NeuS: Learning Neural Implicit Surfaces by Volume Rendering for Multi-view Reconstruction
This addresses the challenge of accurate surface reconstruction for objects and scenes in computer vision, offering a robust solution without requiring mask supervision, though it builds incrementally on existing neural methods like NeRF.
The paper tackles the problem of reconstructing high-fidelity 3D surfaces from 2D images by proposing NeuS, a method that uses a neural signed distance function and a new volume rendering formulation to reduce geometric errors, resulting in state-of-the-art performance on datasets like DTU and BlendedMVS, especially for complex structures and self-occlusion.
We present a novel neural surface reconstruction method, called NeuS, for reconstructing objects and scenes with high fidelity from 2D image inputs. Existing neural surface reconstruction approaches, such as DVR and IDR, require foreground mask as supervision, easily get trapped in local minima, and therefore struggle with the reconstruction of objects with severe self-occlusion or thin structures. Meanwhile, recent neural methods for novel view synthesis, such as NeRF and its variants, use volume rendering to produce a neural scene representation with robustness of optimization, even for highly complex objects. However, extracting high-quality surfaces from this learned implicit representation is difficult because there are not sufficient surface constraints in the representation. In NeuS, we propose to represent a surface as the zero-level set of a signed distance function (SDF) and develop a new volume rendering method to train a neural SDF representation. We observe that the conventional volume rendering method causes inherent geometric errors (i.e. bias) for surface reconstruction, and therefore propose a new formulation that is free of bias in the first order of approximation, thus leading to more accurate surface reconstruction even without the mask supervision. Experiments on the DTU dataset and the BlendedMVS dataset show that NeuS outperforms the state-of-the-arts in high-quality surface reconstruction, especially for objects and scenes with complex structures and self-occlusion.