Key Technology Research And Application Of Multi-character Recognition For Handwriting Input

With the development of technology, handwriting recognition technology has under grown great development. The achievement of computer intelligence and automation has been not only applied on computer related products but also on the field of cell phone and camera. The technology has developed from isolated single character recognition to chapter recognition, whose core module is the recognition of text, which is multi-character recognition. Because of this technology the automation of computer and cell phone has been greatly promoted. Base on the study of multi-character handwriting technology, it has also done deep research on the methods about multi-character handwriting, to find a better product which is easier for operation for customers. No matter in the field of computer or other electronic fields, multi-character handwriting technique has shown us the demand on diversification and intelligence. Furthermore, it also will be developed into comprehensive application, during which other applications and technique research will also be studied, among which it mainly contains as below:This paper mainly does comprehensive explore and study on the need of multi-character handwriting recognition, especially does design and realization of comprehensive multi-character system on Chinese, English and numbers; During the design of multi-character handwriting system, it also emphasized the verification of system’s difficulties, like character segmentation and single-character recognition; It verifies normalization on collected data using multi-character handwriting system, tries to conduct the verification from computing method and theory logic, which means normalizing data of different sizes into standard data for process calculation; Taking geometric evaluation of structure feature, character feature and probability as data base for calculation and reduction, to segment Chinese characters into specified characters and recognize multi-character segmentation of Chinese, English and numbers, to get the main process of realization of multi-character recognition; Through Chinese recognition of sentences and rule repository to improve recognition effectiveness of sentences, and to lower the exploration for difficult problems, to verify the method for accelerating correctness of characters, to promote correctness of the whole segmentation.In the last part, combined with current needs this paper has verified the methods mentioned in this paper using multi-characters handwriting. Through the application on e-reader it can be shown the effectiveness of multi-characters recognition on processing Chinese, English, numbers and signals.