Make Skeleton-based Action Recognition Model Smaller, Faster and Better
arXiv:1907.09658
Abstract
Although skeleton-based action recognition has achieved great success in recent years, most of the existing methods may suffer from a large model size and slow execution speed. To alleviate this issue, we analyze skeleton sequence properties to propose a Double-feature Double-motion Network (DD-Net) for skeleton-based action recognition. By using a lightweight network structure (i.e., 0.15 million parameters), DD-Net can reach a super fast speed, as 3,500 FPS on one GPU, or, 2,000 FPS on one CPU. By employing robust features, DD-Net achieves the state-of-the-art performance on our experimental datasets: SHREC (i.e., hand actions) and JHMDB (i.e., body actions). Our code will be released with this paper later.
6 pages, 5 figures
References in corpus (6)
- Adam: A Method for Stochastic Optimization
- VNect: Real-time 3D Human Pose Estimation with a Single RGB Camera
- Spatial Temporal Graph Convolutional Networks for Skeleton-Based Action Recognition
- An End-to-End Spatio-Temporal Attention Model for Human Action Recognition from Skeleton Data
- SlowFast Networks for Video Recognition
- Can WiFi Estimate Person Pose?
Cited by in corpus (5)
- Stronger, Faster and More Explainable: A Graph Convolutional Baseline for Skeleton-based Action Recognition
- A Survey on 3D Skeleton-Based Action Recognition Using Learning Method
- Efficient Visual Recognition with Deep Neural Networks: A Survey on Recent Advances and New Directions
- Skeleton-based Approaches based on Machine Vision: A Survey
- Memory Group Sampling Based Online Action Recognition Using Kinetic Skeleton Features