VS-Splat: Voxel-Selective feed-forward Gaussian Splatting for end-to-end 3D object reconstruction from sparse-views

2026-09-10

AI Summary

本文提出VS-Splat模型，通过选择性预测物体中心体素内的高斯基元来改进从稀疏视角图像重建3D物体的细节表现。

Abstract

Feed-forward Gaussian splatting models have demonstrated remarkable effectiveness in reconstructing three-dimensional (3D) objects from a few two-dimensional (2D) images, even if they are unseen. As existing methods typically predict Gaussian primitives uniformly across the 3D space, most primitives are placed in non-object regions. This may hinder the representation of fine object details. This paper proposes a Voxel-Selective Gaussian Splatting model (VS-Splat), a new end-to-endfeed-forward Gaussian splatting framework that predicts many primitives only within selected voxels that are likely to belong to an object, without 3D structural supervision. To achieve this, we propose a new learnable voxel selection approach that identifies object-centric voxels only with 2D rendering supervision.Our sparse-view rendering experiments with three benchmark datasets show that proposed VS-Splat outperforms several state-of-the-art methods. We further demonstrate its effectiveness as a backbone for an existing densification method and show that anoptional extension improves its robustness to inaccurate camera pose estimates.

Problem

• Gaussian Splatting
• 3D Object Reconstruction
• Sparse-views
• Voxel Selection
• 2D Rendering Supervision

Innovation

• Voxel-Selective
• Gaussian Splatting
• End-to-end
• Sparse-view
• Learnable Voxel Selection