arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1502.00258cs.CV
收藏

用于人体动作识别的隐时空组合模型学习

Learning Latent Spatio-Temporal Compositional Model for Human Action Recognition

Sun Yat-Sen University(中山大学);IBM Research(IBM研究院)

Xiaodan Liang, Liang Lin, Liangliang Cao

摘要:

动作识别是多媒体理解中的一个重要问题。本文通过建立一个表达性的组合动作模型来解决这个问题。我们在视频中的一个动作实例建模与合奏的时空组成:一些离散的时间锚帧,其中每一个被进一步分解到一个布局的可变形部分。通过这种方式,我们的模型可以识别时空与或图(STAOG)来表示动作的潜在结构,例如三级跳远,摇摆和跳高。STAOG模型包括四层:(i)底部的一批叶节点,用于检测视频补丁内的各种动作部分;(ii)底部的或节点,即开关变量,以激活它们的子叶节点用于结构可变性;(iii)锚定帧内的与节点,用于验证空间组成;以及(iv)顶部的根节点,用于聚合时间锚定帧上的分数。此外,上下文之间的相互作用被定义在空间和时间域的叶节点。对于模型训练,我们开发了一种新的弱监督学习算法,该算法迭代地确定结构配置(例如,与或节点相关联的叶节点的产生)以及多层参数的优化。通过充分利用时空组合和相互作用,我们的方法很好地处理大的类内动作方差(例如,不同的视图,个人外观,时空结构)。具有挑战性的数据库上的实验结果表明,我们的方法优于其他竞争方法的性能。

英文摘要:

Action recognition is an important problem in multimedia understanding. This paper addresses this problem by building an expressive compositional action model. We model one action instance in the video with an ensemble of spatio-temporal compositions: a number of discrete temporal anchor frames, each of which is further decomposed to a layout of deformable parts. In this way, our model can identify a Spatio-Temporal And-Or Graph (STAOG) to represent the latent structure of actions e.g. triple jumping, swinging and high jumping. The STAOG model comprises four layers: (i) a batch of leaf-nodes in bottom for detecting various action parts within video patches; (ii) the or-nodes over bottom, i.e. switch variables to activate their children leaf-nodes for structural variability; (iii) the and-nodes within an anchor frame for verifying spatial composition; and (iv) the root-node at top for aggregating scores over temporal anchor frames. Moreover, the contextual interactions are defined between leaf-nodes in both spatial and temporal domains. For model training, we develop a novel weakly supervised learning algorithm which iteratively determines the structural configuration (e.g. the production of leaf-nodes associated with the or-nodes) along with the optimization of multi-layer parameters. By fully exploiting spatio-temporal compositions and interactions, our approach handles well large intra-class action variance (e.g. different views, individual appearances, spatio-temporal structures). The experimental results on the challenging databases demonstrate superior performance of our approach over other competing methods.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