Datasets:
Kurt Stolle
commited on
Commit
·
dc9db9f
1
Parent(s):
7f95801
Added data and card
Browse filesThis view is limited to 50 files because it contains too many changes.
See raw diff
- README.md +103 -0
- data/train/000000/000004.depth.png +3 -0
- data/train/000000/000004.panoptic.png +3 -0
- data/train/000000/000009.depth.png +3 -0
- data/train/000000/000009.panoptic.png +3 -0
- data/train/000000/000014.depth.png +3 -0
- data/train/000000/000014.panoptic.png +3 -0
- data/train/000000/000019.depth.png +3 -0
- data/train/000000/000019.panoptic.png +3 -0
- data/train/000000/000024.depth.png +3 -0
- data/train/000000/000024.panoptic.png +3 -0
- data/train/000000/000029.depth.png +3 -0
- data/train/000000/000029.panoptic.png +3 -0
- data/train/000001/000004.depth.png +3 -0
- data/train/000001/000004.panoptic.png +3 -0
- data/train/000001/000009.depth.png +3 -0
- data/train/000001/000009.panoptic.png +3 -0
- data/train/000001/000014.depth.png +3 -0
- data/train/000001/000014.panoptic.png +3 -0
- data/train/000001/000019.depth.png +3 -0
- data/train/000001/000019.panoptic.png +3 -0
- data/train/000001/000024.depth.png +3 -0
- data/train/000001/000024.panoptic.png +3 -0
- data/train/000001/000029.depth.png +3 -0
- data/train/000001/000029.panoptic.png +3 -0
- data/train/000002/000004.depth.png +3 -0
- data/train/000002/000004.panoptic.png +3 -0
- data/train/000002/000009.depth.png +3 -0
- data/train/000002/000009.panoptic.png +3 -0
- data/train/000002/000014.depth.png +3 -0
- data/train/000002/000014.panoptic.png +3 -0
- data/train/000002/000019.depth.png +3 -0
- data/train/000002/000019.panoptic.png +3 -0
- data/train/000002/000024.depth.png +3 -0
- data/train/000002/000024.panoptic.png +3 -0
- data/train/000002/000029.depth.png +3 -0
- data/train/000002/000029.panoptic.png +3 -0
- data/train/000003/000004.depth.png +3 -0
- data/train/000003/000004.panoptic.png +3 -0
- data/train/000003/000009.depth.png +3 -0
- data/train/000003/000009.panoptic.png +3 -0
- data/train/000003/000014.depth.png +3 -0
- data/train/000003/000014.panoptic.png +3 -0
- data/train/000003/000019.depth.png +3 -0
- data/train/000003/000019.panoptic.png +3 -0
- data/train/000003/000024.depth.png +3 -0
- data/train/000003/000024.panoptic.png +3 -0
- data/train/000003/000029.depth.png +3 -0
- data/train/000003/000029.panoptic.png +3 -0
- data/train/000004/000004.depth.png +3 -0
README.md
ADDED
@@ -0,0 +1,103 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# Cityscapes VPS
|
2 |
+
|
3 |
+
This dataset is derived from the videos in the *validation* split of the Cityscapes[^1] dataset.
|
4 |
+
It aggregates the images and metadata from Cityscapes[^1], Cityscapes-VPS[^2] and Cityscapes-DVPS[^3] into a single structured format.
|
5 |
+
This comprehensive derivative was created out of the need for a batteries-included variant of the dataset for academic purposes.
|
6 |
+
Specifically, joining samples from the individual datasets in their original structure (each is organized differently) involves a significant amount of boilerplate code.
|
7 |
+
|
8 |
+
[^1]: `Cordts et al., “The Cityscapes Dataset for Semantic Urban Scene Understanding” (CVPR 2016)`
|
9 |
+
|
10 |
+
[^2]: `Kim et al., "Video Panoptic Segmentation" (CVPR 2020)`
|
11 |
+
|
12 |
+
[^3]: `Qiao et al., "Learning Visual Perception with Depth-aware Video Panoptic Segmentation" (CVPR 2021)`
|
13 |
+
|
14 |
+
|
15 |
+
This dataset is relevant to computer vision research areas such as:
|
16 |
+
|
17 |
+
- Segmentation
|
18 |
+
- Depth estimation
|
19 |
+
- Autonomous driving
|
20 |
+
- Video understanding
|
21 |
+
|
22 |
+
## Overview
|
23 |
+
The following variables are included.
