SALT: A Flexible Semi-Automatic Labeling Tool for General LiDAR Point Clouds with Cross-Scene Adaptability and 4D Consistency
Preprint
Abstract
We propose a flexible Semi-Automatic Labeling Tool (SALT) for general LiDAR point clouds with cross-scene adaptability and 4D consistency. Unlike recent approaches that rely on camera distillation, SALT operates directly on raw LiDAR data, automatically generating pre-segmentation results. To achieve this, we propose a novel zero-shot learning paradigm, termed data alignment, which transforms LiDAR data into pseudo-images by aligning with the training distribution of vision foundation models. Additionally, we design a 4D-consistent prompting strategy and 4D non-maximum suppression module to enhance SAM2, ensuring high-quality, temporally consistent presegmentation. SALT surpasses the latest zero-shot methods by 18.4% PQ on SemanticKITTI and achieves nearly 40-50% of human annotator performance on our newly collected low-resolution LiDAR data and on combined data from three LiDAR types, significantly boosting annotation efficiency. We anticipate that SALT's open-sourcing will catalyze substantial expansion of current LiDAR datasets and lay the groundwork for the future development of LiDAR foundation models. Code is available at https://github.com/Cavendish518/SALT.Keywords
All Related Versions
This publication has 51 references indexed in Scilit:
- Medical SAM adapter: Adapting segment anything model for medical image segmentationMedical Image Analysis, 2025
- Deep Hierarchical Learning for 3D Semantic SegmentationInternational Journal of Computer Vision, 2025
- SFPNet: Sparse Focal Point Network for Semantic Segmentation on General LiDAR Point Clouds2024
- Better Call SAL: Towards Learning to Segment Anything in Lidar2024
- Learning to Adapt SAM for Segmenting Cross-Domain Point Clouds2024
- RAPiD-Seg: Range-Aware Pointwise Distance Distribution Networks for 3D LiDAR Segmentation2024
- SAM3D: zero-shot 3D object detection via the segment anything modelScience China Information Sciences, 2024
- HVDistill: Transferring Knowledge from Images to Point Clouds via Unsupervised Hybrid-View DistillationInternational Journal of Computer Vision, 2024
- SqueezeSegV3: Spatially-Adaptive Convolution for Efficient Point-Cloud Segmentation2020
- Deep Clustering for Unsupervised Learning of Visual Features2018