SLP Header

Novel Feature Extraction for Face Recognition using Multiscale Principal Component Analysis

IJCSEC Front Page

Abstract
A method of face recognition based on multiscale principal component analysis (MSPCA) is presented in this paper. Initially face area is extracted from the given face image using Adaboost face detection algorithm. From the face area, regions of interest such as eyes, nose and mouth part are extracted by dividing it along horizontal and vertical directions. Then MSPCA is employed on these regions of interest to extract the features. Multiscale Principal Component Analysis (MSPCA) combines the ability of PCA to decor relate the variables by extracting a linear relationship with that of wavelet analysis to extract deterministic features and approximately decor relate the auto correlated measurements. MSPCA computes the principal component analysis (PCA) of the wavelet coefficients at each scale, followed by combining the results at relevant scales. K-Nearest Neighbor (k-NN) classifier is used for recognition. The proposed methodology exhibits better recognition rate when compared to conventional principal component analysis.
Keyword:Face Recognition, Feature extraction, PCA, MSPCA, k-NN Classifier.
I.Introduction
Recently face recognition is attracting much attention in society of network multimedia information access. Areas such as network security, content indexing and retrieval, and video compression benefits from face recognition technology. Network access control via face recognition not only makes hackers virtually impossible to steal one’s “password”, but also increases the user-friendliness in human-computer interaction. Indexing and/or retrieving video data based on the appearances of particular persons will be useful for users such as news reporters and moviegoers. For the applications of videophone and teleconferencing, the assistance of face recognition also provides a more efficient coding scheme. A good survey on face recognition is found in [1]. Face recognition methods can be roughly divided into two different groups: geometrical features matching and template matching. In the first method, some geometrical measures about distinctive facial features such as eyes, mouth, nose and chin are extracted [1]. In the second method the face image is represented as a two-dimensional array of intensity values and is compared to a single or several templates representing a whole face. The earliest methods for template matching are correlation-based which are computationally very expensive and require great amount of storage. Since a few years, the Principal Components Analysis (PCA) method also known as Karhunen-Loeve method, is successfully used as feature extraction technique and also used to perform dimensionality reduction [2, 3, 4, 5]. The problem with PCA is high computation complexity.
In this paper, we propose a new method for face recognition based on multiscale principal component analysis (MSPCA). From the input face image face area is extracted using Adaboost face detection algorithm. Regions of interest (ROI) such as eyes, nose and mouth part are extracted by dividing the detected face area along horizontal and vertical directions. Features are extracted by employing MSPCA on these regions. Then k-NN classifier is used for classification by considering different values for k. Experimental results are presented for ORL, Grimace and Faces94 databases and demonstrated the efficiency of the proposed method.

References:

  1. W. Zhao, R. Chellappa, A. Rosenfeld, P.J. Phillips, “Face Recognition: A Literature Survey”, ACM Computing Surveys, pp 399- 458, 2003.
  2. M. Kirby, L. Sirovich, “Application of the Karhunen-Loeve Procedure and the Characterization of Human Faces”, IEEE Transactions on Pattern Analysis and Machine Intelligence, 12(1):103–108, 1990.
  3. A. Pentland, R.W. Picard, S. Sclaroff., “Photobook:Content-Based Manipulation of Image Databases”,In Proceedings of the SPIE Storage and Retrievaland Video Databases II, No. 2185, San Jose, 1994.
  4. M. Turk, A. Pentland.,”Eigenfaces for Recognition”, Journal of Cognitive Science, 3(1):71–86, 1991.
  5. C. L.Wilson, C. S. Barnes, R. Chellappa, S. A. Sirohey, “Face Recognition Technology for Law Enforcement Applications”, NISTIR 5465, U.S. Department of Commerce, 1994.
  6. Bakshi, B, "Multiscale PCA with application to Multivariate Statistical Process Monitoring", AIChE Journal, 44, pp. 1596-1610,1998.
  7. Aminghafari, M.; Cheze, N.; Poggi, J-M. (2006), "Multivariate de-noising using wavelets and principal component analysis", Computational Statistics & Data Analysis, 50, pp. 2381-2398, 2006.
  8. T.M. Cover and P.E. Hart, “Nearest neighbor pattern classification”, IEEE transaction in information theory, Vol 13, pp. 21-27, Jan 1967.
  9. Yu-Long Qiao; Jeng-Shyang Pan; Sheng- He Sun, “Improved K nearest neighbor classification algorithm”, Circuits and Systems, 2004, Proceedings. IEEE Asia-Pacific Conference, vol.2, pp. 1101- 1104, Dec. 2004
  10. K.T.Talele, Sunil Kadam and Atul Tikare. “ Efficient Face Detection using Adaboost”. IJCA Proceedings on International Conference in Computational Intelligence (ICCIA2012) iccia (10), March 2012.
  11. http://www.cl.cam.ac.uk/Research/DTG/attarchive:pub/data/att_- faces.zip.
  12. http://cswww.essex.ac.uk/mv/allfaces/grimace.zip.
  13. http://cswww.essex.ac.uk/mv/allfaces/faces94.html.