NeRFReN: Neural Radiance Fields with Reflections
This addresses a specific limitation in view synthesis for scenes with reflections, enabling better rendering and editing applications, but it is incremental as it builds directly on NeRF.
The paper tackled the problem of modeling complex reflections like glass and mirrors in Neural Radiance Fields (NeRF), which previously led to inaccurate depth and blurry renderings, and introduced NeRFReN to split scenes into transmitted and reflected components, achieving high-quality novel view synthesis and physically sound depth estimation in experiments.
Neural Radiance Fields (NeRF) has achieved unprecedented view synthesis quality using coordinate-based neural scene representations. However, NeRF's view dependency can only handle simple reflections like highlights but cannot deal with complex reflections such as those from glass and mirrors. In these scenarios, NeRF models the virtual image as real geometries which leads to inaccurate depth estimation, and produces blurry renderings when the multi-view consistency is violated as the reflected objects may only be seen under some of the viewpoints. To overcome these issues, we introduce NeRFReN, which is built upon NeRF to model scenes with reflections. Specifically, we propose to split a scene into transmitted and reflected components, and model the two components with separate neural radiance fields. Considering that this decomposition is highly under-constrained, we exploit geometric priors and apply carefully-designed training strategies to achieve reasonable decomposition results. Experiments on various self-captured scenes show that our method achieves high-quality novel view synthesis and physically sound depth estimation results while enabling scene editing applications.