Similarity between polygonal shapes

Guorong Fang, University of Nevada, Las Vegas


In this thesis we review existing algorithms for measuring shape similarity between polygons. We present a new approach to measure similarity based on the notion of annular profile. We also present the implementation of three shape measuring algorithms: signature function, turning function, and annular profile. The implementation is done by using the Visual C++ programming language. Finally, we discuss the comparative performances of the above three methods for capturing shape similarity. Measurement of shape similarity has applications in pattern recognition and artificial intelligence.