Vehicular shape-based objects classification using Fourierdescriptor technique

Abstract

This paper reports classification of vehicular shape-based objects using Fourier descriptor (FD) technique. Retrieval andclassification of different types of shapes was done for closest match, comparing set of feature vectors of query object to eachobject of every class. Centroid distance based shape signatures were used for derivation of feature vectors. Euclidean distance wascalculated as a similarity measure parameter for shape classification. Classification of noisy objects was carried out using FDtechnique. FD performs better than wavelet descriptor (WD) technique.