We study local similarity based distance measures for point-patterns. Such measures can be used for matching point-patterns under non-uniform transformations — a problem that naturally arises in image comparison problems. A general framework for the matching problem is introduced. We show that some of the most obvious instances of this framework lead to NP–hard optimization problems and are not ap-proximable within any constant factor. We also give a relaxation of the framework that is solvable in polynomial time and works well in practice in our experiments with two–dimensional protein electrophoresis gel images.
CITATION STYLE
Mäkinen, V., & Ukkonen, E. (2002). Local similarity based point-pattern matching. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 2373, pp. 115–132). Springer Verlag. https://doi.org/10.1007/3-540-45452-7_11
Mendeley helps you to discover research relevant for your work.