Abstract
A trainable recognition system intended for the detection of features in satellite imagery and for potential application to, for example, production line inspection has been constructed. A genetic search algorithm is used to find linear discriminant functions which will partition the pattern space and isolate the required features. The partitions are built up hierachically and represented as a classification tree. The training phase generates programs for the Linear Array Processor permitting subsequent images to be processed rapidly. It is shown that the system can generate a relaxation process to exploit contextual information.
This work is supported by a SERC CASE studentship in conjunction with the National Physical Laboratory, Teddington. Hilary Adams is now at Cambridge Electronic Design Ltd., Science Park, Cambridge CB4 4FE. Paper submitted for the 4th International Conference on Pattern Recognition, March, 1988.
Preview
Unable to display preview. Download preview PDF.
References
Curran P J, Principles of Remote Sensing, Longman (1985).
Plummer A P, The Linear Array Processor, The National Physical Laboratory (1986). DITC h 24.
Plummer P and Dale F, The Picture Processing Language Compiler, The National Physical Laboratory. (1984).
Devinjer P A and Kittler J, Pattern Recognition: A statistical Approach, Prentice/Hall Int. (1982).
Breiman L, Friedman J, Olshen R, and Stone C, Classification and regression trees., Wadsworth Int. Group (1984).
Quinlan J R, “Induction of Decision Trees”, Machine Learning 1 (1)pp. 81–106 (1986).
Poole I, Lapwing — A Prototype Image Recognition System for the Linear Array Processor, Dept. Computer Science, UCL (1987). Internal Note 2157.
Holland J H, “Genetic Algorithms and the Optimal Allocation of Trials”, Trials”, SIAM J. of Computing 2 (2) (1973).
Aleksander, I and Holland, J H, “Adaptation in natural and artificial systems.”, Pattern Recognition Letters 1pp. 724, University of Michigan Press. (1975).
Smith T R and DeJong K A, “Genetic algorithms applied to information driven models of US migration.”, 12th Annual Pittsburgh Conference on Modelling and Simulation. 3pp. 955–959 (1981).
Goldberg D E, “A tale of two problems: broad and efficient optimisation using genetic algorithms.”, Proceedings of the 1986 Summer Computer Simulation Conference, Reno., pp. 44–48 (1986).
Smith S F, “Flexible learning of problem solving heuristics through adaptive search.”, Proceedings 8th International Joint Conference on Artifial Intelligence., pp. 422–425 (1983).
Holland J H, “Escaping brittleness.”, Proceedings 2nd International Workshop on Machine Learning., pp. 92–95, University of Illinois (1983).
Lenat D B, “The role of heuristcs in learning by discovery.”, In Machine Learning, an artificial intelligence approach. Michalski R S, Carbonell J G & Mitchell T M., Tioga Press. (1983).
Grefenstette J J and Pettey C B, “Approaches to machine learning using genetic algorithms.”, IEEE International Conference on Systems, Man and Cybernetics 1pp. 55–60 (1986).
Adams H P, The Genetic Algorithm–A Search Strategy: Review, Implementation and Use., Dept Computer Science, University of York (1987). MSc thesis..
Davis L S and Rosenfield, “Co-operating Processes for Low-Level Vision: A Survey.”, Artificial Intelligence. 17pp. 245–263 (1981).
Kittler J and Illingworth J, “Relaxation labelling algorithms — A review”, Image and Vision Computing 3 (4)pp. 206–216 (1985).
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1988 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Poole, I., Adams, H. (1988). Lapwing - A trainable image recognition system for the linear array processor. In: Kittler, J. (eds) Pattern Recognition. PAR 1988. Lecture Notes in Computer Science, vol 301. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-19036-8_29
Download citation
DOI: https://doi.org/10.1007/3-540-19036-8_29
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-19036-3
Online ISBN: 978-3-540-38947-7
eBook Packages: Springer Book Archive