Title :
Maximization of mutual information for offline Thai handwriting recognition
Author :
Nopsuwanchai, R. ; Biem, A. ; Clocksin, W.F.
Author_Institution :
Inf. Technol. Lab., Kasei Corp., Atsugi
Abstract :
This paper aims to improve the performance of an HMM-based offline Thai handwriting recognition system through discriminative training and the use of fine-tuned feature extraction methods. The discriminative training is implemented by maximizing the mutual information between the data and their classes. The feature extraction is based on our proposed block-based PCA and composite images, shown to be better at discriminating Thai confusable characters. We demonstrate significant improvements in recognition accuracies compared to the classifiers that are not discriminatively optimized
Keywords :
feature extraction; handwritten character recognition; hidden Markov models; learning (artificial intelligence); optimisation; principal component analysis; HMM; Thai confusable characters; block-based PCA; composite images; discriminative training; fine-tuned feature extraction methods; hidden Markov models; mutual information maximization; offline Thai handwriting recognition; principal component analysis; Character recognition; Clocks; Feature extraction; Handwriting recognition; Hidden Markov models; Maximum likelihood estimation; Mutual information; Optical character recognition software; Optimization methods; Principal component analysis; Character recognition; Hidden Markov Model; PCA; Thai handwriting recognition.; discriminative training; feature extraction; Algorithms; Artificial Intelligence; Automatic Data Processing; Computer Simulation; Documentation; Handwriting; Image Enhancement; Image Interpretation, Computer-Assisted; Information Storage and Retrieval; Likelihood Functions; Models, Statistical; Online Systems; Pattern Recognition, Automated; Thailand;
Journal_Title :
Pattern Analysis and Machine Intelligence, IEEE Transactions on
DOI :
10.1109/TPAMI.2006.167
