search

hustvl / Symphonies

[CVPR 2024] Symphonies (Scene-from-Insts): Symphonize 3D Semantic Scene Completion with Contextual Instance Queries
https://arxiv.org/abs/2306.15670
MIT License
142 stars 8 forks source link

How to organize the KITTI-360 dataset? #23

Closed 3bobo closed 1 day ago

3bobo commented 5 days ago

Thanks for your great work! I want to train your model on the KITTI-360 dataset. I download the dataset following this, and run the code match_kitti_360_stereo.py. However, it does not work. So I want to ask how to organize the KITTI-360 dataset and output their depth.

npurson commented 5 days ago

It appears that the SSCBench-KITTI-360 dataset has updated the stereo images in recent commits. As a result, you can simply follow the same pre-processing steps of stereo depth estimation for SemanticKITTI, with the exception of adapting the disparity value.

Note that the match_kitti_360_stereo.py script is intended to find matched stereo pairs, which was previously necessary when these pairs were not provided.

3bobo commented 5 days ago

Thank you for your reply! I'll take a try.