Grouped Spatial-Temporal Aggregation for Efficient Action Recognition
arXiv:1909.13130
Abstract
Temporal reasoning is an important aspect of video analysis. 3D CNN shows good performance by exploring spatial-temporal features jointly in an unconstrained way, but it also increases the computational cost a lot. Previous works try to reduce the complexity by decoupling the spatial and temporal filters. In this paper, we propose a novel decomposition method that decomposes the feature channels into spatial and temporal groups in parallel. This decomposition can make two groups focus on static and dynamic cues separately. We call this grouped spatial-temporal aggregation (GST). This decomposition is more parameter-efficient and enables us to quantitatively analyze the contributions of spatial and temporal features in different layers. We verify our model on several action recognition tasks that require temporal reasoning and show its effectiveness.
ICCV 2019
References in corpus (9)
- Two-Stream Convolutional Networks for Action Recognition in Videos
- UCF101: A Dataset of 101 Human Actions Classes From Videos in The Wild
- SlowFast Networks for Video Recognition
- Hidden Two-Stream Convolutional Networks for Action Recognition
- Object Level Visual Reasoning in Videos
- On the effectiveness of task granularity for transfer learning
- Attention is All We Need: Nailing Down Object-centric Attention for Egocentric Activity Recognition
- Long-Term Feature Banks for Detailed Video Understanding
- Learning Discriminative Motion Features Through Detection