PG-NeuS: Robust and Efficient Point Guidance for Multi-View Neural Surface Reconstruction
arXiv:2310.07997 · doi:10.1109/TVCG.2024.3514748
Abstract
Recently, learning multi-view neural surface reconstruction with the supervision of point clouds or depth maps has been a promising way. However, due to the underutilization of prior information, current methods still struggle with the challenges of limited accuracy and excessive time complexity. In addition, prior data perturbation is also an important but rarely considered issue. To address these challenges, we propose a novel point-guided method named PG-NeuS, which achieves accurate and efficient reconstruction while robustly coping with point noise. Specifically, aleatoric uncertainty of the point cloud is modeled to capture the distribution of noise, leading to noise robustness. Furthermore, a Neural Projection module connecting points and images is proposed to add geometric constraints to implicit surface, achieving precise point guidance. To better compensate for geometric bias between volume rendering and point modeling, high-fidelity points are filtered into a Bias Network to further improve details representation. Benefiting from the effective point guidance, even with a lightweight network, the proposed PG-NeuS achieves fast convergence with an impressive 11x speedup compared to NeuS. Extensive experiments show that our method yields high-quality surfaces with high efficiency, especially for fine-grained details and smooth regions, outperforming the state-of-the-art methods. Moreover, it exhibits strong robustness to noisy data and sparse data.
References in corpus (13)
- Instant Neural Graphics Primitives with a Multiresolution Hash Encoding
- What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?
- Learning to Predict 3D Objects with an Interpolation-based Differentiable Renderer
- MonoSDF: Exploring Monocular Geometric Cues for Neural Implicit Surface Reconstruction
- Neural Dual Contouring
- Geo-Neus: Geometry-Consistent Neural Implicit Surfaces Learning for Multi-view Reconstruction
- HF-NeuS: Improved Surface Reconstruction Using High-Frequency Details
- Neural-Pull: Learning Signed Distance Functions from Point Clouds by Learning to Pull Space onto Surfaces
- Direct Voxel Grid Optimization: Super-fast Convergence for Radiance Fields Reconstruction
- Voxurf: Voxel-based Efficient and Accurate Neural Surface Reconstruction
- View Synthesis with Sculpted Neural Points
- VQ-NeRF: Neural Reflectance Decomposition and Editing with Vector Quantization
- Depth-NeuS: Neural Implicit Surfaces Learning for Multi-view Reconstruction Based on Depth Information Optimization