A Novel One-Parameter Regularized Linear Discriminant Analysis for Solving Small Sample Size Problem in Face Recognition

2004 ◽  
pp. 320-329 ◽  
Author(s):  
Wensheng Chen ◽  
Pong C Yuen ◽  
Jian Huang ◽  
Daoqing Dai
Keyword(s):  
Face Recognition ◽  
Discriminant Analysis ◽  
Sample Size ◽  
Linear Discriminant Analysis ◽  
Small Sample Size ◽  
Small Sample ◽  
Small Sample Size Problem ◽  
Linear Discriminant ◽  
Size Problem ◽  
Regularized Linear Discriminant Analysis

A NEW REGULARIZED LINEAR DISCRIMINANT ANALYSIS METHOD TO SOLVE SMALL SAMPLE SIZE PROBLEMS

2005 ◽  
Vol 19 (07) ◽  
pp. 917-935 ◽  
Author(s):  
WEN-SHENG CHEN ◽  
PONG C. YUEN ◽  
JIAN HUANG
Keyword(s):  
Face Recognition ◽  
Discriminant Analysis ◽  
Sample Size ◽  
Linear Discriminant Analysis ◽  
Small Sample Size ◽  
Small Sample ◽  
Scatter Matrix ◽  
Linear Discriminant ◽  
Matrix Formula ◽  
Parameter Regularization

This paper presents a new regularization technique to deal with the small sample size (S3) problem in linear discriminant analysis (LDA) based face recognition. Regularization on the within-class scatter matrix Sw has been shown to be a good direction for solving the S3 problem because the solution is found in full space instead of a subspace. The main limitation in regularization is that a very high computation is required to determine the optimal parameters. In view of this limitation, this paper re-defines the three-parameter regularization on the within-class scatter matrix [Formula: see text], which is suitable for parameter reduction. Based on the new definition of [Formula: see text], we derive a single parameter (t) explicit expression formula for determining the three parameters and develop a one-parameter regularization on the within-class scatter matrix. A simple and efficient method is developed to determine the value of t. It is also proven that the new regularized within-class scatter matrix [Formula: see text] approaches the original within-class scatter matrix Sw as the single parameter tends to zero. A novel one-parameter regularization linear discriminant analysis (1PRLDA) algorithm is then developed. The proposed 1PRLDA method for face recognition has been evaluated with two public available databases, namely ORL and FERET databases. The average recognition accuracies of 50 runs for ORL and FERET databases are 96.65% and 94.00%, respectively. Comparing with existing LDA-based methods in solving the S3 problem, the proposed 1PRLDA method gives the best performance.

scholarly journals Simulasi Ektraksi Citra Wajah dengan Deskriptor Global Direct Fractional-Step Linear Discriminant Analysis (DF-LDA) untuk Pengenalan Wajah

2015 ◽  
Vol 14 (01) ◽  
pp. 59-66
Author(s):  
Ikuthen Gabriel Barus ◽  
Riko Arlando Saragih
Keyword(s):  
Face Recognition ◽  
Discriminant Analysis ◽  
Sample Size ◽  
Linear Discriminant Analysis ◽  
Weighting Function ◽  
Small Sample Size ◽  
Small Sample ◽  
Fractional Step ◽  
Linear Discriminant

Tulisan ini memaparkan simulasi ekstraksi citra wajah secara global dengan menggunakan salah satu teknik Linear Discriminant Analysis (LDA), yaitu Direct Fractional-Step LDA (DF-LDA) untuk pengenalan wajah. Tujuan tulisan ini adalah untuk mengevaluasi unjuk kerja teknik ini terhadap masalah small sample size (SSS) yang sering muncul di dalam pengenalan wajah. Pada dasarnya teknik berbasis LDA ini (DF-LDA) merupakan kombinasi dari teknik D-LDA dan F-LDA, dimana untuk merepresentasikan citra wajah secara global secara efisien dapat ditambahkan sebuah fungsi pembobot (weighting function) dengan bertahap secara langsung dan fraksional pada proses LDA. Proses pencocokan dilakukan dengan mencari jarak Euclidean minimum antara ciri citra wajah uji terhadap ciri citra wajah latih yang terdapat di dalam database. Dari hasil simulasi untuk Database Face Recognition Data dan Database Mahasiswa Maranatha diperoleh akurasi pengenalan wajah yang lebih baik untuk kondisi jumlah citra wajah satu per orang di dalam proses pelatihan jika database wajah diproses secara terpisah.

Sign in / Sign up

Export Citation Format

Share Document