Automatic gait recognition via statistical approaches for extendedtemplate features

摘要

A gait recognition system using extended template features isnpresented. A proposed statistical approach is applied for featurenextraction from spatial and temporal templates. This method can be usednto reduce data dimensionality and to optimize the class separability ofndifferent gait sequences simultaneously. Dimensionality reduction isnachieved by template extraction followed by principal componentnanalysis. Gait recognition is achieved in the canonical space using anmeasure of accumulated distance as the metric. By incorporating spatialnand temporal information into an extended feature, gait recognitionnbecomes more robust and accurate than using spatial or temporal featuresnalone