In this dissertation, on-line hand-written Uyghur word segmentation and character recognition on the basis of the structural features of Uyghur are studied. The main contents of dissertation are as follows:1. The dissertation introduces the development and the importance of Uyghur character recognition, and analyzes the difficulty of on-line handwritten Uyghur character recognition in details.2. A brief introduction of Uyghur and the characteristics of Uyghur are given, which contain the Uyghur writing style, the form of connection, disconnected writing of the main-stroke, joined-up writing of the point-strokes etc.3. Both on-line word segmentation and off-line word segmentation of Uyghur are introduced in this paper. By comparing of the two segmentation methods, we obtain the strengths and weaknesses of them. The on-line segmentation contains excessive segmentation and post-processing. The excessive segmentation is completed by two steps: extraction of the connected main-stroke, segmentation of the connected main-stroke.4. This paper focuses on the research of feature extraction, which contains structural features of stroke and meshing direction feature. By the analysis of the experimental results, the reason that results low recognizable rate of the structural features of stroke is obtained.5. In the Windows Mobile platform, we developed a program on-line handwritten Uyghur character recognition and word segmentation with VS2005. |