复杂背景下基于定位的人体动作识别算法.

Most existing human action recognition approaches require a large amount training data or MoCAP to handle multiple viewpoints, and often rely on clean actor silhouettes. The paper presents an approach to recognize single actor human actions in complex backgrounds. The method tracks the actor pose by sampling from 3D action models. The action models in our approach are obtained by annotating key poses in 2D, lifting them to 3D stick figures and then computing the transformation matrices between the 3D key pose figures. In addition, poses sampled from coarse action models may not fit the observations well, to overcome this difficulty, we propose an approach for efficiently localizing a pose by generating a pose-specific part model (PSPM), which captures appropriate kinematic and occlusion constraints in a tree-structure. In addition, our approach does not require pose silhouettes. We show improvements to previous results on two publicly available datasets as well as on a novel, augmented dataset with dynamic backgrounds. [ABSTRACT FROM AUTHOR]