Action Recognition Based on Multi-Level Topological Channel Attention of Human Skeleton
Kai Hu, Chaowen Shen, Tianyan Wang, Shuai Shen, Chengxue Cai, Huaming Huang, Min Xia, Cosimo Distante · 2023
Read Action Recognition Based on Multi-Level Topological Channel Attention of Human Skeleton by Kai Hu, Chaowen Shen, Tianyan Wang, Shuai Shen, Chengxue Cai, Huaming Huang, Min Xia, Cosimo Distante (2023) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-10765606.txt.
Description
Kai Hu and a team of co-authors teach computers to name human actions from skeleton data alone. Working from body joints strips away background and lighting noise that confuse normal video. Graph networks suit the skeleton's shape, but older ones underuse the body's known structure and how limbs coordinate.
Their network stacks three ideas. A topology module reads the body plan from coarse to fine while weighing channel importance. A coordination module models how opposite and same-side limbs move together. A global module then catches patterns across space and time at several scales. On the two big NTU benchmarks, accuracy reached the low-to-mid nineties across the standard splits.
Catalog record
| CREATOR | Kai Hu, Chaowen Shen, Tianyan Wang, Shuai Shen, Chengxue Cai, Huaming Huang, Min Xia, Cosimo Distante |
|---|---|
| YEAR | 2023 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 62 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |