Data preprocessing: Unsigned Distance

[stalhabukhari]

The paper mentions the use of signed distance for each object. However, the preprocessing code makes use of unsigned distance. Is this correct? (absolute=True is used)

grasp_diffusion/scripts/create_data/generate_mesh_sdf.py

Lines 43 to 58 in 59b633c

	def generate_mesh_sdf(mesh, absolute=True, normalize=False, n_points=200000):
	print (mesh)
	q_sdf, pcl = sample_sdf_near_surface(mesh, number_of_points=n_points, return_gradients=False)
	query_points, sdf = q_sdf[0], q_sdf[1]
	if absolute:
	neg_sdf_idxs = np.argwhere(sdf<0)[:,0]
	sdf[neg_sdf_idxs] = -sdf[neg_sdf_idxs]
	if normalize:
	sdf_max = sdf.max()
	sdf_min = sdf.min()
	sdf = (sdf - sdf_min) / (sdf_max - sdf_min)
	return query_points, sdf