Title :
On improvement of feature extraction algorithms for discriminative pattern classification
Author :
Gao, Jiang ; Ding, Xiaoqing
Author_Institution :
Dept. of Electron. Eng., Tsinghua Univ., Beijing, China
Abstract :
Two new feature extraction strategies-modified multiple discriminant analysis (MMDA) and difference principal component analysis (DPCA)-are presented and derived. The proposed algorithms are especially useful in automatic feature extraction from patterns in a small category set. Experimental results for recognition of Chinese character fonts and handwritten numerals using MMDA and DPCA are presented. Compared with the traditional algorithms, MMDA and DPCA provide more effective feature metrics for pattern discrimination in some settings
Keywords :
character recognition; covariance matrices; feature extraction; pattern classification; principal component analysis; Chinese character fonts; automatic feature extraction; difference principal component analysis; discriminative pattern classification; feature extraction algorithms; feature metrics; handwritten numerals; modified multiple discriminant analysis; Character recognition; Covariance matrix; Eigenvalues and eigenfunctions; Feature extraction; Handwriting recognition; Humans; Image processing; Pattern classification; Principal component analysis; Shape;
Conference_Titel :
Pattern Recognition, 2000. Proceedings. 15th International Conference on
Conference_Location :
Barcelona
Print_ISBN :
0-7695-0750-6
DOI :
10.1109/ICPR.2000.906026
