Closed huixiancheng closed 3 years ago
The original KITTI dataset provides remission for each lidar point. (The data is stored as x,y,z, remission).
We are currently working on a way to also provide image-based labels, but this is still in the early stages.
You should have a look at KITTI360 (http://www.cvlibs.net/datasets/kitti-360/index.php). On the semantic segmentation benchmark of KITTI, http://www.cvlibs.net/datasets/kitti/eval_semantics.php, there are also annotations from other researchers.
Thanks, I'll try!Tribute to your team's work!
As shown in the question.It's look like don't have. Furthermore,will your team have the plan to annotation the image? Since the KITTI dataset don't have big 2D sem.seg dataset.Ony one is a small dataset with 200 pics as I know. looking forward to reply and thank your team’s for the great work!