SamsungLabs / fcaf3d

[ECCV2022] FCAF3D: Fully Convolutional Anchor-Free 3D Object Detection
MIT License
231 stars 37 forks source link

output all predict bbox coordinate position #34

Closed PengboLi1998 closed 2 years ago

PengboLi1998 commented 2 years ago

Hello,I successfully run your code,but I want to output all predict bbox coordinate position.I have tried print the bbox_list in single_stage_mono3d.py,but when i run the tools/test.py ,I found it did not call the single_stage_mono3d.py.Could you please help me ?

filaPro commented 2 years ago

Hi @PengboLi1998 ,

As you use fcaf3d you should try to print boxes in SingleStageSparse3DDetector.

PengboLi1998 commented 2 years ago

Thank you for your reply.I will have a try!

PengboLi1998 commented 2 years ago

Hello,I have printed the predicted bbox coordinate ,but I found the coordinate system has changed from the original input point cloud.And what should I do to keep same with original input point cloud coordinate system.

filaPro commented 2 years ago

Actually, the point cloud as an input of the first layer and the predicted boxes are in the same coordinate system. You can check our visualization in meshlab.

Your problem may be caused by a single rotation of the ScanNet point cloud here.