Quality assessment of 3D human animation: Subjective and objective evaluation
This work addresses a gap in quality assessment for virtual human animations, which is important for applications in virtual and augmented reality, but it is incremental as it builds on existing task-oriented evaluation metrics.
The paper tackles the problem of assessing the quality of 3D human animations, particularly those not generated with parametric body models, by introducing a data-driven framework that includes a dataset with subjective realism scores and a linear regressor achieving a 90% correlation, outperforming a deep learning baseline.
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.