garvita-tiwari/PoseNDF

Bug in create_data.py

Alan-LanFeng opened this issue · 3 comments

sampled_pose = sampled_pose +sigma*sampled_pose
sampled_pose = sampled_pose / np.linalg.norm(sampled_pose, axis=2, keepdims=True)

I think there's something wrong with the code snippet. The two lines of code will not add any noise to the data.

L79: Adds noise to poses(in quaternion)
L80: normalizes the pose

Thanks for answering. But the sigma here is a float such as 0.01. If I understood correctly, what the code does is just multiplying sample_pose by 1.01 then normalize it.

Hi,
Thanks for pointing this out.
It should be : sampled_pose +sigma*np.random.rand(21,4)*sampled_pose