基于多级特征的行为识别.ppt

Action Recognition Using Multilevel Features andLatent Structural SVM 信息与通信工程 曹刘彬 2016.3.21 Abstract 1.We first propose a new low-level visual feature,called spatio-temporal context distribution feature of interest points,to describe human actions. 2.Each action video is expressed as a set of relative XYT coordinates between pairwise interest points in a local region. 3. In order to capture the spatio-temporal relationships at different levels,multiple GMMs are utilized to describe the context distributions of interest points over multiscale local regions. 4. Motivated by the observation that some actions share similar motion patterns, we additionally propose a novel mid-level class correlation feature to capture the semantic correlations between different action classes. 5. Moreover,human actions are often associated with some specific natural environments and also exhibit high correlation with particular scene classes. 6. It is therefore beneficial to utilize the contextual scene information for action recognition. 7. In this paper, we build the high-level co-occurrence relationship between action classes and scene classes to discover the mutual contextual constraints between action and scene. 8. By treating the scene class label as a latent variable, we propose to use the latent structural SVM (LSSVM) model to jointly capture the compatibility between multilevel action feature and action classes, the compatibility between multilevel scene features and scene classes, and the contextual relationship between action classes and scene classes. 9. Extensive experiments on UCF Sports, YouTube and UCF50 datasets demonstrate the effectiveness of the proposed multilevel features and action-scene interaction based LSSVM model for human action recognition. Multilevel Features for Action Description Low-Level Spatio-Temporal Context Distribution Feature Fig. 1 shows some examples of detected interest points of human actions. The subfigures on the left c