Boundary and Region based Moments Analysis for Image Pattern Recognition

Abstract

In a number of pattern recognition application, moments have been used to recognize image patterns. The recognition process involves effective shape representation method. So, the present paper analyzes the recognition rate in two different shape representation methods called external and internal representation. With the proper shape representation of the given image pattern, the present paper has computed 7 boundary (external) based and 10 region (internal) based Hu moments. The experimental results on five different pattern image groups (Brick, Circle, Curve, Line and Zigzag) are precisely recognized by both boundary based and region based moments of order 1 and 10 respectively.