Learning spatio-temporal co-occurrence correlograms for efficient human action classification

Spatio-temporal interest point (STIP) based features show great promises in human action analysis with high efficiency and robustness. However, they typically focus on bag-of-visual words (BoVW), which omits any correlation among words and shows limited discrimination in real-world videos. In this paper, we propose a novel approach to add the spatio-temporal co-occurrence relationships of visual words to BoVW for a richer representation. Rather than assigning a particular scale on videos, we adopt the normalized google-like distance (NGLD) to measure the words´ co-occurrence semantics, which grasps the videos´ structure information in a statistical way. All pairwise distances in spatial and temporal domain compose the corresponding NGLD correlograms, then their united form is incorporated with BoVW by training a multi-channel kernel SVM classifier. Experiments on real-world datasets (KTH and UCF sports) validate the efficiency of our approach for the classification of human actions.