Chinese Character Recognition Method Based on Image Processing and Hidden Markov Model

Abstract

On the basis of image processing, the Chinese character recognition model of non specific people is researched based on hidden Markov model (HMM), and an improved isolated Chinese character recognition model is proposed based on HMM and scale invariant feature transform (SIFT) algorithm. The SIFT algorithm is used to extract the feature points of Chinese character, the SIFT feature point extraction algorithm for Chinese character is realized. The adaptive median filtering method and the gray histogram equalization method are used for the image preprocessing. The affine model is used to solve the strokes of a Chinese character moving parameters, and the Kalman filter is taken for the scanning filtering of the isolated Chinese character, the corresponding angular points of Chinese character image and the strokes feature are calculated. The frame window and pre-emphasis methods are used to extract the feature parameters. Finally, the HMM training is taken for the parameters. The template matching is implemented. The Chinese character recognition of uncertain people is achieved. Simulation result shows that the method can recognize the rare Chinese character image accurately, the recognition precision is improved greatly, and the image enhancement and noise reduction effect is obvious, the recognition model is stable with robustness. It has good application value in practice.