SIGNeRF: Scene Integrated Generation for Neural Radiance Fields
This addresses the need for efficient scene editing in 3D generation for applications like virtual reality or content creation, though it appears incremental by building on existing NeRF and diffusion model techniques.
The paper tackles the problem of editing existing photorealistic 3D scenes and generating objects within them, which is challenging for current object-centric generative 3D methods, and achieves fast, controllable editing without iterative optimization by using a multi-view reference sheet and depth-conditioned diffusion models.
Advances in image diffusion models have recently led to notable improvements in the generation of high-quality images. In combination with Neural Radiance Fields (NeRFs), they enabled new opportunities in 3D generation. However, most generative 3D approaches are object-centric and applying them to editing existing photorealistic scenes is not trivial. We propose SIGNeRF, a novel approach for fast and controllable NeRF scene editing and scene-integrated object generation. A new generative update strategy ensures 3D consistency across the edited images, without requiring iterative optimization. We find that depth-conditioned diffusion models inherently possess the capability to generate 3D consistent views by requesting a grid of images instead of single views. Based on these insights, we introduce a multi-view reference sheet of modified images. Our method updates an image collection consistently based on the reference sheet and refines the original NeRF with the newly generated image set in one go. By exploiting the depth conditioning mechanism of the image diffusion model, we gain fine control over the spatial location of the edit and enforce shape guidance by a selected region or an external mesh.