Enregistré dans:
| Auteurs principaux: | , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2024
|
| Sujets: | |
| Accès en ligne: | https://arxiv.org/abs/2411.17532 |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866913586341740544 |
|---|---|
| author | Li, Chengjian Shu, Xiangbo Cui, Qiongjie Yao, Yazhou Tang, Jinhui |
| author_facet | Li, Chengjian Shu, Xiangbo Cui, Qiongjie Yao, Yazhou Tang, Jinhui |
| contents | Diffusion models achieve impressive performance in human motion generation. However, current approaches typically ignore the significance of frequency-domain information in capturing fine-grained motions within the latent space (e.g., low frequencies correlate with static poses, and high frequencies align with fine-grained motions). Additionally, there is a semantic discrepancy between text and motion, leading to inconsistency between the generated motions and the text descriptions. In this work, we propose a novel diffusion-based FTMoMamba framework equipped with a Frequency State Space Model (FreqSSM) and a Text State Space Model (TextSSM). Specifically, to learn fine-grained representation, FreqSSM decomposes sequences into low-frequency and high-frequency components, guiding the generation of static pose (e.g., sits, lay) and fine-grained motions (e.g., transition, stumble), respectively. To ensure the consistency between text and motion, TextSSM encodes text features at the sentence level, aligning textual semantics with sequential features. Extensive experiments show that FTMoMamba achieves superior performance on the text-to-motion generation task, especially gaining the lowest FID of 0.181 (rather lower than 0.421 of MLD) on the HumanML3D dataset. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2411_17532 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | FTMoMamba: Motion Generation with Frequency and Text State Space Models Li, Chengjian Shu, Xiangbo Cui, Qiongjie Yao, Yazhou Tang, Jinhui Computer Vision and Pattern Recognition Diffusion models achieve impressive performance in human motion generation. However, current approaches typically ignore the significance of frequency-domain information in capturing fine-grained motions within the latent space (e.g., low frequencies correlate with static poses, and high frequencies align with fine-grained motions). Additionally, there is a semantic discrepancy between text and motion, leading to inconsistency between the generated motions and the text descriptions. In this work, we propose a novel diffusion-based FTMoMamba framework equipped with a Frequency State Space Model (FreqSSM) and a Text State Space Model (TextSSM). Specifically, to learn fine-grained representation, FreqSSM decomposes sequences into low-frequency and high-frequency components, guiding the generation of static pose (e.g., sits, lay) and fine-grained motions (e.g., transition, stumble), respectively. To ensure the consistency between text and motion, TextSSM encodes text features at the sentence level, aligning textual semantics with sequential features. Extensive experiments show that FTMoMamba achieves superior performance on the text-to-motion generation task, especially gaining the lowest FID of 0.181 (rather lower than 0.421 of MLD) on the HumanML3D dataset. |
| title | FTMoMamba: Motion Generation with Frequency and Text State Space Models |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2411.17532 |