Eigenfaces vs. Fisherfaces: Recognition Using Class Specific Linear Projection
IEEE Transactions on Pattern Analysis and Machine Intelligence
Face Recognition by Elastic Bunch Graph Matching
IEEE Transactions on Pattern Analysis and Machine Intelligence
IEEE Transactions on Pattern Analysis and Machine Intelligence
IEEE Transactions on Pattern Analysis and Machine Intelligence
IEEE Transactions on Pattern Analysis and Machine Intelligence
Face recognition with one training image per person
Pattern Recognition Letters
Journal of Cognitive Neuroscience
Face recognition from a single image per person: A survey
Pattern Recognition
On solving the face recognition problem with one training sample per subject
Pattern Recognition
Selecting discriminant eigenfaces for face recognition
Pattern Recognition Letters
Robust face recognition from one training sample per person
ICNC'05 Proceedings of the First international conference on Advances in Natural Computation - Volume Part I
Recent advances in subspace analysis for face recognition
SINOBIOMETRICS'04 Proceedings of the 5th Chinese conference on Advances in Biometric Person Authentication
Survey: Subspace methods for face recognition
Computer Science Review
Adaptive discriminant learning for face recognition
Pattern Recognition
Improving fusion with optimal weight selection in Face Recognition
Integrated Computer-Aided Engineering
Hi-index | 0.00 |
Many face recognition algorithms/systems have been developed in thelast decade and excellent performances are also reported when thereis sufficient number of representative training samples. In manyreal-life applications, only one training sample is available.Under this situation, the performance of existing algorithms willbe degraded dramatically or the formulation is incorrect, which inturn, the algorithm cannot be implemented. In this paper, wepropose a component-based linear discriminant analysis (LDA) methodto solve the one training sample problem. The basic idea of theproposed method is to construct local facial feature componentbunches by moving each local feature region in four directions. Inthis way, we not only generate more samples, but also consider theface detection localization error while training. After that, weemploy a sub-space LDA method, which is tailor-made for smallnumber of training samples, for the local feature projection tomaximize the discrimination power. Finally, combining thecontributions of each local feature draws the recognition decision.FERET database is used for evaluating the proposed method andresults are encouraging.