Mobile Device Oriented Chinese Handwritten Character Recognition System

With the rapid development of the Internet, especially the rapid quick booming ofmobile devices, mobile devices have become part of daily. tablet computer and smartphones with full-size touch screen make Chinese character handwritten recognitionmore convenient. For Chinese character input, the development of mobile devices inhandwritten recognition brings commercial value at the same time, with greatlychallenges, we need our recognition to be rapid and accurate. Let the users to accept,make the user friendly.The main work is to complete mobile equipment handwritten Chinese characterrecognition system and different equipment have indiscriminate：(1) The realization of the system with C-MQDFBecause the memory space of mobile devices is limited, so we need a kind of lowconsumption of Chinese classifier, which will not makes mobile platform identificationappear crash or a lot of recognition problems. The new classifier method reduces thespace complexity, but the recognition rate will have little effect. In three handwrittenlibrary evaluation, we can obviously reduce the space complexity, and at the same timeit reduce the accuracy of the small.(2) suitability analysis of handwritten Chinese characters mobile equipmentFirstly, in order to complete mobile equipment of the handwriting recognitionalgorithm, we use two kinds of language (C++and JAVA) to implement thehandwriting recognition algorithm and provide the interface for identification. Forrecognition time evaluation, a little progress of handwriting recognition time is veryvaluable. Secondly, in different equipments, the same writing can cause identificationdifferent accuracy, then the chi-square test method is used to measure differentequipments validation, and experiments show different equipments opponenthandwriting recognition is discriminating. At last, different equipments are differentopponent handwriting identification which is the fundamental cause of the equipment ofdifferent mining point number is different. Mining point position is also different. Thesame person for the different words implements a large number of collection, and indifferent equipments, calculation of mining points expectations and evaluation for therecognition rate of the influence are very big.