The Method Of Online Handwritten Chinese Character Recognition For IOS And Its Realization

With the rise of mobile internet, smart phones have become a part of daily necessities. Meanwhile, people have changed their interactive habits with mobile equipment from relying on the traditional keyboard to the new touch Screen. Consequently, handwriting input method, as a tool of people-equipment interaction, is becoming more and more important, and the percentage of population using it is also going up year by year. In terms of handwriting input method, text recognition can identify more characters at a time compared with single-character handwriting recognition, which may make more natural and smoother interaction experiences. Apple opened its handwriting method API in 2014, which had made the use of third-party input methods in iOS possible. Therefore, designing a handwriting input software for iOS equipment is valuable.Based on the above, this paper puts forward an online written Chinese character recognition method for iOS, and conducts further study on the realization of it. The main work and contributions are as follows:1.Research on unconstrained online handwritten Chinese character recognition algorithm, during which the linguistic model, confidence coefficient conversion, geometric model, routing evaluation and routing search algorithm of the recognition algorithm and its application in iOS input methods are further studied;2.Deeply study the input method framework and realization mechanism in iOS and creatively design a SCUT gPen handwriting input method App, which has the functions including single-character recognition, text recognition and overlapping handwriting recognition, and provides handwriting keypad, numeric keypad, symbolic keypad, English keypad, emoticon keyboard and emoji emoticons keypad according to the different needs of input. At the same time, auxiliary functions like appearance changing, simplified-style and complex-font Chinese characters switching and handwriting beautification have been added to the software. So far, the software has been put into use in App Store, and its accumulative downloads have surpassed 50,000 with being downloaded hundreds of times every day;3.Considering the timing sequence of associative phrases, this input software employs the algorithm of Long Short Term Memory Recurrent Neural Networks to realize the dynamic generation of associative phrases in line with users’ former input, instead of searching for the associative phrases based on fixed phrase data inefficiently.