search

yuxumin / PoinTr

[ICCV 2021 Oral] PoinTr: Diverse Point Cloud Completion with Geometry-Aware Transformers
MIT License
583 stars 109 forks source link

Some questions about KITTI #141

Closed wsw233 closed 5 months ago

wsw233 commented 8 months ago

Can I inference the pretrained model on pointcloud data in KITTI which have no bounding boxes(velodyne in KITTI 3d object detection)?And are the bboxs in KITTI necessary for completion?