On-Line Hand-written Character Recognition Based On The Mobile Platform

Language is the main medium to communicate, exchange, and record information. With popularity of the personal computer and development of information technology, processing characters with computer has become an important research field. Chinese character reorganization is an important sub-field of pattern recognition, which is also a most difficult issue to be fixed, involve the pattern recognition, image processing, digital signal processing, natural language understanding, artificial intelligence, vague mathematics, informationism, computer science, Chinese language information processing and so on. Chinese Character recognition is a integrated technology. And it has the practical value and theoretical significance on many high technology fields such as office automation, machine translation, and artificial intelligence. In this dissertation, preprocessing, coarse classification, feature extraction and designing of class selector in directional element feature studied to recognize not only regular written characters but also characters with stoke-connected. The main contents of dissertation are as follows:1. The dissertation reviews the development of on-line handwritten Chinese character recognition and analyzes the on-line handwritten Chinese character in details.2. In this paper I refined the character image pre-processing steps, and compared a variety of normalization methods on the impact of the recognition results, and then using linear and nonlinear combination of methods to improve the anti-noise performance of this system. The final adoption of trajectory points, re-sampling and filtering to eliminate noise and enhance the system robustness.3. This paper focuses on the research of feature extraction and the contrastive analysis between characters structural features and statistical features. Based on this content, the scheme for on-line handwritten Chinese character recognition is proposed by directional element feature through a variety of methods on meshing and the weighted (grid, the weighted grid, information of stroke, etc.).4. Other focuses of this dissertation has been placed on how to conduct within-class clustering, how to determine representative point of classes and how to select the sample quantity and quality in learning process.5. In the Windows Mobile platform, using the VC2005 we developed a set of practical on-line handwritten Chinese character recognition software.