Support subset for KITTI 3D format

[sooahleex]

Summary

• Support subset structure for KITTI 3d format
• Update media type from PointCloud to Image

How to test

• Add unit test

Checklist

• I have added unit tests to cover my changes.​
• I have added integration tests to cover my changes.​
• I have added the description of my changes into CHANGELOG.​
• I have updated the documentation accordingly

License

• I submit my code changes under the same MIT License that covers the project.
  Feel free to contact the maintainers if that's a concern.
• I have updated the license header for each file (see an example below).

# Copyright (C) 2024 Intel Corporation
#
# SPDX-License-Identifier: MIT

[kprokofi]

I have a quick question. Will it support "train", "val" and "test" folders right? What kind of structure is expected as input?

[sooahleex]

I have a quick question. Will it support "train", "val" and "test" folders right? What kind of structure is expected as input?
@kprokofi I added the test asset for subset like below kitti_3d_with_subset

[kprokofi]

I have a quick question. Will it support "train", "val" and "test" folders right? What kind of structure is expected as input?
@kprokofi I added the test asset for subset like below kitti_3d_with_subset

Ok, I see. Thank you! I will prepare the same structure then.

[codecov]

Codecov Report

Attention: Patch coverage is 91.89189% with 3 lines in your changes missing coverage. Please review.

Please upload report for BASE (releases/1.10.0@e96dbe2). Learn more about missing BASE report.

Files with missing lines	Patch %	Lines
...c/datumaro/plugins/data_formats/kitti_3d/format.py	83.33%	1 Missing and 1 partial ⚠️
src/datumaro/plugins/data_formats/kitti_3d/base.py	94.73%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@                Coverage Diff                 @@
##             releases/1.10.0    #1621   +/-   ##
==================================================
  Coverage                   ?   81.12%           
==================================================
  Files                      ?      281           
  Lines                      ?    32773           
  Branches                   ?     6653           
==================================================
  Hits                       ?    26587           
  Misses                     ?     4723           
  Partials                   ?     1463           

Flag	Coverage Δ
ubuntu-20.04_Python-3.10	81.10% <91.89%> (?)
windows-2022_Python-3.10	81.10% <91.89%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[kprokofi]

@wonjuleee , could you review?

[sovrasov]

Hi @wonjuleee @sooahleex is there any reason not to support mmdet's kitty layout?
https://mmdetection3d.readthedocs.io/en/v0.17.1/datasets/kitti_det.html

mmdetection3d
├── mmdet3d
├── tools
├── configs
├── data
│   ├── kitti
│   │   ├── ImageSets
│   │   ├── testing
│   │   │   ├── calib
│   │   │   ├── image_2
│   │   │   ├── velodyne
│   │   ├── training
│   │   │   ├── calib
│   │   │   ├── image_2
│   │   │   ├── label_2
│   │   │   ├── velodyne

I think many people'd prefer to use mmdet's data preparation script

[sooahleex]

Hi @wonjuleee @sooahleex is there any reason not to support mmdet's kitty layout? https://mmdetection3d.readthedocs.io/en/v0.17.1/datasets/kitti_det.html

mmdetection3d
├── mmdet3d
├── tools
├── configs
├── data
│   ├── kitti
│   │   ├── ImageSets
│   │   ├── testing
│   │   │   ├── calib
│   │   │   ├── image_2
│   │   │   ├── velodyne
│   │   ├── training
│   │   │   ├── calib
│   │   │   ├── image_2
│   │   │   ├── label_2
│   │   │   ├── velodyne

I think many people'd prefer to use mmdet's data preparation script

@sovrasov I only considered the structure of the kitti dataset like kitti-3d-object-detection-dataset, and the use case for our product is limited to monocular 3D object detection. So I did not support point clouds from the lidar sensor or any additional information, but I fully agree with our opinion that the public prefers to use the kitti layout of mmdetection. I will consider supporting this structure in the next release.