Abstract: This paper proposed an algorithm using skeleton information to do shape recognition and classification. Using the information of skeleton junction nodes,the problem of shape similarity is converted to distance of skeleton junction nodes. Using method of neural network, the classification results can be got at the output of the network. The method can solve the recognition of nonrigid objects. The experiment results show that our method has better recognition results than other method.

Key words: Skeleton, Neural network, Shape recognition, Shape classification