Abstract
The suggested model for image quantization transforms the images (the image space) in low gray levels or quantized values. It use the class membership of the images. The model is based on multi-level thresholding or irregular quantization for each pixel. The problems stated are reduced to solving integer-valued systems of equations and inequalities as well as integer-valued optimization problems. These problems are NP-complete. From the condition that the image classes are non-intersecting it follows that the minimum quantization of the images exists. Possible algorithms for solving the suggested problems are briefly discussed.
This is a preview of subscription content, log in via an institution to check access.
Preview
Unable to display preview. Download preview PDF.
Similar content being viewed by others
References
V.Valev, Ju.I.Zhuravlev: Integer-valued problems of transforming the training tables in k-valued code in pattern recognition problems. Pattern Recognition 24, 283–288 (1991)
V.Valev: Pictorial pattern recognition by Boolean formulas. In: R. Klette, W. G. Kropatsch (eds.): Proc. of 5th Workshop on Theoretical Foundation of Computer Vision, Germany, Buckow, March 29–April 3, 1992. Mathematical Research, Vol. 69 Akademie Verlag 1992, pp.241–250
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1993 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Valev, V. (1993). A model-based image quantization technique for supervised image recognition. In: Chetverikov, D., Kropatsch, W.G. (eds) Computer Analysis of Images and Patterns. CAIP 1993. Lecture Notes in Computer Science, vol 719. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-57233-3_17
Download citation
DOI: https://doi.org/10.1007/3-540-57233-3_17
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-57233-6
Online ISBN: 978-3-540-47980-2
eBook Packages: Springer Book Archive