On-Line Sketchy Graphic Recognition Based on Adaptive HMM

User adaptation is a crux for on-line sketchy graphic recognition. The purpose of this paper is to make a comprehensive discuss about some critical issues of user adaptation in on-line sketch recognition supporting normal graphics design, we also hope to make some achievements in adaptive ouline sketch recognition. A method of on-line sketchy graphics recognition for user adaptation is proposed in this paper. We make a preparatory research on selec- tion mechanism of feature descriptor and propose a combined feature based on curvature, velocity and geometrical character of stroke. We make some improvement in traditional HMM and propose an adaptive HMM based on vari- able state number for the purpose of description of multi-stroke sketch. Experiments prove the proposed method both effective and efficient.