WildAvatar: Learning In-the-wild 3D Avatars from the Web
This addresses the scalability and real-world representation issues in avatar creation for applications in 3D human modeling and related fields, though it is incremental as it builds on existing methods with new data and filtering.
The authors tackled the problem of creating 3D avatars from real-world web videos by proposing an automatic annotation pipeline and curating WildAvatar, a dataset with over 10,000 human subjects that is at least 10 times richer than previous datasets and improves state-of-the-art methods on benchmarks.
Existing research on avatar creation is typically limited to laboratory datasets, which require high costs against scalability and exhibit insufficient representation of the real world. On the other hand, the web abounds with off-the-shelf real-world human videos, but these videos vary in quality and require accurate annotations for avatar creation. To this end, we propose an automatic annotating pipeline with filtering protocols to curate these humans from the web. Our pipeline surpasses state-of-the-art methods on the EMDB benchmark, and the filtering protocols boost verification metrics on web videos. We then curate WildAvatar, a web-scale in-the-wild human avatar creation dataset extracted from YouTube, with $10000+$ different human subjects and scenes. WildAvatar is at least $10\times$ richer than previous datasets for 3D human avatar creation and closer to the real world. To explore its potential, we demonstrate the quality and generalizability of avatar creation methods on WildAvatar. We will publicly release our code, data source links and annotations to push forward 3D human avatar creation and other related fields for real-world applications.