用视觉模型自动标注3D框，无需反复训练
OpenBox: Annotate Any Bounding Boxes in 3D
- 先对齐2D图像与3D点云，再按物体状态生成适配框
- 在三个数据集上精度更高，速度更快，无需自训练
- 适合自动驾驶场景中低成本标注新物体
无监督且开放词汇的3D目标检测近年受到关注，尤其在自动驾驶领域，降低标注成本和识别未见物体对安全与可扩展性至关重要。然而，现有方法通常统一标注3D边界框，忽略物体物理状态，且需多次自训练迭代进行标注优化，导致质量不佳且计算开销大。为此，我们提出OpenBox，一个两阶段自动标注流程，利用2D视觉基础模型。第一阶段通过跨模态实例对齐，将视觉基础模型处理的2D图像中的实例线索关联至对应的3D点云。第二阶段根据刚性和运动状态分类实例，并基于类别特定尺寸统计生成自适应边界框。结果表明，OpenBox无需自训练即可生成高质量3D边界框标注。在Waymo Open Dataset、Lyft Level 5 Perception数据集及nuScenes数据集上的实验显示，其性能优于基线方法，兼具更高准确率与效率。
原文摘要 · Abstract (English)
Unsupervised and open-vocabulary 3D object detection has recently gained attention, particularly in autonomous driving, where reducing annotation costs and recognizing unseen objects are critical for both safety and scalability. However, most existing approaches uniformly annotate 3D bounding boxes, ignore objects' physical states, and require multiple self-training iterations for annotation refinement, resulting in suboptimal quality and substantial computational overhead. To address these challenges, we propose OpenBox, a two-stage automatic annotation pipeline that leverages a 2D vision foundation model. In the first stage, OpenBox associates instance-level cues from 2D images processed by a vision foundation model with the corresponding 3D point clouds via cross-modal instance alignment. In the second stage, it categorizes instances by rigidity and motion state, then generates adaptive bounding boxes with class-specific size statistics. As a result, OpenBox produces high-quality 3D bounding box annotations without requiring self-training. Experiments on the Waymo Open Dataset, the Lyft Level 5 Perception dataset, and the nuScenes dataset demonstrate improved accuracy and efficiency over baselines.
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。