Research report on Bengla OCR training and testing methods

Abstract

In this paper we present the training and recognition mechanism of a Hidden Markov Model (HMM) based multi-font Optical Character Recognition (OCR) system for Bengali character. In our approach, the central idea is to separate the HMM model for each segmented character or word. The system uses HTK toolkit for data preparation, model training and recognition. The Features of each trained character are calculated by applying the Discrete Cosine Transform (DCT) to each pixel value of the character image where the image is divided into several frames according to its size. The extracted features of each frame are used as discrete probability distributions which will be given as input parameters to each HMM model. In the case of recognition, a model for each separated character or word is built up using the same approach. This model is given to the HTK toolkit to perform the recognition using the Viterbi Decoding method. The experimental results show significant performance over models using neural network based training and recognition systems.