Use of Region-Oriented Segmentation in Coin Recognition


Article Preview

Region-oriented segmentation is a simple relatively robust method for coin recognition. In this paper we present the use of Region-oriented segmentation for Coin Recognition. We use an improved K-Means Clustering Algorithm, which has the advantage to speed up the automatic determining of the optimal number of classes, to group all the gray-levels into several clusters. With the help of this cluster algorithm a label image of original coin image is obtained. In turn, the features such as area, perimeter, compactness and polar distance are extracted from the label image. The coins presented in the image could be recognized by matching the classifiers stored in the database. Several common segmentation approaches are also presented here in comparing to the region-oriented segmentation.



Key Engineering Materials (Volumes 277-279)

Edited by:

Kwang Hwa Chung, Yong Hyeon Shin, Sue-Nie Park, Hyun Sook Cho, Soon-Ae Yoo, Byung Joo Min, Hyo-Suk Lim and Kyung Hwa Yoo






M. Ma et al., "Use of Region-Oriented Segmentation in Coin Recognition", Key Engineering Materials, Vols. 277-279, pp. 312-317, 2005

Online since:

January 2005




In order to see related information, you need to Login.

In order to see related information, you need to Login.