An improved clustering for action recognition in online video

A new method for human action recognition in online video sequences using Latent Dirichlet Markov Clustering (LDMC) is proposed. Video sequences are represented by a novel "bag-of-words" representation, and each frame corresponds to a "word". LDMC builds on Hidden Markov Models (HMMs) and Latent Dirichlet Allocation, and it overcome their low recognition rate, robustness and high computational complexity. A collapsed Gibbs sampler is designed for offline learning with unlabeled training data, and a new approximation to online Bayesian inference is formulated to enable human action recognition in new online video sequence in real-time. The strength of this model is demonstrated by unsupervised learning of human action categories and detecting salient actions in one complex and crowded public scenes.