Structural pattern recognition of biological textures with growing deformations: A case of cattle's muzzle patterns

Abstract

AbstractTextured patterns on living organisms (hereafter referred to as biological textures), which are typified by human fingerprints or cattle muzzle patterns, have features that differ according to the individual organism, and it is well known that these biological textures can be used for identification. However, the pattern structure of cattle muzzle patterns, for example, is more complex than that of human fingerprints, and since the structure features are changed or deformed during the growing stage, these pattern structures cannot be skillfully recognized by using a technique like the one used for conventional fingerprint comparison. In this paper, the authors propose a new recognition technique that is applicable even to biological textures with these kinds of growing deformations. Specifically, they focus on cattle muzzle patterns as an example of a biological texture and select graph matching based on a horizontal search as the basic strategy to create this new recognition technique consisting of a technique for determining a pair of Hough‐transformed search‐starting cycles and a muzzle pattern recognition method that improves search efficiency by incorporating processing for merging non‐searched cycles. The authors performed evaluation experiments to verify the effectiveness of the proposed technique and obtained good results. © 2004 Wiley Periodicals, Inc. Electron Comm Jpn Pt 2, 87(5): 54–66, 2004; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjb.20076