Danila Rukhovich
Danila Rukhovich
I probably don't quite understand your last question. In my understanding we need extrinsics for monocular 3d detection here as the ground truth box is parametrized by only one (not...
Please follow [#27](https://github.com/SamsungLabs/fcaf3d/issues/27#issuecomment-1100297060) .
You can follow https://github.com/chrischoy/SpatioTemporalSegmentation-ScanNet
I think, this idea is called `pseudo-lidar` and is very well explored on outdoor datasets e.g. KITTI. You can also find some works with this keyword on SUN RGB-D, but...
Hi @nss6planner , I don't see any errors here, just warnings. Can you provide the full log?
Looks like the problem not with `fcaf3d`, but with your `pytorch` installation... Can you check that `pytorch` and for example `mmcv` works fine on your machine?
For now even save and load weight layouts are inconsistent with each other:( @traveller59 , can this be fixed? Thanks in advance.
@daizhirui PR is great, thx!
KeyError: Caught KeyError in DataLoader worker process 0 when training imvoxelnet on ScanNet Dataset
Have you preprocessed scannet with this repo or with master branch of mmdetection3d? It should be this repo, as preprocessing has changed in mmdetection3d since our release.
Yes, I don't see difference between lidar2img and world2img. We only compute projection matrix [here](https://github.com/saic-vul/imvoxelnet/blob/master/mmdet3d/models/detectors/imvoxelnet.py#L115) and use it [here](https://github.com/saic-vul/imvoxelnet/blob/master/mmdet3d/models/detectors/imvoxelnet.py#L145). Just check that your points from world coordinate system are projected...