Closed IsaacGuan closed 6 years ago
I think I have found one possible approach. With a simple tool called plyfile, we can read the points from .ply file as a numpy array. Then we simply write these arrays into HDF5 files using h5py library. This is how I did: https://github.com/IsaacGuan/PointNet-Plane-Detection/blob/master/data/write_hdf5.py
Hello All, I wanted to train PointNet for classification only. I have pointcloud data(x, y, z, i) in .bin format and corresponding labels in .txt format.
Can anyone please help me how to train PointNet with this data?
Have you solved your question? If you like, can I ask you questions My qq1922797937 thank you very much
Thank you so much for sharing the code! In a previous issue How to prepare my own data set for the segmentation training? you mentioned that we can reference the script here: https://github.com/charlesq34/3dmodel_feature/blob/master/io/write_hdf5.py to generate the HDF5 files. I looked into that code, it only takes two types of information to the HDF5 file, which are
h5_batch_data
andh5_batch_label
. However, for part segmentation we should also introduce an additional dimension which includes the part information. And actually the structure of HDF5 file for part segmentation looks like this: The view of "pid" contains the part information for each point cloud. But the problem is how can this be written to HDF5 files? As I have some ready made segmented point clouds with extra channel representing the parts in the structure of (X, Y, Z, Pid), I really want to use your implementation to explore further on my own dataset. The "write_hdf5.py" has very few comments in it, so I would be really appreciated if you could give me some hints on this issue!