Title :
Probabilistic two-dimensional canonical correlation analysis for face recognition
Author :
Afrabandpey, Homayun ; Safayani, Mehran ; Mirzaei, Abdolreza
Author_Institution :
Dept. of Electr. & Comput. Eng. (ECE), Isfahan Univ. of Technol. (IUT), Isfahan, Iran
Abstract :
Recently, two-dimensional canonical correlation analysis (2DCCA) proved to be an efficient technique for image feature extraction. In this paper we present a method of 2DCCA with probabilistic framework called probabilistic 2DCCA (P2DCCA), which is robust to noise and is able to cope with missing data problems. The experimental recognition results on three subsets of AR face database show the robustness of the proposed algorithm in face recognition in different illumination conditions, facial expressions and occlusion.
Keywords :
correlation methods; face recognition; feature extraction; probability; AR face database; P2DCCA; face recognition; facial expressions; illumination conditions; image feature extraction; missing data problems; occlusion; probabilistic 2DCCA method; probabilistic framework; probabilistic two-dimensional canonical correlation analysis; Data models; Face; Feature extraction; Noise; Principal component analysis; Probabilistic logic; Transforms; Canonical Correlation Analysis (CCA); Feature extraction; Two-dimensional analysis;
Conference_Titel :
Computer and Knowledge Engineering (ICCKE), 2014 4th International eConference on
Conference_Location :
Mashhad
Print_ISBN :
978-1-4799-5486-5
DOI :
10.1109/ICCKE.2014.6993337
