I've an h5 file which has [r,g,b,x,y,z] information for each point in the point cloud. The point cloud has labels stored in a .bin file. Each row has a corresponding label. This is a classification problem.
I'm stuck at data processing stage. How to train/test split 700k rows? Also, this is one file, I have to include other h5 files also for training. Whats the best approach to such a problem? I'm trying to run PointNet++ with this data to get classification.
Asked
Active
Viewed 78 times
1
h612
- 544
- 2
- 11