RH. Stereo matching technique based on the theory of possibility. Patt Recognition Letters 1994; 13: 735-744 · Zbl 05474489
[21] Hoff W, Ahuja N. Surface from stereo: integrating feature matching, disparity estimation, and contour detection. IEEE Trans Patt Anal Machine Intell 1989; 11: 121-136 · Zbl 05110678
[22] Wuescher DM, Boyer KL. Robust contour decomposition using a constraint curvature criterion. IEEE Trans Patt Anal Machine Intell 1991; 13(1): 41-51 · Zbl 05110471
[23] Breuel TM. Finding lines under bounded error. Pattern Recognition 1996; 29 (1): 167-178 · Zbl 05475974
[24] Zhou Y, Chellapa R. Artificial Neural Networks for Computer Vision, Springer-Verlag, Berlin, 1992
[25] Yu SS, Tsai WH. Relaxation by the Hopfield neural network. Pattern Recognition 1992; 25(2): 197-209
[26] Ruichek Y, Postaire JG. A neural matching algorithm for 3-D reconstruction from stereo pairs of linear images. Pattern Recognition Letters 1996; 17: 387-398 · Zbl 05476849
[27] Lew MS, Huang TS, Wong K. Learning and feature selection in stereo matching. IEEE Trans Patt Anal Machine Intell 1994; 16(9): 869-881 · Zbl 05112329
[28] Huertas A, Medioni G. Detection of intensity changes with subpixel accuracy using Laplacian-Gaussian masks. IEEE Trans Patt Anal Machine Intell 1986; 8(5): 651-664
[29] Leu JG, Yau HL. Detecting the dislocations in metal crystals from microscopic images. Pattern Recognition 1991; 24(1): 41-56
[30] Krotkov EP. Active Computer Vision by Cooperative focus and Stereo, Springer-Verlag, Berlin, 1989 · Zbl 0825.68033
[31] Kahn P, Kitchen L, Riseman EM. A fast line finder for vision-guided robot navigation. IEEE Trans Patt Anal Machine Intell 1990; 12(11): 1098-1102 · Zbl 05112702
[32] Maravall D. Reconocimiento de Formas y Visión Artificial, RAMA, Madrid, 1993
[33] Tanaka S, Kak AC. A rule-based approach to binocular stereopsis. In: Jain RC, Jain AK (eds) Analysis and Interpretation of Range Images, Springer-Verlag, 1990
[34] Nevatia R, Babu KR. Linear feature extraction and description. Computer Vision, Graphics and Image Process 1980; 13: 257-269
[35] Mardia KV. Statistics of Directional Data, Academic Press, London, 1972 · Zbl 0244.62005
[36] Duda RO, Hart PE. Pattern Classification and Scene Analysis. Wiley, New York, 1973 · Zbl 0277.68056
[37] Kosko B. Neural Networks and Fuzzy Systems, Prentice-Hall, Englewood Cliffs, NJ, 1992 · Zbl 0755.94024
[38] Martin-Smith P, Pelayo FJ, Diaz A, Ortega J, Prieto A. A learning algorithm to obtain self-organizing maps using fixed neighbourhood Kohonen networks. In: Mira J, Cabestany J, Prieto A (eds), New Trends in Neural Computation, Springer-Verlag, 1993, 297-304
[39] Haykin S. Neural Networks: A Comprehensive Foundation, Macmillan. New York, 1994 · Zbl 0828.68103
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.