|
24 |
+
|
25 |
+
1. **Images.** The input data captured by the left camera from Cityscapes[^1], in 8-bit format. Every sequence has 30 frames.
|
26 |
+
2. **Segmentation labels.** Derived from Cityscapes[^1] and Cityscapes-DVPS[^3], these labels provide detailed semantic segmentation and instance segmentation information for 6 frames of every sequence.
|
27 |
+
3. **Depth maps.** Improved depth information from Cityscapes-DVPS[^3], offering enhanced quality over the disparity package from Cityscapes[^1], provided for the same samples as the segmentation labels above.
|
28 |
+
4. **Camera calibrations.** Includes the intrinsic and extrinsic parameters provided by Cityscapes[^1] for each sequence.
|
29 |
+
5. **Vehicle odometry.** Odometry data for each frame, a subset of those provided in Cityscapes[^1].
|
30 |
+
|
31 |
+
Files are grouped by split, sequence and frame.
|
32 |
+
This leads to the following structure:
|
33 |
+
```text
|
34 |
+
data
|
35 |
+
train
|
36 |
+
000000
|
37 |
+
000000.image.png
|
38 |
+
000000.panoptic.png
|
39 |
+
000000.depth.png
|
40 |
+
000000.vehicle.json
|
41 |
+
000000.timestamp.txt
|
42 |
+
000000.camera.json
|
43 |
+
000001.image.png
|
44 |
+
000001.panoptic.png
|
45 |
+
000001.depth.png
|
46 |
+
000001.vehicle.json
|
47 |
+
000001.timestamp.txt
|
48 |
+
000001.camera.json
|
49 |
+
...
|
50 |
+
000001
|
51 |
+
...
|
52 |
+
val
|
53 |
+
000000
|
54 |
+
...
|
55 |
+
000001
|
56 |
+
....
|
57 |
+
test
|
58 |
+
000000
|
59 |
+
...
|
60 |
+
000001
|
61 |
+
....
|
62 |
+
|
63 |
+
```
|
64 |
+
|
65 |
+
The `data` directory in this repository only contains the segmentation and depth map annotations.
|
66 |
+
The remaining data should be downloaded from official sources using the provided preparation script.
|
67 |
+
|
68 |
+
|
69 |
+
## Preparation
|
70 |
+
|
71 |
+
1. Install the [Cityscapes developer kit](https://github.com/mcordts/cityscapesScripts) using `pip`.
|
72 |
+
|
73 |
+
```bash
|
74 |
+
python -m pip install cityscapesscripts
|
75 |
+
```
|
76 |
+
|
77 |
+
2. Run the preparation script provided in this repository.
|
78 |
+
Note that this may prompt your [Cityscapes account](https://cityscapes-dataset.com/login/) login credentials.
|
79 |
+
|
80 |
+
```bash
|
81 |
+
python scripts/download.py downloads data manifest.csv
|
82 |
+
```
|
83 |
+
|
84 |
+
3. Remove the downloaded Cityscapes archive files to save disk space (optional).
|
85 |
+
|
86 |
+
```bash
|
87 |
+
rm -r downloads
|
88 |
+
```
|
89 |
+
|
90 |
+
|
91 |
+
## License
|
92 |
+
|
93 |
+
Please refer to the Cityscapes license for more details.
|
94 |
+
|
95 |
+
|
96 |
+
## Citation
|
97 |
+
|
98 |
+
If you use this dataset in your research, please cite the original
|
99 |
+
[Cityscapes](https://cityscapes-dataset.com),
|
100 |
+
[Cityscapes-VPS](https://github.com/mcahny/vps), and
|
101 |
+
[Cityscapes-DVPS](https://github.com/joe-siyuan-qiao/ViP-DeepLab) datasets.
