Abstract: Facsimileing of existing painting artworks is one of the most effective methods to learn traditional ink painting.However,this method requires a lot of immovable and duplicated work thus is too boring for children.To allow natural and intuitive example-based learning,this paper proposes a new Shu-Kong facsimile interaction scheme that employs motion sensors instead of traditional painting brushes to strengthen the interaction.First,stokes from ink paintings are extracted and their corresponding hand gestures are collected as training samples.Then,HMM is employed as underlying model for gesture training and recognition.Finally,the recognized stokes are deformed according to the quantitative difference between the input and sample gestures.We evaluate the Shu-Kong based painting interface by having children as target users on a test set of Chinese ink painting artworks.Experiment results show that our approach can inspire the interests of the children on learning ink painting more effectively than the traditional pen-based interface.