StNet: Local and Global Spatial-Temporal Modeling for Action Recognition
This work addresses action recognition in videos, offering a novel architecture that improves accuracy while managing complexity, though it is incremental in advancing spatial-temporal modeling techniques.
The paper tackles the problem of designing effective network architectures for spatial-temporal modeling in videos by introducing StNet, which uses super-images and temporal convolution to capture local and global relationships, achieving state-of-the-art performance on the Kinetics dataset with a balance between accuracy and model complexity.
Despite the success of deep learning for static image understanding, it remains unclear what are the most effective network architectures for the spatial-temporal modeling in videos. In this paper, in contrast to the existing CNN+RNN or pure 3D convolution based approaches, we explore a novel spatial temporal network (StNet) architecture for both local and global spatial-temporal modeling in videos. Particularly, StNet stacks N successive video frames into a \emph{super-image} which has 3N channels and applies 2D convolution on super-images to capture local spatial-temporal relationship. To model global spatial-temporal relationship, we apply temporal convolution on the local spatial-temporal feature maps. Specifically, a novel temporal Xception block is proposed in StNet. It employs a separate channel-wise and temporal-wise convolution over the feature sequence of video. Extensive experiments on the Kinetics dataset demonstrate that our framework outperforms several state-of-the-art approaches in action recognition and can strike a satisfying trade-off between recognition accuracy and model complexity. We further demonstrate the generalization performance of the leaned video representations on the UCF101 dataset.