No abstract

No abstract

Gait, as a typical feature of biology, highlights such advantages as non-contact, non-invasive, difficulty of hiding and disguise etc. so that the gait recognition has become one of the hottest research directions in the biological feature recognition. In this paper, a multi-feature hybrid recognition algorithm is proposed based on the ICP algorithm for a specific gait database. Firstly, the feature point sets of the gait database are classified and the feature curves are generated; then, the ICP matching for each feature curve is implemented in a step by step manner. Finally, an appropriate threshold is selected according to the ROC curve. The experimental results show that the algorithm still has good recognition effect in the condition of low resolution.