基于稀疏体素表示的高保真曲面通用神经重建
Generalizable Neural Reconstruction of High-Fidelity Surfaces via Sparse Volumetric Representations
AI总结：
针对通用神经曲面重建中稠密体素导致的内存与分辨率瓶颈，提出SVRecon框架，利用占用率驱动的稀疏体素和两阶段架构，在32GB硬件上实现高达512³分辨率的高保真重建，在稀疏视图下表现更优。
摘要：
神经隐式表示最近在新视图合成和多视图三维重建中取得了令人瞩目的成果，但基于NeRF和高斯泼溅的方法都需要逐场景优化，这使得它们效率低下。通用神经曲面重建（GNSR）方法通过学习直接从输入图像预测的特征表示，旨在消除这一需求。然而，这些方法通常依赖稠密特征体素，由于高昂的内存成本，严重限制了可达到的分辨率和保真度。我们提出了稀疏体素重建（SVRecon），一种新的GNSR框架，通过学习驱动的占用率稀疏性，以比早期在GNSR中引入稀疏性的方法更有效的方式，实现了高分辨率、内存高效的重建。我们的方法采用嵌套的两阶段架构：（1）一个占用率预测网络，用于识别包含表面的体素；（2）一个仅在这些占用区域内定义的高分辨率稀疏体素渲染框架，以及专门为光线采样、特征聚合和查询设计的稀疏化算法。这种设计使得能够进行精细的曲面重建，同时避免了稠密网格的巨大内存占用。SVRecon在标准32GB硬件上可运行高达$512^3$的分辨率，远高于先前的通用方法，并在多种数据集上，尤其是在稀疏视图设置下，生成更平滑、更精确的重建结果。
英文摘要：
Neural implicit representations have recently achieved impressive results in novel view synthesis and multi-view 3D reconstruction, yet both NeRF- and Gaussian Splatting-based methods require per-scene optimization, which makes them inefficient. Generalizable Neural Surface Reconstruction (GNSR) methods have been proposed to remove this need by learning feature representations directly predicted from input images. However, their typical reliance on dense feature volumes severely limits achievable resolution and fidelity due to prohibitive memory costs. We introduce Sparse Volumetric Reconstruction (SVRecon), a new GNSR framework that unlocks high-resolution, memory-efficient reconstruction through learned occupancy-driven sparsity, in a more effective way than earlier approaches to introducing sparsity in GNSRs. Our approach uses a nested two-stage architecture: (1) an occupancy prediction network that identifies surface-containing voxels, and (2) a high-resolution sparse volume rendering framework defined only within these occupied regions, together with specialized sparsified algorithms for ray sampling, feature aggregation, and querying. This design enables fine-grained surface reconstruction while avoiding the heavy memory footprint of dense grids. SVRecon operates at resolutions up to $512^3$ on standard 32GB hardware---substantially higher than prior generalizable methods---and delivers smoother and more precise reconstructions across diverse datasets, particularly in sparse-view settings.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
