Gait Recognition Using CHLAC Features and Hidden Markov Model

摘要

Variations in speed have a strong impact on gait recognition performance. In this paper, we propose a gait recognition method that is robust against speed variation. Our method is based on combining Fisher Discriminant Analysis (FDA)-based Cubic Higher-order Local Auto-Correlation (CHLAC) and the statistical framework provided by hidden Markov models (HMMs). This combination is aimed at better retaining the spatio-temporal characteristics of gait sequences. We compared the performance of the system with a combination of Principal Component Analysis (PCA) features and HMMs, and also with a combination of CHLAC and k-Nearest Neighbour (k-NN). We found that CHLAC-FDA combined with HMMs gave better results. Applying a Gaussian mixture to the HMMs also provided improvement in the performance.