fuxiao0719 / panopticnerf Goto Github PK

I find this will largely enforce NeRF to converge to proper depth. While intersections are from 3D bboxes, so how do you deal with road surface? I cannot see from here. Thanks in advance : )

Thanks for your excellent work!!!

When I was processing data, I encountered the visible_id file, but the kitti-360 official website does not have this file. How should I generate it, please?

Best wishes

Thank you for this great work!

In the provided dataset, there are lidar depth files: lidar_depth/[seq_id]/*.npy
Are these files converted from the raw SICK scans in the original Kitti360 dataset?
If so, how can I convert them (from Kitti360 to the provided format)? Would you like to share sample code?
Thank you!

As titled.

能处理动态汽车吗

Thanks for your great work!
I was confused by the code on this line
The xyz is in real world space coordinate. The cft.dist denotes the distance from the origin to the furthest bounding primitive.

When I change the "batch_size" under "panopticnerf_test.yaml"——"train", an error occured.

Traceback (most recent call last):
File "train_net.py", line 95, in
main()
File "train_net.py", line 92, in main
train(cfg, network)
File "train_net.py", line 47, in train
trainer.train(epoch, train_loader, optimizer, recorder)
File "/root/data1/PanopticNeRF/lib/train/trainers/trainer.py", line 58, in train
for iteration, batch in enumerate(data_loader):
File "/opt/conda/lib/python3.7/site-packages/torch/utils/data/dataloader.py", line 521, in next
data = self._next_data()
File "/opt/conda/lib/python3.7/site-packages/torch/utils/data/dataloader.py", line 1203, in _next_data
return self._process_data(data)
File "/opt/conda/lib/python3.7/site-packages/torch/utils/data/dataloader.py", line 1229, in _process_data
data.reraise()
File "/opt/conda/lib/python3.7/site-packages/torch/_utils.py", line 434, in reraise
raise exception
KeyError: Caught KeyError in DataLoader worker process 0.
Original Traceback (most recent call last):
File "/opt/conda/lib/python3.7/site-packages/torch/utils/data/_utils/worker.py", line 287, in _worker_loop
data = fetcher.fetch(index)
File "/opt/conda/lib/python3.7/site-packages/torch/utils/data/_utils/fetch.py", line 52, in fetch
return self.collate_fn(data)
File "/opt/conda/lib/python3.7/site-packages/torch/utils/data/_utils/collate.py", line 74, in default_collate
return {key: default_collate([d[key] for d in batch]) for key in elem}
File "/opt/conda/lib/python3.7/site-packages/torch/utils/data/_utils/collate.py", line 74, in
return {key: default_collate([d[key] for d in batch]) for key in elem}
File "/opt/conda/lib/python3.7/site-packages/torch/utils/data/_utils/collate.py", line 74, in default_collate
return {key: default_collate([d[key] for d in batch]) for key in elem}
File "/opt/conda/lib/python3.7/site-packages/torch/utils/data/_utils/collate.py", line 74, in
return {key: default_collate([d[key] for d in batch]) for key in elem}
File "/opt/conda/lib/python3.7/site-packages/torch/utils/data/_utils/collate.py", line 74, in
return {key: default_collate([d[key] for d in batch]) for key in elem}
KeyError: 89

Could you please look at this problem?

Thanks for sharing your work and code!
I noticed there is center_pose in config files. I am working on a different dataset. Could you please share how you got the center_pose values in your config files? Thanks.

Thanks for your great work! I found that you used real world space coordinate instead of common NDC space. So I want to know how you supervise sky depth cause it is inf depth in real world. I found that you filted it out when training here

您好 ，目前这个工作可以用在bev图的场景分割吗？

Many thanks for this great work!

When evaluating, there are three files needed, gt_2d_semantics, gt_2d_panoptics and lidar_depth.

How could I prepare these files from the original Kitti360 data? Could there be a code release to generate these files?