去除视频中动态物体干扰，提升3D重建精度
T-3DGS: Removing Transient Objects for 3D Scene Reconstruction
- 利用重建过程中的训练动态差异，无监督区分动态与静态物体
- 结合分割模型与双向追踪，精准定位并保持动态物体边界一致性
- 在稀疏与密集采集数据上均优于现有方法，适合真实复杂场景
视频序列中的瞬时物体严重降低3D场景重建质量。为此，我们提出T-3DGS框架，通过高斯点阵（Gaussian Splatting）实现对瞬时干扰物的鲁棒过滤。该框架包含两步：首先，采用无监督分类网络，基于重建过程中动态与静态元素的训练差异识别瞬时物体；其次，通过集成现成分割方法与双向追踪模块，进一步优化初始检测结果，提升边界精度与时间一致性。在稀疏与密集采集的视频数据集上的评估表明，T-3DGS显著优于当前最优方法，可在真实复杂场景中实现高保真3D重建。
原文摘要 · Abstract (English)
Transient objects in video sequences can significantly degrade the quality of 3D scene reconstructions. To address this challenge, we propose T-3DGS, a novel framework that robustly filters out transient distractors during 3D reconstruction using Gaussian Splatting. Our framework consists of two steps. First, we employ an unsupervised classification network that distinguishes transient objects from static scene elements by leveraging their distinct training dynamics within the reconstruction process. Second, we refine these initial detections by integrating an off-the-shelf segmentation method with a bidirectional tracking module, which together enhance boundary accuracy and temporal coherence. Evaluations on both sparsely and densely captured video datasets demonstrate that T-3DGS significantly outperforms state-of-the-art approaches, enabling high-fidelity 3D reconstructions in challenging, real-world scenarios.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。