T-3DGS: Removing Transient Objects for 3D Scene Reconstruction
This addresses the challenge of high-fidelity 3D reconstruction in real-world scenarios with moving objects, but it is incremental as it builds on existing Gaussian Splatting methods.
The paper tackled the problem of transient objects degrading 3D scene reconstructions from videos, proposing T-3DGS to filter them out, and demonstrated that it significantly outperforms state-of-the-art approaches in evaluations on video datasets.
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.