Construction and Experiment Analysis of Uygur Handwriting Recognition Database

Uygur is one of the official languages of Xinjiang Uygur Autonomous Region. At present, the development of its handwriting technology still lags behind, and the text input is still in the keyboard coding stage. Based on the handwriting technologies for Chinese, English and Arabic, by combining the structure, word-formation characteristics and input habits of Uygur, this paper studies the construction of basic large-scale Uygur handwriting recognition sample database, the processing of samples, the extraction and classification technology of features and the organization form of database, and analyzes the influence of training sample number on the recognition performance through experiments.