offline detector for auto-labeling

  1. obtain precise dense annotations with manually provided coarse information, such as bounding boxes and extreme points.

    1. PolygonRNN 17, PolygonRNN++ 18, DEXTR [16]

  2. Several open-sourced annotation tools like CVAT as well as commercial tools (Roboflow [19], Labelbox [20]) support SAM [21] to boost the efficiency of annotating.

  4. Once Detected, Never Lost: Surpassing Human Performance in Offline LiDAR based 3D Object Detection, iccv23 oral

  5. OpenAnnotate3D: Open-Vocabulary Auto-Labeling System for Multi-modal 3D Data, 23

  7. Offboard 3D Object Detection from Point Cloud Sequences, cvpr21

  8. Auto4D: Learning to Label 4D Objects from Sequential Point Clouds, 21

  9. Autolabeling 3D Objects With Differentiable Rendering of SDF Shape Priors, cvpr19

    image + lidar

  10. Weakly Supervised 3D Object Detection from Lidar Point Cloud, eccv 20

  11. Deep Active Learning for Efficient Training of a LiDAR 3D Object Detector, IV19

  12. Leveraging Pre-Trained 3D Object Detection Models for Fast Ground Truth Generation, ICITS 18

  13. [DEXTR] Deep Extreme Cut: From Extreme Points to Object Segmentation, cvpr17

  14. Multi-label Point Cloud Annotation by Selection of Sparse Control Points, 3dv17

  15. Annotating Object Instances with a Polygon-RNN, cvpr17

  16. Efficient Interactive Annotation of Segmentation Datasets with Polygon-RNN++, cvpr18