Segmentation of online sketching using geometric feature

The stroke segmentation is one of the important factors for checking recognition rate of online freehand sketching that is an effective way for expressing, communicating and recording ideas during conceptual design. Strokes are divided into convex and concave stroke depending on whether is change the vector direction of preprocessing points. The paper proposes two kinds of segmentation methods for the former and analyzes their differences. To the latter, groups the preprocessing points and follows the rule that is the most of square number of the preprocessing points in unit length to segment the stroke. This paper develops a human-computer interface prototype system FSR, which makes system interface easy and friendly to user. The FRS system has been tested with a number of freehand sketches; the test results show that the algorithm achieves a satisfactory segmenting efficiency. Our method that tolerates high degrees of uncertainty and vague ideas does not rely on any domain-specific knowledge, and therefore it can be easily integrated with other high-level applications.