A Cross-Dataset Study for Text-based 3D Human Motion Retrieval
Fuente:
arXiv
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911888812539904 |
|---|---|
| author | Bensabath, Léore Petrovich, Mathis Varol, Gül |
| author_facet | Bensabath, Léore Petrovich, Mathis Varol, Gül |
| contents | We provide results of our study on text-based 3D human motion retrieval and particularly focus on cross-dataset generalization. Due to practical reasons such as dataset-specific human body representations, existing works typically benchmarkby training and testing on partitions from the same dataset. Here, we employ a unified SMPL body format for all datasets, which allows us to perform training on one dataset, testing on the other, as well as training on a combination of datasets. Our results suggest that there exist dataset biases in standard text-motion benchmarks such as HumanML3D, KIT Motion-Language, and BABEL. We show that text augmentations help close the domain gap to some extent, but the gap remains. We further provide the first zero-shot action recognition results on BABEL, without using categorical action labels during training, opening up a new avenue for future research. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2405_16909 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | A Cross-Dataset Study for Text-based 3D Human Motion Retrieval Bensabath, Léore Petrovich, Mathis Varol, Gül Computer Vision and Pattern Recognition We provide results of our study on text-based 3D human motion retrieval and particularly focus on cross-dataset generalization. Due to practical reasons such as dataset-specific human body representations, existing works typically benchmarkby training and testing on partitions from the same dataset. Here, we employ a unified SMPL body format for all datasets, which allows us to perform training on one dataset, testing on the other, as well as training on a combination of datasets. Our results suggest that there exist dataset biases in standard text-motion benchmarks such as HumanML3D, KIT Motion-Language, and BABEL. We show that text augmentations help close the domain gap to some extent, but the gap remains. We further provide the first zero-shot action recognition results on BABEL, without using categorical action labels during training, opening up a new avenue for future research. |
| title | A Cross-Dataset Study for Text-based 3D Human Motion Retrieval |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2405.16909 |