Abstract
In this paper, we consider how machine learning can be used to help solve the problem of identifying objects or structures composed of parts as they occur in complex scenes. We first discuss an automatic conditional rule generation technique (CRG) that is designed to describe structures via part attributes and their relations. It does so by generating part-indexed decision trees where the branches define the types of pattern structures necessary to identify and to generalize from the different training examples. We then show how the resultant rules can be used for region labeling, and we examine grouping and constraint propagation techniques that are required for the identification of objects in complex scenes.
Supported by Grant OGP38521 from the Natural Sciences and Engineering Research Council of Canada.
Chapter PDF
References
Bischof, W. F., Caelli, T.: Learning structural descriptions of patterns: A new technique for conditional clustering and rule generation. Pattern Recognition 27 (1994) 689–698
Caelli, T., Dreier, A.: Variations on the evidenced-based object recognition theme. Pattern Recognition 27 (1994) 185–204
Grimson, W. E. L.: Object Recognition by Computer. Cambridge: MIT Press (1990)
Jain, A. K., Hoffman, D.: Evidence-based recognition of objects. IEEE Transactions on Pattern Analysis and Machine Intelligence 10 (1988) 783–802
Quinlan, J. R.: C4.5 Programs for Machine Learning. San Mateo: Morgan Kaufman (1993)
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1995 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Bischof, W.F., Caelli, T. (1995). Learning how to find patterns or objects in complex scenes. In: Braccini, C., DeFloriani, L., Vernazza, G. (eds) Image Analysis and Processing. ICIAP 1995. Lecture Notes in Computer Science, vol 974. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-60298-4_272
Download citation
DOI: https://doi.org/10.1007/3-540-60298-4_272
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-60298-9
Online ISBN: 978-3-540-44787-0
eBook Packages: Springer Book Archive