Attribute-based learning for gait recognition using spatio-temporal interest points

This paper proposes a new method to extract a gait feature from a raw gait video directly. The Space-Time Interest Points (STIPs) are detected where there are significant movements of human body along both spatial and temporal directions in local spatio-temporal volumes of a raw gait video. Then, a histogram of STIP descriptors (HSD) is constructed as a gait feature. Display Omitted A new gait feature is extracted from a raw gait video directly.The proposed gait feature extraction is performed in the spatio-temporal domain.The attribute-based learning is used to enhance the SVM-based gait classification.The proposed method has outstanding performances under changes of clothing types.The proposed method has outstanding performances under changes of carry conditions.