Abstract
Handwriting identification is one of the most important branches of computer vision and pattern recognition, and an effective means of identification that is important for surveillance and security, telecommunications, digital libraries, video meeting, and human-computer intelligent interactions. A deep research on pre-processing of Chinese handwriting identification is conducted in this paper. Computer can automatically find the threshold by using a new iteration algorithm in binarization, as it can retain more texture features of image. In noise eliminating, we also use an improved median filter to automatically determine the noise threshold, which is effective to get rid of the scattered noise. After experimentally repeated for many times in Matlab7.0, it’s proved that this preprocess is vary fast and stable.