Jon Barron Profile picture
AI researcher at Google DeepMind, NeRF person, eigenvector enthusiast. Views are my own.
2 subscribers
Mar 3, 2022 5 tweets 3 min read
Very glad I can finally talk about our newly-minted #CVPR2022 paper. We extended mip-NeRF to handle unbounded "360" scenes, and it got us ~photorealistic renderings and beautiful depth maps. Explainer video: and paper: arxiv.org/abs/2111.12077 We "contract" Euclidean space into a bounded domain, which gets hard because we need to warp the mip-NeRF Gaussians that model 3D volumes of space. The trick for making this work is linearizing the contraction (thanks JAX!) and using the same math as an extended Kalman filter.