Design And Implementation Of Chinese Handwritten Input Method Based On ARM9and Qt

Abstract:As there is no handwritten input method on the embedded platform based on ARM9and Qt, and the current handwritten Chinese character recognition core has large resource occupies, slow recognition speed, development inconvenience and other shortcomings when applied to the platform, the development of the Chinese handwritten input method based on ARM9and Qt platform has a broad application prospects. Besides, explore a possible ways to realize simple and efficient large scale handwritten Chinese character recognition on resource constrained embedded platform also has profound significance in theory.This thesis focuses on the preprocessing, feature extraction and classifier design problems in online handwritten Chinese character recognition on the embedded platform based on ARM9and Qt. Because there is no floating point calculation hardware on ARM9processor, and generally, the embedded platform based on ARM9and Qt has small memory, so the amount of calculating and resource consumption must be considering when designed the recognition characteristics and classification. This thesis, which enterprise project as the background, researched the current handwritten Chinese character recognition technology and Qt input method framework, based the features of the ARM9and Qt embedded platform, designed and implemented a Chinese handwriting input method on the platform. Specifically, the main work of this thesis is as follows: ①Researched the basic steps, the metrics, the principle of the recognition of handwritten Chinese characters, and the input method on embedded Qt platform.②Build a cross development environment; transplant the bottom software, including the BootLoader, the Linux kernel, the root file system as well as the Qt/Embedded graphics library.③Researched the input method used in PC and telephone, designed the system function, the UI, and the software architecture.④) Researched the mainstream handwritten Chinese character recognition technology, designed a simple and efficient algorithm for handwritten Chinese character recognition. The algorithm preprocessed the user’s writing points, combined Chinese characters’structural and statistical features, used the serial and parallel structure of classifier, and consolidated the results finally.⑤Realized the phrase association function, created the desired phrase library, and no restriction on association phrase length.⑥Using varieties of optimization techniques to achieve a practical input method software on ARM9and Qt platform. The software has lower resources consumption, highly scalable, good portability, and can meet the daily input requirements on ARM9and Qt platform.