An Analysis On The Algorithm Handwritten Chinese Character Recognition And System Realization

Handwritten Chinese character recognition is one of the hotspots in image pattern recognition,which includes many key technologies such as-image preprocessing,feature extraction and recognition classification design,Which is the most critical part of Chinese character extraction.The effective feature extraction method can quickly and accurately identify the overall characteristics of Chinese characters.The specific work of this paper is as follows:1.The pre-treatment of Chinese character.This paper introduces several commonly used methods of Chinese character image preprocessing,including gray scale,smoothing and drying,binarization,Chinese character segmentation and acquisition of Chinese character contour image.Preconditioning of Chinese character images makes it easier to recognize Chinese characters.2.Feature extraction of handwritten Chinese characters.This paper focuses on several commonly used Chinese character extraction methods.For different Chinese characters have different contour features,this paper uses Fourier descriptors to extract the characteristics of Chinese characters.Firstly,the boundary of the Chinese character contour is obtained,and then the boundary points are separated from each other.Then,the boundary points are normalized,and the discrete Fourier transform(DFT)is made to the normalized points.The Fourier transform results are normalized to obtain a descriptor with the rotation,translation and scale invariance,ie the Fourier descriptor.When using the low frequency component of DFT,this method can effectively suppress high frequency clutter,enhance robustness,and play the characteristic of feature compression,and it is insensitive to the rotation,translation and size of Chinese characters.3.Classification of handwritten Chinese characters.This paper mainly introduces the four classification methods of hidden Markov model,nearest neighbor(kNN)classification,adaptive enhancement algorithm(adaboost)and artificial neural network classification.This paper analyzes the principle and characteristics of the four classifiers and draws lessons from the previous experience,and finally selects the BP neural network as the classifier to realize the handwritten Chinese character recognition.4.The system realization of handwritten Chinese characters and the analysis of Its experimental results.Using MATLAB to design human-computer interaction interface,to achieve the recognition of handwritten Chinese characters,and statistical results.