yanx27 / JS3C-Net

Sparse Single Sweep LiDAR Point Cloud Segmentation via Learning Contextual Shape Priors from Scene Completion (AAAI 2021)
MIT License
210 stars 34 forks source link

KITTI test #12

Closed Fayeben closed 2 years ago

Fayeben commented 2 years ago

Hi~

May I ask the voxels used in test for SSC need to be generated with the labels predicted by the segmentation?

yanx27 commented 2 years ago

Yes, we map the predicted labels from segmentation to the original SSC inputs.

Fayeben commented 2 years ago

Thank you for your patient reply!

Fayeben commented 2 years ago

When I use ./gen_data to generate voxels with setting.cfg, "unknown parameter: future scans" appears. How can I correctly utilize the future scans to produce voxels? Thank you !

Fayeben commented 2 years ago

Solved. It should be "past scans".