A rotation and location invariant face identification and localization with or without occlusion using modified RBFN

Author

Bhakta, Dhananjoy ; Sarker, Goutam

Author_Institution

Comput. Sci. & Eng. Dept., NIT Durgapur, Durgapur, India

fYear

2013

fDate

9-11 Dec. 2013

Firstpage

533

Lastpage

538

Abstract

This paper presents a new modified Radial Basis Function Network (RBFN) for identifying and localizing faces with or without occlusion for single images as well as for multiple image frame. The present method of facial identification is completely rotation and location invariant in the image frame. The technique of using the modified RBFN to perform learning of the different facial images and subsequent identification and location invariant localization of the clear, rotated and occluded faces is efficient, effective and fast. Also the identification rate of faces in single and multi-frame is quiet moderate.

Keywords

face recognition; learning (artificial intelligence); radial basis function networks; facial images; image frame; location invariant face identification; location invariant face localization; machine learning; modified RBFN; modified radial basis function network; multiple image frame; occlusion; rotation invariant face identification; rotation invariant face localization; Clustering algorithms; Conferences; Databases; Face; Neural networks; Training; Vectors; ANN; BP Networks; Face Identification; Face Localization; Identification Rate; Machine Learning; OCA; RBFN;

fLanguage

English

Publisher

ieee

Conference_Titel

Image Information Processing (ICIIP), 2013 IEEE Second International Conference on

Conference_Location

Shimla

Print_ISBN

978-1-4673-6099-9

Type

conf

DOI

10.1109/ICIIP.2013.6707649

Filename

6707649