| Summary: | Moment-based Angular Radial Transform, Legendre moment invariants and Zernike moments are a family of orthogonal functions which allow the generation of non-redundant descriptors by the projection of an image onto an orthogonal basis. These descriptors can be used for classification, such as in face recognition. Zernike moments and Legendre moments have already been used for this purpose.This paper proposes to use moment-based Angular Radial Transform for extracting the face characteristics that feed a Support Vector Machine or a Nearest Neighbor Classifier for face recognition. Facial images from the ORL database, Essex Faces94 database, Essex Faces96 database, and Yale database were used for testing the proposed approach. The experimental results obtained show that the proposed method is more efficient, in terms of recognition rate, than the methods based on Zernike and Legendre moments. It is also found that its performance is comparable to that of the best state-of-the-arts methods. Keywords: Angular Radial Transform (ART), Legendre moment invariants (LMI), Euclidean distance (ED), Pseudo-Zernike moments (PZM), Nearest Neighbor Classifier (NNC), Support Vector Machines (SVM)
|
|---|
