Eigenfaces vs. Fisherfaces: Recognition Using Class Specific Linear Projection
IEEE Transactions on Pattern Analysis and Machine Intelligence
Recognizing faces with PCA and ICA
Computer Vision and Image Understanding - Special issue on Face recognition
Face recognition: A literature survey
ACM Computing Surveys (CSUR)
Face Recognition Using Laplacianfaces
IEEE Transactions on Pattern Analysis and Machine Intelligence
Journal of Cognitive Neuroscience
Enhanced independent component analysis and its application to content based face image retrieval
IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics
Face recognition by independent component analysis
IEEE Transactions on Neural Networks
Face recognition using kernel direct discriminant analysis algorithms
IEEE Transactions on Neural Networks
A structure-preserved local matching approach for face recognition
Pattern Recognition Letters
Content-based facial image retrieval using constrained independent component analysis
Information Sciences: an International Journal
Pattern Recognition Letters
| Hi-index | 0.00 |
We present a new dimensionality reduction method for face recognition, which is called independent component based neighborhood preserving analysis (IC-NPA). In this paper, NPA is firstly proposed which can keep the strong discriminating power of LDA while preserving the intrinsic geometry of the in-class data samples. As NPA depends on the second-order statistical structure between pixels in the face images, it cannot find the important information contained in the high-order relationships among the image pixels. Therefore, we propose IC-NPA method which combines ICA and NPA. In this method, NPA is performed on the reduced ICA subspace which is constructed by the statistically independent components of face images. IC-NPA can fully consider the statistical property of the input feature. Furthermore, it can find an embedding that preserves local information. In this way, IC-NPA shows more discriminating power than the traditional subspace methods when dealing with the variations resulting from changes in lighting, facial expression, and pose. The feasibility of the proposed method has been successfully tested on both frontal and pose-angled face recognition, using two data sets from the FERET database and the CAS-PEAL database, respectively. The experiment results indicate that the IC-NPA shows better performance than the popular method, such as the Eigenface method, the ICA method, the LDA-based method and the Laplacianface method.