Open Source 4DAnyone Generates Photorealistic 4D Humans from a Single Video
thisguyknowsai · x · 2026-08-31
A research team released an open-source system called 4DAnyone. Creating a photorealistic 4D human typically requires a calibrated wall of synchronized cameras, but this system works from just one casual video. It generates footage from angles the camera never captured, providing the views needed to build a moving 3D version of a person. Users can simply record a person in a normal video to create a dynamic 3D version observable from various perspectives.
Related event: 4DAnyone: Reconstructing 4D Humans from a Single Phone Video(3 posts)→
More from Multimodal
- Tessellation-free geometric queries on displaced surfaces cut caustics variance 10x — ssh4net · 2026-10-16
- Complex multi-link CAD mechanism finally animates, dev celebrates the breakthrough — zetalyrae · 2026-10-16
- OpenMOSS releases MOSS-Image-SFT: Apache 2.0 text-to-image model on Hugging Face — bdsqlsz · 2026-10-16
- Hyperflow open-sourced: 49 sampling steps down to 8, ~3x faster video generation — mhdfaran · 2026-10-16
- Cinematic 3D fantasy character prompt template with salvaged-robot armor — azed_ai · 2026-10-16
- Midjourney vintage woodcut style: --sref 2446102396 beaver with a cleaver — michaelrabone · 2026-10-16