Reconstructing and decomposing dynamic urban scenes is crucial for autonomous driving, urban planning, and scene editing. However, existing methods fail to perform instance-aware decomposition without manual annotations, which is crucial for instance-level scenethis http URLpropose UnIRe, a 3D Gaussian Splatting (3DGS) based approach that decomposes a scene into a static background and individual dynamic instances using only RGB images and LiDAR point clouds. At its core, we introduce 4D superpoints, a novel representation that clusters multi-frame LiDAR points in 4D space, enabling unsupervised instance separation based on spatiotemporal correlations. These 4D superpoints serve as the foundation for our decomposed 4D initialization, i.e., providing spatial and temporal initialization to train a dynamic 3DGS for arbitrary dynamic classes without requiring bounding boxes or objectthis http URL, we introduce a smoothness regularization strategy in both 2D and 3D space, further improving the temporalthis http URLon benchmark datasets show that our method outperforms existing methods in decomposed dynamic scene reconstruction while enabling accurate and flexible instance-level editing, making it a practical solution for real-world applications.
View on arXiv@article{mao2025_2504.00763,
title={ UnIRe: Unsupervised Instance Decomposition for Dynamic Urban Scene Reconstruction },
author={ Yunxuan Mao and Rong Xiong and Yue Wang and Yiyi Liao },
journal={arXiv preprint arXiv:2504.00763},
year={ 2025 }
}