search

XuyangBai / PPF-FoldNet

PyTorch reimplementation for "PPF-FoldNet: Unsupervised Learning of Rotation Invariant 3D Local Descriptors" https://arxiv.org/abs/1808.10322
99 stars 19 forks source link

Unsupported image format in Prepare the data #2

Closed roxanasoto closed 4 years ago

roxanasoto commented 4 years ago

Hello When I prepared the train data using script/fuse_fragments_3DMatch.py to fuse rgbd fragments obtained an error in console:

Start making fragments
2 scenes
  Start scene rgbd-scenes-v2-scene_05
  1 sequences
    Start seq-01
[ScalableTSDFVolume::Integrate] Unsupported image format.
[ScalableTSDFVolume::Integrate] Unsupported image format.
[ScalableTSDFVolume::Integrate] Unsupported image format.
[ScalableTSDFVolume::Integrate] Unsupported image format.
...
[ScalableTSDFVolume::Integrate] Unsupported image format.
Write PLY failed: point cloud has 0 points.
    Finished seq-01
  Finished scene rgbd-scenes-v2-scene_06
Finished making fragments

And I don't find the .ply files with the points clouds

XuyangBai commented 4 years ago

Hi, sorry for the confusion. Some of the scenes(name begin with 'analysis') in 3DMatch RGBD datasets have no color images. So the release code here is for the scene without color images, so if you use it for the scenes with color images, the image format is uncorrect. I have updated the code for both with depth and without depth situation. Try the newest version.

roxanasoto commented 4 years ago

ok, Thanks so much, I don't have any problem concerning the fusion, but when I executed the code "input_preparation.py" obtained in my console: Finish rgbd-scenes-v2-scene_05, time: 8.916854858398438e-05 I see that It didn't read point cloud and voxel downsample, neither choose reference point(or interest point) from the point cloud. Neither the steps that you specified in the README.md . I think that is because I need any file with the name .color into the file "./data/3DMatch/{scene}/seq-01".?

Also, the variable downsample isn't defined and it's generated an error when executed only "input_preparation.py" or "train.py"

An exception of type <class 'NameError'> occurred for rgbd-scenes-v2-scene_05/seq-01/cloud_bin_0, Argument: ("name 'downsample' is not defined",)

XuyangBai commented 4 years ago

Sorry. The main function of input_preparation.py is buggy, I mixed it with my previous data pipeline. I shouldn't check the filename.__contains__('color') because, after the fusion we only have a .ply and a .pose.npy for each point cloud fragment.

And also I should add one default value to the variable downsample. I will update the code later. Thank you :)