Title :
A new feature extraction method based on Fourier transform in handwriting digits recognition
Author :
Liu, Gang ; Wu, Di ; Zhang, Hong-Gang ; Guo, Jun
Author_Institution :
Sch. of Inf. Eng., Beijing Univ. of Posts & Telecommun., China
Abstract :
Fourier transform (FT) is a kind of efficient feature extraction method. In this paper, we mainly discuss the most frequently used feature extraction methods based on FT in the field of handwriting digits recognition and propose a new method, called the one dimensional FT feature extraction method based on sum and difference. We have implemented this method to off-line recognition of handwriting digits, and it has been proved by experiments that this method is effective.
Keywords :
Fourier transforms; feature extraction; handwritten character recognition; hidden Markov models; Fourier transform; feature extraction; filter bank analysis; handwritten digit recognition; hidden Markov model; sum difference method; Feature extraction; Fourier transforms; Handwriting recognition; Hidden Markov models; Karhunen-Loeve transforms; Lagrangian functions; Optical character recognition software; Pattern recognition; Telecommunications; Wavelet transforms;
Conference_Titel :
Machine Learning and Cybernetics, 2002. Proceedings. 2002 International Conference on
Print_ISBN :
0-7803-7508-4
DOI :
10.1109/ICMLC.2002.1167396
