Query about test point labels for the ParisLille3D Dataset

[vijay-jaisankar]

Checklist

• I have searched for similar issues.
• I have tested with the latest development wheel.
• I have checked the release documentation and the latest documentation (for main branch).

My Question

Hello!

While loading the ParisLille3D dataset, it appears that there are no labels provided for the test point clouds. Do you know where we can find these semantic segmentation labels for the test point clouds? Thanks!

Steps to reproduce

We can observe the distribution of labels across all 10 classes of points in the val split of the dataset:

import open3d.ml.torch as ml3d  
import pandas as pd

# construct a dataset by specifying dataset_path
dataset = ml3d.datasets.ParisLille3D(dataset_path=DATASET_DIR)

# Get the validation split
val_split = dataset.get_split('val')

# Print the value counts of the label
for idx in range(len(val_split)):
    print(dict(pd.Series(val_split.get_data(idx)['label']).value_counts()))

Produces the following output:

{1: 12042099, 2: 7193259, 9: 917631, 8: 770451, 0: 179841, 5: 115885, 3: 109906, 6: 54818, 7: 11233, 4: 7298}

However, we note a "blank" output for all point clouds in the test section:

import open3d.ml.torch as ml3d  
import pandas as pd

# construct a dataset by specifying dataset_path
dataset = ml3d.datasets.ParisLille3D(dataset_path=DATASET_DIR)

# Get the test split
test_split = dataset.get_split('test')

# Print the value counts of the label
for idx in range(len(test_split)):
    print(dict(pd.Series(test_split.get_data(idx)['label']).value_counts()))

Produces the following output:

{0: 10000000}
{0: 10000000}
{0: 10000000}