Cattle marks similarity measure by shape distributions

Abstract

This paper reports the investigation results addressed to design of a di­gital image processing method for automate the registration and control process of livestock marks according to Colombian regulations livestock sector. It allows automate the process of search and comparison required for livestock mark uniqueness assurance on a computer assisted system. First, a histogram is generated based on statistical distribution of geome­trical information, which permit to detect similarities between registered marks by means Minkowski-based metric. The results show that the method allows discriminating between images using a digital signature from a Shape Distribution the geometric structure reducing ambiguity and ensuring the uniqueness of registered marks. We show the results and make an analysis of system application