Technology & AIAug 20, 2026
4DAnyone: Create Anyone in 4D from a Casual Monocular Video
We present 4DAnyone, a framework for reconstructing 4D humans from an uncalibrated monocular video by generating reconstruction-grade multiview-consistent videos and lifting them into 4D Gaussian Splatting (4DGS).
We present 4DAnyone, a framework for reconstructing 4D humans from an uncalibrated monocular video by generating reconstruction-grade multiview-consistent videos and lifting them into 4D Gaussian Splatting (4DGS). Existing camera-controlled video diffusion models synthesize…
Sign in to learn & save →
The frontier is open to all. Sign in to learn this from first principles and save it to your knowledge base.