arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2004.01278cs.CV
收藏

知道什么,在哪里,什么时候看:有效的视频动作建模与注意力

Knowing What, Where and When to Look: Efficient Video Action Modeling with Attention

Samsung AI Centre(三星人工智能中心);University of Surrey(萨里大学)

Juan-Manuel Perez-Rua, Brais Martinez, Xiatian Zhu, Antoine Toisoul, Victor Escorcia, Tao Xiang

摘要:

由于无约束视频在时间和空间上具有丰富的冗余信息,因此注意视频建模对于无约束视频中的动作识别至关重要。然而,由于两个原因,在深度神经网络中引入注意力以进行动作识别是具有挑战性的。首先,一个有效的注意力模块需要学习什么(对象及其局部运动模式),在哪里(空间),以及何时(时间)关注。其次,视频注意力模块必须是有效的,因为现有的动作识别模型已经遭受高计算成本。为了解决这两个挑战,提出了一种新的What-Where-When(W3)视频注意模块。从现有的替代方案出发,我们的W3模块模型的所有三个方面的视频注意力联合。至关重要的是,它通过将高维视频特征数据分解为低维有意义的空间(“什么”的1D通道向量和“哪里”的2D空间张量),然后进行轻量级的时间注意力推理,效率非常高。大量的实验表明,我们的注意力模型带来了显着的改进,现有的动作识别模型,实现了新的国家的最先进的性能在一些基准。

英文摘要:

Attentive video modeling is essential for action recognition in unconstrained videos due to their rich yet redundant information over space and time. However, introducing attention in a deep neural network for action recognition is challenging for two reasons. First, an effective attention module needs to learn what (objects and their local motion patterns), where (spatially), and when (temporally) to focus on. Second, a video attention module must be efficient because existing action recognition models already suffer from high computational cost. To address both challenges, a novel What-Where-When (W3) video attention module is proposed. Departing from existing alternatives, our W3 module models all three facets of video attention jointly. Crucially, it is extremely efficient by factorizing the high-dimensional video feature data into low-dimensional meaningful spaces (1D channel vector for `what' and 2D spatial tensors for `where'), followed by lightweight temporal attention reasoning. Extensive experiments show that our attention model brings significant improvements to existing action recognition models, achieving new state-of-the-art performance on a number of benchmarks.

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。

↑