google-research / kubric

A data generation pipeline for creating semi-realistic synthetic multi-object videos with rich annotations such as instance segmentation masks, depth maps, and optical flow.
Apache License 2.0
2.34k stars 233 forks source link

How was the depth map rendered in the dataset? #295

Open z-yf17 opened 1 year ago

z-yf17 commented 1 year ago

Hi, I’m interested in the depth map of the dataset and I have a question about it. I’m wondering if the author used a linear transformation to convert the depth values to a more interpretable range. If so, could you please provide more details about the transformation? This is important for me because I want to reconstruct point clouds from the depth maps. Thank you!