Abstract:To efficiently utilize the discriminant information in the range space of locality preserving total scatter, this paper proposes a complete discriminant locality preserving projections (CDLPP) algorithm for face recognition.Since Fisher discriminant analysis and locality preserving projections (LPP) have been widely used in face recognition, CDLPP algorithm integrates them together and analyzes the discriminant information contained in the principal spaces and null spaces of locality preserving within-class scatter, locality preserving between-class scatter and locality preserving total scatter.First, CDLPP algorithm removes the null space of locality preserving total scatter, in which no discriminant information is contained, using singular value decomposition (SVD).Then, regular discriminant features and irregular discriminant features of CDLPP are extracted severally in the null space and principal space of the locality preserving within-class scatter.Finally, both regular discriminant features and irregular discriminant features are concatenated to be used for face recognition.Extensive experiments on ORL face database, FERET subset and PIE subset illustrate that the performances of CDLPP outperform those of current subspace face recognition algorithms, such as LDA, LPP and discriminant LPP, which proves the effectiveness of the proposed algorithm.

Orthogonal complete discriminant locality preserving projections of kernel

A Kernel Orthogonal Discriminant Locality Preserving Projections Method

Face recognition with discriminant locality preserving projections in complete kernel space

Kernel Orthogonal Unsupervised Discriminant Projection with Applications to Face Recognition

An Orthogonal Discriminant Locality Preserve Projections with Schur Decomposition

Kernel Orthogonal and Uncorrelated Neighborhood Preservation Discriminant Embedding Algorithm

Kernel Based Orthogonal Locality Preserving Projections for Face Recognition

Supervised Kernel Optimized Locality Preserving Projection with Its Application to Face Recognition and Palm Biometrics

Improved kernel direct Fisher discriminant analysis and face recognition

Kernel locality preserving discriminant analysis for face recognition

Complete Discriminant Locality Preserving Projections for Face Recognition

Local Feature Discriminant Projection

A Recognition Method Using Uncorrelated Discriminant Locality Preserving Projections

Kernel optimization-based discriminant analysis for face recognition

Discriminant Locality Preserving Projections Based on Neighborhood Maximum Margin

Kernel Neighborhood Preserving Projections for Face Recognition

Optimal Locality Preserving Projection for Face Recognition.

Direct Orthogonal Neighborhood Preserving Discriminant Analysis

Orthogonal Locality Minimizing Globality Maximizing Projections for Feature Extraction

Discriminant Local Median Preserving Projections with its Application to Face Recognition

An alternative formulation of kernel LPP with application to image recognition