Giters

TRI-ML / sdflabel

Official PyTorch implementation of CVPR 2020 oral "Autolabeling 3D Objects With Differentiable Rendering of SDF Shape Priors"

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

css network training data

slinghe0321 opened this issue · comments

commented

Hi, could you provide the whole dataset(8k patches image_rgb and image_uvw) for css network training? There are only 5 examples in the 'data/db/crops'. Thanks a lot.

commented

Hi there!
Unfortunately we are not able to release the training dataset since it was generated from a proprietary dataset provided by Parallel Domain. A possible alternative would be to use CARLA to extract patches and poses from there. Then using the respective car models and poses, one car render NOCS (e.g. using nocs-renderer).