T2M Mamba: Motion Periodicity-Saliency Coupling Approach for Stable Text-Driven Motion Generation
Xingzu Zhan, Chen Xie, Honghang Chen, Yixun Lin, Xiaochun Mai · 2026
Read T2M Mamba: Motion Periodicity-Saliency Coupling Approach for Stable Text-Driven Motion Generation by Xingzu Zhan, Chen Xie, Honghang Chen, Yixun Lin, Xiaochun Mai (2026) 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/arxiv-2602-01352.txt.
Description
Xingzu Zhan and co-authors build T2M Mamba, a text-to-motion plan tying beat and keyframe weight together. Past motion plans treat beat and keyframe weight apart, drifting on long runs, and crumble when rewordings shift text vectors. Their beat-salience Mamba gauges keyframe weights via boosted density-peak grouping and FFT-sped self-match, catching linked motion cheaply. A periodic align module then firms text-motion links. Tests on HumanML3D and KIT-ML confirm gains, with an FID of 0.068 and lifts elsewhere.
Catalog record
| CREATOR | Xingzu Zhan, Chen Xie, Honghang Chen, Yixun Lin, Xiaochun Mai |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 34938 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602.01352