NeRF++: Analyzing and Improving Neural Radiance Fields
This work addresses a parametrization issue for 360-degree captures in unbounded scenes, which is an incremental improvement for view synthesis applications.
The paper tackles the problem of applying Neural Radiance Fields (NeRF) to 360-degree captures of objects in large-scale, unbounded 3D scenes, improving view synthesis fidelity in this challenging scenario.
Neural Radiance Fields (NeRF) achieve impressive view synthesis results for a variety of capture settings, including 360 capture of bounded scenes and forward-facing capture of bounded and unbounded scenes. NeRF fits multi-layer perceptrons (MLPs) representing view-invariant opacity and view-dependent color volumes to a set of training images, and samples novel views based on volume rendering techniques. In this technical report, we first remark on radiance fields and their potential ambiguities, namely the shape-radiance ambiguity, and analyze NeRF's success in avoiding such ambiguities. Second, we address a parametrization issue involved in applying NeRF to 360 captures of objects within large-scale, unbounded 3D scenes. Our method improves view synthesis fidelity in this challenging scenario. Code is available at https://github.com/Kai-46/nerfplusplus.