Face recognition using discriminant locality preserving projections

Kernel Fisher Discriminant Analysis with Locality Preserving for Feature Extraction and Recognition

“…KDA and its variations [23][24][25][26][27][28][29][30] only consider global geometrical structure and neglect local geometrical structure. In this section, we will develop a new KDA framework which can incorporate the local geometrical structure of data samples.…”

“…All such methods attempt to embed the original data into a submanifold by preserving the local geometrical structure. Different from LLE, Isomap and Laplacian eigenmap, LPP is a linear algorithm which is quite simple and easy to realize, thus has received much attention in the research community [23][24][25][26][27][28][29][30]. He et al [23] applied LPP on the face recognition and demonstrated the effectiveness of LPP in exploring the local geometrical structure of the data.…”

“…He et al [23] applied LPP on the face recognition and demonstrated the effectiveness of LPP in exploring the local geometrical structure of the data. In [24], a discriminant locality preserving projection (DLPP) algorithm was proposed to improve the classification performance of LPP. Yang et al [25] developed an unsupervised discriminant projection (UDP) technique for dimensionality reduction.…”

Face Recognition Based on Multi-classifierWeighted Optimization and Sparse Representation

“…For example, on some special occasions such as passport verification, law enforcement and ID authentication, one can only get very few training samples, and sometimes even only one single sample, thus, the number of dimensions of image samples is much larger than the number of training samples. This will result in the failure of the distance measurement in pattern recognition, thus quickly decreasing or depriving of the recognition performance of traditional PCA, LDA, LPP face recognition methods and their derivative ones [3].…”

Dual Two-Dimensional Fuzzy Class Preserving Projections for Facial Expression Recognition

“…LPP aims to find an embedding that preserves local information, and obtains a face subspace that best detects the essential face manifold structure. There have been various improvements made to LPP: Orthogonal LPP [3], discriminant LPP [4], Supervised kernel LPP [5] etc. Although LPP is widely used in many domains, it suffers form the singular problem in the high dimensional image space, which makes the direct implementation of the LPP algorithm impossible.…”