Title :
Historical Hand-Written String Recognition by Non-linear Discriminant Analysis using Kernel Feature Selection
Author :
Inoue, Ryo ; Nakayama, Hidehisa ; Kato, Nei
Author_Institution :
Graduate Sch. of Inf. Sci., Tohoku Univ., Sendai
Abstract :
In this paper, we propose a method to compose a classifier by non-linear discriminant analysis using kernel method combined with kernel feature selection for holistic recognition of historical hand-written string. Through experiments using historical hand-written string database HCD2, we show that our approach can obtain high recognition accuracy comparable to that of individual character recognition
Keywords :
feature extraction; handwritten character recognition; history; pattern classification; HCD2; historical handwritten string database; historical handwritten string recognition; kernel feature selection; nonlinear discriminant analysis; Character recognition; Feature extraction; Image recognition; Information analysis; Kernel; Noise shaping; Pattern recognition; Shape; Smoothing methods; Spatial databases;
Conference_Titel :
Pattern Recognition, 2006. ICPR 2006. 18th International Conference on
Conference_Location :
Hong Kong
Print_ISBN :
0-7695-2521-0
DOI :
10.1109/ICPR.2006.629
