Gait authentication based on multi-criterion model of acceleration features

A novel gait authentication method is developed, in this research work, based on multi-acceleration features, vertical acceleration signal is collected by a waist-mounted MEMS accelerometer and represented as a seven-feature-tuple extracted by wavelet transform. A multi-criterion model is designed to match the feature sequences using dynamic time wrapping algorithm. The proposed method is verified with the gait information from 24 volunteers. The test result shows that the equal error rate of authentication has achieved 5% with the proposed method which is superior to the results of 6.4% and 6.7% in the previous work.