|
102 |
+
|
103 |
+
|
data/train/000000/000004.depth.png
ADDED
Git LFS Details
|
data/train/000000/000004.panoptic.png
ADDED
Git LFS Details
|
data/train/000000/000009.depth.png
ADDED
Git LFS Details
|
data/train/000000/000009.panoptic.png
ADDED
Git LFS Details
|
data/train/000000/000014.depth.png
ADDED
Git LFS Details
|
data/train/000000/000014.panoptic.png
ADDED
Git LFS Details
|
data/train/000000/000019.depth.png
ADDED
Git LFS Details
|
data/train/000000/000019.panoptic.png
ADDED
Git LFS Details
|
data/train/000000/000024.depth.png
ADDED
Git LFS Details
|
data/train/000000/000024.panoptic.png
ADDED
Git LFS Details
|
data/train/000000/000029.depth.png
ADDED
Git LFS Details
|
data/train/000000/000029.panoptic.png
ADDED
Git LFS Details
|
data/train/000001/000004.depth.png
ADDED
Git LFS Details
|
data/train/000001/000004.panoptic.png
ADDED
Git LFS Details
|
data/train/000001/000009.depth.png
ADDED
Git LFS Details
|
data/train/000001/000009.panoptic.png
ADDED
Git LFS Details
|
data/train/000001/000014.depth.png
ADDED
Git LFS Details
|
data/train/000001/000014.panoptic.png
ADDED
Git LFS Details
|
data/train/000001/000019.depth.png
ADDED
Git LFS Details
|
data/train/000001/000019.panoptic.png
ADDED
Git LFS Details
|
data/train/000001/000024.depth.png
ADDED
Git LFS Details
|
data/train/000001/000024.panoptic.png
ADDED
Git LFS Details
|
data/train/000001/000029.depth.png
ADDED
Git LFS Details
|
data/train/000001/000029.panoptic.png
ADDED
Git LFS Details
|
data/train/000002/000004.depth.png
ADDED
Git LFS Details
|
data/train/000002/000004.panoptic.png
ADDED
Git LFS Details
|
data/train/000002/000009.depth.png
ADDED
Git LFS Details
|
data/train/000002/000009.panoptic.png
ADDED
Git LFS Details
|
data/train/000002/000014.depth.png
ADDED
Git LFS Details
|
data/train/000002/000014.panoptic.png
ADDED
Git LFS Details
|
data/train/000002/000019.depth.png
ADDED
Git LFS Details
|
data/train/000002/000019.panoptic.png
ADDED
Git LFS Details
|
data/train/000002/000024.depth.png
ADDED
Git LFS Details
|
data/train/000002/000024.panoptic.png
ADDED
Git LFS Details
|
data/train/000002/000029.depth.png
ADDED
Git LFS Details
|
data/train/000002/000029.panoptic.png
ADDED
Git LFS Details
|
data/train/000003/000004.depth.png
ADDED
Git LFS Details
|
data/train/000003/000004.panoptic.png
ADDED
Git LFS Details
|
data/train/000003/000009.depth.png
ADDED
Git LFS Details
|
data/train/000003/000009.panoptic.png
ADDED
Git LFS Details
|
data/train/000003/000014.depth.png
ADDED
Git LFS Details
|
data/train/000003/000014.panoptic.png
ADDED
Git LFS Details
|
data/train/000003/000019.depth.png
ADDED
Git LFS Details
|
data/train/000003/000019.panoptic.png
ADDED
Git LFS Details
|
data/train/000003/000024.depth.png
ADDED
Git LFS Details
|
data/train/000003/000024.panoptic.png
ADDED
Git LFS Details
|
data/train/000003/000029.depth.png
ADDED
Git LFS Details
|
data/train/000003/000029.panoptic.png
ADDED
Git LFS Details
|
data/train/000004/000004.depth.png
ADDED
Git LFS Details
|