New Variants of Global-Local Partial Least Squares Discriminant Analysis for Appearance-Based Face Recognition

Abstract

We propose new appearance-based face recognition methods based on global-local structurepreserving partial least squares discriminant analysis. Two variants of the method are described in this paper: the neighbourhood-preserving partial least squares discriminant analysis (NPPLS-DA) and the uncorrelated NPPLS-DA (UNNPPLS-DA). In contrast to standard partial least squares discriminant analysis (PLS-DA), which effectively only recognizes the global Euclidean structure of the face space, both NPPLS-DA and UNNPPLS-DA are designed to find an embedding that preserves both the global and local neighbourhood information and obtain a face subspace that best detects the essential manifold structure of the face space. Unlike global-local features extracted using other methods, the global-local PLS-DA features are obtained by maximizing covariance between data matrix and a response matrix which is coded with the class structure of the data. Furthermore, in UNPPLS-DA, an uncorrelated constraint is introduced into the objective function of NPPLS-DA to extract uncorrelated features that are important in many pattern recognition problems. We compare the proposed NPPLS-DA and UNPPLS-DA methods with several competing methods on six different face databases. The experimental results show that the proposed NPPLS-DA and UNPPLS-DA methods provide better representation and consistently achieve higher recognition rates in face recognition than the other competing methods.

Parameter-free nonlinear partial least squares regression model for image classification

“…In the classification task, images are treated as pixel vectors and represented as individual points in high dimensions space. The complexity of classification and identification increases with the increase of image dimensionality in the dataset 1 . Dimensionality reduction and feature extraction techniques could improve the accuracy of image classification and the efficiency of machine learning algorithms.…”

“…Although PLSR models have achieved some success in distinguishing and preserving linear structure, 1 the performance of linear PLSR models generally degrades when faced with data with nonlinear structure 16 . Two main types of nonlinear PLSR models have been proposed recently.…”

“…The complexity of classification and identification increases with the increase of image dimensionality in the dataset. 1 Dimensionality reduction and feature extraction techniques could improve the accuracy of image classification and the efficiency of machine learning algorithms. Regression models are generally used for dimension reduction and feature extraction.…”