Summary of Adaptive Hyper-graph Convolution Network For Skeleton-based Human Action Recognition with Virtual Connections, by Youwei Zhou and Tianyang Xu and Cong Wu and Xiaojun Wu and Josef Kittler
Adaptive Hyper-Graph Convolution Network for Skeleton-based Human Action Recognition with Virtual Connections
by Youwei Zhou, Tianyang Xu, Cong Wu, Xiaojun Wu, Josef Kittler
First submitted to arxiv on: 22 Nov 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Machine Learning (cs.LG)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper presents an adaptive graph convolutional network (GCN) for action recognition that leverages the shared topology of human skeletons. Unlike existing GCNs that rely on binary connections between neighboring joints, this study explores the benefits of constructing multi-vertex convolution structures using hyper-graphs. The proposed Hyper-GCN adaptively optimizes hyper-graphs during training to uncover latent relationships within actions. Additionally, virtual connections are introduced to support efficient feature aggregation and extend dependencies within skeletons. Experiments on NTU-60, NTU-120, and NW-UCLA datasets demonstrate the superiority of our approach, achieving 90.5% and 91.7% top-1 recognition accuracy on X-Sub and X-Set. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper develops a new way to recognize human actions using graphs. Currently, most action recognition systems rely on simple connections between joints in the skeleton. However, this study shows that by using more complex relationships between multiple joints, we can achieve better results. This is achieved through an “adaptive” graph convolutional network (Hyper-GCN) that learns how to combine information from different parts of the skeleton during training. The result is a more accurate way to recognize actions. |
Keywords
» Artificial intelligence » Convolutional network » Gcn
