Invariant Pattern Recognition - An Overview

Dr R Mukundan

Dept. of CSSE, University of Canterbury

Fri Sep 15 15:10:00 NZST 2006 in Room 031, MSCS

Abstract

Moment invariants are feature descriptors that can be used for shape classification and pattern recognition. While translation and scale invariants can be obtained without much computational cost, rotation invariants often require complex functions and are more susceptible to measurement errors. This talk introduces various types of invariants used in pattern recognition, and discusses some recent developments in the field.