FreeMask3D: Zero-Shot Point Cloud Instance Segmentation Without 3D Training

doi:10.1109/LRA.2025.3621977

	FreeMask3D: Zero-Shot Point Cloud Instance Segmentation Without 3D Training
	Zhou, Mingquan 1,2; Wu, Xiaodong 1,2; He, Chen 1,2; Wang, Ruiping 1,2; Chen, Xilin 1,2
	2025-12-01
发表期刊	IEEE ROBOTICS AND AUTOMATION LETTERS
ISSN	2377-3766
卷号	10 期号:12 页码:12301-12308
摘要	Point cloud instance segmentation is crucial for 3D scene understanding in robotics. However, existing methods heavily rely on learning-based approaches that require large amounts of annotated 3D data, resulting in high annotation costs. Therefore, developing cost-effective and data-efficient solutions is essential. To this end, we propose FreeMask3D, a novel approach that achieves 3D point cloud instance segmentation without requiring any 3D annotation or additional training. Our method consists of two main steps: instance localization and instance recognition. For instance localization, we leverage pre-trained 2D instance segmentation models to perform instance segmentation on corresponding RGB-D images. These results are then mapped to 3D space and fused across frames to generate the final 3D instance masks. For instance recognition, the OpenSem module infers the category of each instance by leveraging the generalization capabilities of cross-modal large models, such as CLIP, to enable open-vocabulary semantic recognition. Experiments and ablation studies on four challenging benchmarks-ScanNetv2, ScanNet200, S3DIS, and Replica-demonstrate that FreeMask3D achieves competitive or superior performance compared to state-of-the-art methods, despite without 3D supervision. Qualitative results highlight its open-vocabulary capabilities based on color, affordance, or uncommon phrase description.
关键词	Three-dimensional displays Instance segmentation Point cloud compression Semantics Training Solid modeling Annotations Visualization Cameras Adaptation models Object detection segmentation and categorization deep learning for visual perception embodied cognitive science
DOI	10.1109/LRA.2025.3621977
收录类别	SCI
语种	英语
资助项目	National Key R&D Program of China[2021ZD0111901] ; National Key R&D Program of China[2023YFF1105104] ; Natural Science Foundation of China[62495082] ; Natural Science Foundation of China[62461160331] ; Natural Science Foundation of China[U21B2025]
WOS研究方向	Robotics
WOS类目	Robotics
WOS记录号	WOS:001600704200005
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
引用统计
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/41579
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Wang, Ruiping
作者单位	1.Chinese Acad Sci, Inst Comp Technol, Key Lab AI Safety, Beijing 100190, Peoples R China 2.Univ Chinese Acad Sci, Beijing 100049, Peoples R China
推荐引用方式 GB/T 7714	Zhou, Mingquan,Wu, Xiaodong,He, Chen,et al. FreeMask3D: Zero-Shot Point Cloud Instance Segmentation Without 3D Training[J]. IEEE ROBOTICS AND AUTOMATION LETTERS,2025,10(12):12301-12308.
APA	Zhou, Mingquan,Wu, Xiaodong,He, Chen,Wang, Ruiping,&Chen, Xilin.(2025).FreeMask3D: Zero-Shot Point Cloud Instance Segmentation Without 3D Training.IEEE ROBOTICS AND AUTOMATION LETTERS,10(12),12301-12308.
MLA	Zhou, Mingquan,et al."FreeMask3D: Zero-Shot Point Cloud Instance Segmentation Without 3D Training".IEEE ROBOTICS AND AUTOMATION LETTERS 10.12(2025):12301-12308.