Quality assessment of 3D human animation: Subjective and objective evaluation
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914158092484608 |
|---|---|
| author | Rekik, Rim Wuhrer, Stefanie Hoyet, Ludovic Zibrek, Katja Olivier, Anne-Hélène |
| author_facet | Rekik, Rim Wuhrer, Stefanie Hoyet, Ludovic Zibrek, Katja Olivier, Anne-Hélène |
| contents | Virtual human animations have a wide range of applications in virtual and augmented reality. While automatic generation methods of animated virtual humans have been developed, assessing their quality remains challenging. Recently, approaches introducing task-oriented evaluation metrics have been proposed, leveraging neural network training. However, quality assessment measures for animated virtual humans that are not generated with parametric body models have yet to be developed. In this context, we introduce a first such quality assessment measure leveraging a novel data-driven framework. First, we generate a dataset of virtual human animations together with their corresponding subjective realism evaluation scores collected with a user study. Second, we use the resulting dataset to learn predicting perceptual evaluation scores. Results indicate that training a linear regressor on our dataset results in a correlation of 90%, which outperforms a state of the art deep learning baseline. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_23301 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Quality assessment of 3D human animation: Subjective and objective evaluation Rekik, Rim Wuhrer, Stefanie Hoyet, Ludovic Zibrek, Katja Olivier, Anne-Hélène Graphics Computer Vision and Pattern Recognition Virtual human animations have a wide range of applications in virtual and augmented reality. While automatic generation methods of animated virtual humans have been developed, assessing their quality remains challenging. Recently, approaches introducing task-oriented evaluation metrics have been proposed, leveraging neural network training. However, quality assessment measures for animated virtual humans that are not generated with parametric body models have yet to be developed. In this context, we introduce a first such quality assessment measure leveraging a novel data-driven framework. First, we generate a dataset of virtual human animations together with their corresponding subjective realism evaluation scores collected with a user study. Second, we use the resulting dataset to learn predicting perceptual evaluation scores. Results indicate that training a linear regressor on our dataset results in a correlation of 90%, which outperforms a state of the art deep learning baseline. |
| title | Quality assessment of 3D human animation: Subjective and objective evaluation |
| topic | Graphics Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2505.23301 |