MrNeRF (@janusch_patas)
2024-11-21 | โค๏ธ 123 | ๐ 14
Generating 3D-Consistent Videos from Unposed Internet Photos
Abstract (excerpt) A handful of input images serve as keyframes, and our model interpolates between them to simulate a path moving between the cameras.
Given random images, a modelโs ability to capture underlying geometry, recognize scene identity, and relate frames in terms of camera position and orientation reflects a fundamental understanding of 3D structure and scene layout.
However, existing video models such as Luma Dream Machine fail at this task.
We design a self-supervised method that takes advantage of the consistency of videos and variability of multiview internet photos to train a scalable, 3D-aware video model without any 3D annotations such as camera parameters.
โฆ We also show our model benefits applications that enable camera control, such as 3D Gaussian Splatting.
๋ฏธ๋์ด
