How does the GT dense depth map come ?

TRAILab / CaDDN

Categorical Depth Distribution Network for Monocular 3D Object Detection (CVPR 2021 Oral)

Apache License 2.0

359 stars 62 forks source link

How does the GT dense depth map come ? #74

Closed Cc-Hy closed 2 years ago

Cc-Hy commented 2 years ago

HELLO, I noticed the GT depth maps that KITTI offers are in form of points projection, but the training depth maps used are dense images. So I wonder how to get the dense depth maps from the projection form. Thank you!

codyreading commented 2 years ago

Hello!

Please refer to the following issues: https://github.com/TRAILab/CaDDN/issues/22, https://github.com/TRAILab/CaDDN/issues/28, https://github.com/TRAILab/CaDDN/issues/43. We use a depth completion method called IP-Basic to get the dense depth images.

Cc-Hy commented 2 years ago

Thanks.

Senwang98 commented 2 years ago

@Cc-Hy Can you share the projection script? I have no idea about convert velo into depth map