visualize ParisLille3D

[Neural-Net-Ninja]

Checklist

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

Describe the issue

I am trying to visualize parislille3d data, the code identifies the presence of the dataset, but does not visualise it.

Steps to reproduce the bug

import os
import glob
import open3d.ml.tf as ml3d
import open3d.ml as _ml3d
from open3d.ml.tf.models import KPFCNN
from open3d.ml.tf.pipelines import SemanticSegmentation

# construct a dataset by specifying dataset_path
dataset = ml3d.datasets.ParisLille3D(dataset_path='/home/jarvis/Open3D-ML/data/Paris_Lille')

# get the 'validation' split that combines validation
validation_split = dataset.get_split('validation')

# print the attributes of the first datum
print(validation_split.get_attr(0))

# print the shape of the first point cloud
print(validation_split.get_data(0)['point'].shape)

# show the first 100 frames using the visualizer
vis = ml3d.vis.Visualizer()
vis.visualize_dataset(dataset, 'validation', indices=range(100))

Error message

INFO - 2022-05-02 13:35:44,460 - parislille3d - Found 1 pointclouds for validation {'idx': 0, 'name': 'Lille2', 'path': '/home/jarvis/Open3D-ML/data/Paris_Lille/training_10_classes/Lille2.ply', 'split': 'validation'} INFO - 2022-05-02 13:35:45,937 - parislille3d - Found 1 pointclouds for validation (21402421, 3)

IndexError Traceback (most recent call last) /home/jarvis/Open3D-ML/paris.ipynb Cell 6' in <cell line: 15>() 13 # show the first 100 frames using the visualizer 14 vis = ml3d.vis.Visualizer() ---> 15 vis.visualize_dataset(dataset, 'validation', indices=range(100))

File ~/anaconda3/envs/interceptor/lib/python3.8/site-packages/open3d/_ml3d/vis/visualizer.py:1689, in Visualizer.visualize_dataset(self, dataset, split, indices, width, height) 1686 self.set_lut("labels", lut) 1688 self._consolidate_bounding_boxes = True -> 1689 self._init_dataset(dataset, split, indices) 1690 self._visualize("Open3D - " + dataset.name, width, height)

File ~/anaconda3/envs/interceptor/lib/python3.8/site-packages/open3d/_ml3d/vis/visualizer.py:770, in Visualizer._init_dataset(self, dataset, split, indices) 769 def _init_dataset(self, dataset, split, indices): --> 770 self._objects = DatasetModel(dataset, split, indices) 771 self._modality = dict() 772 self._modality['use_lidar'] = True

File ~/anaconda3/envs/interceptor/lib/python3.8/site-packages/open3d/_ml3d/vis/visualizer.py:292, in DatasetModel.init(self, dataset, split, indices) 290 path2idx[self._dataset.path_list[i]] = i 291 real_indices = [path2idx[p] for p in sorted(path2idx.keys())] --> 292 indices = [real_indices[idx] for idx in indices] 294 # SemanticKITTI names its items <sequence#>_<timeslice#>, 295 # "mm_nnnnnn". We'd like to use the hierarchical feature of the tree 296 # to separate the sequences. We cannot change the name in the dataset 297 # because this format is used to report algorithm results, so do it 298 # here. 299 underscore_to_slash = False

File ~/anaconda3/envs/interceptor/lib/python3.8/site-packages/open3d/_ml3d/vis/visualizer.py:292, in (.0) 290 path2idx[self._dataset.path_list[i]] = i 291 real_indices = [path2idx[p] for p in sorted(path2idx.keys())] --> 292 indices = [real_indices[idx] for idx in indices] 294 # SemanticKITTI names its items <sequence#>_<timeslice#>, 295 # "mm_nnnnnn". We'd like to use the hierarchical feature of the tree 296 # to separate the sequences. We cannot change the name in the dataset 297 # because this format is used to report algorithm results, so do it 298 # here. 299 underscore_to_slash = False

IndexError: list index out of range

Expected behavior

visualisation of data

Open3D, Python and System information

- Operating system: WSL 20.04 on windows 11
- Python version: Python 3.8.13
- Open3D version: Open3D 0.15.2
- System type: x86_64
- Is this a remote workstation?: no
- How did you install Open3D?: pip

Additional information

No response

[Neural-Net-Ninja]

pcd = o3d.io.read_point_cloud("/home/jarvis/Open3D-ML/data/Paris_Lille/training_10_classes/Lille2.ply")
print(pcd)
o3d.visualization.draw_geometries ([pcd])

Simple visualisation like this seems to work.