人脸识别论文：TR_pami_llda.pdfVIP

Locally Linear Discriminant Analysis for Multi-modally Distributed Classes For Face Recognition with a Single Model Image Tae-Kyun Kim , Josef Kittler : Department of Engineering, University of Cambridge, Cambridge, CB2 1PZ, UK. : Centre for Vision, Speech and Signal Processing, University of Surrey, Guildford, GU2 7XH, UK. This was published in IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol.27, No.3, pp.318–327, 2005. This can be best viewed in color printing or electronic form. 1 IEEE Trans. on Pattern Analysis and Machine Intelligence, Vol.27, No.3, 2005 Locally Linear Discriminant Analysis for Multi-modally Distributed Classes For Face Recognition with a Single Model Image Tae-Kyun Kim and Josef Kittler : Department of Engineering, University of Cambridge, Cambridge, CB2 1PZ, UK. : Centre for Vision, Speech and Signal Processing, University of Surrey, Guild- ford, GU2 7XH, UK. Abstract– We present a novel method of learning a set of locally linear trans- formations called ”Locally Linear Discriminant Analysis (LLDA)” for nonlinear classification problems. The underlying idea is that global nonlinear data struc- tures are locally linear and local structures can be linearly aligned. Input vectors are projected into each local feature space by linear transformations found to yield locally linearly transformed classes that maximize the between-class covariance while minimizing the within-class covariance in the aligned output space. This maximizes the separability of classes locally while promoting consistency between the multiple local representations of single class objects. In face recogniti