different point cloud shapes

[ElsheikhAhmed]

I am having a real world data which is sometimes with (xyz) , (xyz, norm), and (xyz, norm, RGB)
i don't know how to solve this problem.

would be possiblt to make the netowrk only tkae (xyz) values and ignore the other values??

another thing that i have also some data which have less than 1024 points how can i fix this issue too ?

Thanks in advance

[ldaowen]

in the config file- just set normal to False if you want (xyz) instead of (xyz,norm)

[BelieferQAQ]

in the config file- just set normal to False if you want (xyz) instead of (xyz,norm)

hello , if my point is (x,y,z,norm,rgb) ,this size is n*9 ,would be possiblt to make the netowrk work？

[qq456cvb]

For xyz-only input, set normal: False in the config. The loader then keeps just the 3 coordinate channels and input_dim becomes 3. RGB isn't wired in by default, so for that you'd extend dataset.py to read the extra channels and bump input_dim. For clouds with fewer than num_point points, the dataset already resamples with replacement (np.random.choice(..., replace=True)), so smaller clouds get upsampled by duplication. For your own inference just pad up to num_point the same way.