Datasets:
Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,134 @@
|
|
1 |
-
|
2 |
-
|
3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# IITKGP_Fence dataset
|
2 |
+
|
3 |
+
## Overview
|
4 |
+
The **IITKGP_Fence dataset** is designed for tasks related to fence-like occlusion detection, defocus blur, depth mapping, and object segmentation. The captured data vaies in scene composition, background defocus, and object occlusions.
|
5 |
+
The dataset comprises both labeled and unlabeled data, as well as additional video and RGB-D data. The contains ground truth occlusion masks (`GT`) for the corresponding images. We created the ground truth occlusion labels in a semi-automatic way with user interaction.
|
6 |
+
|
7 |
+
|
8 |
+
## Dataset Structure
|
9 |
+
<details>
|
10 |
+
<summary>Click to expand directory structure</summary>
|
11 |
+
|
12 |
+
<pre style="font-size: 12px;">
|
13 |
+
ββββIITKGP_Fence dataset
|
14 |
+
ββββLabeled
|
15 |
+
β ββββGT
|
16 |
+
β β (1).png
|
17 |
+
β β ...
|
18 |
+
β β (175).png
|
19 |
+
β β
|
20 |
+
β ββββImgs
|
21 |
+
β β (1).png
|
22 |
+
β β ...
|
23 |
+
β β (175).png
|
24 |
+
β β
|
25 |
+
β ββββOthers
|
26 |
+
β ββββBird
|
27 |
+
β β c1.png
|
28 |
+
β β ...
|
29 |
+
β β f1.png
|
30 |
+
β β
|
31 |
+
β ββββBlurred_bg
|
32 |
+
β β ββββGT
|
33 |
+
β β β gt1.jpg
|
34 |
+
β β β ...
|
35 |
+
β β β gt88.jpg
|
36 |
+
β β β
|
37 |
+
β β ββββImgs
|
38 |
+
β β im (1).jpg
|
39 |
+
β β ...
|
40 |
+
β β im (88).jpg
|
41 |
+
β β
|
42 |
+
β ββββ...
|
43 |
+
ββββUnlabeled
|
44 |
+
ββββDefocusBlurred
|
45 |
+
β ββββBlurred_fg
|
46 |
+
β β BlMov_01.mov
|
47 |
+
β β ...
|
48 |
+
β β BlMov_46.mov
|
49 |
+
β β
|
50 |
+
β ββββImgs
|
51 |
+
β BlJPG_01.jpg
|
52 |
+
β ...
|
53 |
+
ββββRGB
|
54 |
+
β ββββImgs
|
55 |
+
β β BlJPG_001.jpg
|
56 |
+
β β ...
|
57 |
+
β β BlJPG_205.jpg
|
58 |
+
β β
|
59 |
+
β ββββVids
|
60 |
+
β ββββutils
|
61 |
+
β ββββZoo
|
62 |
+
β Zoo_001.mp4
|
63 |
+
β ...
|
64 |
+
β Zoo_205.mp4
|
65 |
+
ββββRGBD
|
66 |
+
β data1.mat
|
67 |
+
β ...
|
68 |
+
ββββOthersamples
|
69 |
+
β ββββDATA1
|
70 |
+
β ββββ...
|
71 |
+
ββββutils
|
72 |
+
</pre>
|
73 |
+
|
74 |
+
</details>
|
75 |
+
|
76 |
+
### Dataset Description
|
77 |
+
Here's an overview of its structure and contents:
|
78 |
+
|
79 |
+
1. **Labeled Data**:
|
80 |
+
- **GT (Ground Truth)**: Contains 175 PNG images representing the ground truth labels for corresponding input images.
|
81 |
+
- **Imgs**: Contains 175 PNG images, which are RGB images that correspond to the ground truth.
|
82 |
+
- **Others**:
|
83 |
+
- **Multiple Scenes**: Contains various scenes (e.g., `Bird/`, `Tennis/`, etc.). Each scene consists of four pairs of RGB images and their corresponding ground truth masks.
|
84 |
+
- **Blurred_bg**: This folder includes images with blurred backgrounds and corresponding ground truth occlusion segmentation labels.
|
85 |
+
- **GT**: Contains 88 ground truth occlusion mask.
|
86 |
+
- **Imgs**: Contains 88 blurred background images in JPG format.
|
87 |
+
|
88 |
+
2. **Unlabeled Data**:
|
89 |
+
- **DefocusBlurred**: Focused on data related to blurred foreground occlusions.
|
90 |
+
- **Blurred_fg**: Contains 46 video files and `Imgs/'.
|
91 |
+
- **RGB**: Contains regular RGB images and videos.
|
92 |
+
- **Imgs**: 205 JPEG images of various scenes with occlusions.
|
93 |
+
- **Vids**: Includes a total of 214 video files.
|
94 |
+
- **RGBD**: Contains data for scenes with RGB images and depth data.
|
95 |
+
- **MAT files**: These files store all the data values and additional camera information for various samples.
|
96 |
+
- **Othersamples**: Includes additional data samples captured in laboratory.
|
97 |
+
|
98 |
+
### Key Dataset Features:
|
99 |
+
- **Fence Detection**: Designed for detecting fences or fence-like structures that might occlude objects.
|
100 |
+
- **Defocus Blur**: Contains images and videos with blurred objects, likely to challenge detection and segmentation algorithms.
|
101 |
+
- **RGBD Data**: Offers depth information alongside RGB images, which can be used for tasks like 3D reconstruction or occlusion handling.
|
102 |
+
- **Unlabeled and Labeled Data**: Facilitates both supervised and unsupervised learning tasks. The `Labeled` <a href="https://huggingface.co/datasets/NeuroVizv0yaZ3R/IITKGP_Fence_dataset/tree/main/Labeled" style="color: darkorange;">folder</a> data provides ground truth occlusion masks, while the `Unlabeled` <a href="https://huggingface.co/datasets/NeuroVizv0yaZ3R/IITKGP_Fence_dataset/tree/main/Unlabeled" style="color: teal;">folder</a> data allows for further experimentation or self-supervised methods.
|
103 |
+
|
104 |
+
### Dataset Repository
|
105 |
+
|
106 |
+
- **GitHub Repository:** <a href="https://github.com/Moushumi9medhi/Occlusion-Removal" style="color: magenta;">Occlusion-Removal</a>
|
107 |
+
- **Paper:** [https://arxiv.org/abs/placeholder]
|
108 |
+
- **Authors:** [Sankaraganesh Jonna](https://www.linkedin.com/in/ganeshjonna/), Moushumi Medhi, [Rajiv Ranjan Sahay](https://www.iitkgp.ac.in/department/EE/faculty/ee-rajiv)
|
109 |
+
|
110 |
+
## Usage
|
111 |
+
```bash
|
112 |
+
pip install datasets
|
113 |
+
|
114 |
+
from datasets import load_dataset
|
115 |
+
dataset = load_dataset('NeuroVizv0yaZ3R/IITKGP_Fence_dataset')
|
116 |
+
```
|
117 |
+
|
118 |
+
## Citation
|
119 |
+
|
120 |
+
**BibTeX:**
|
121 |
+
Will be updated soon!
|
122 |
+
|
123 |
+
## License
|
124 |
+
|
125 |
+
For license details, refer to the [LICENSE](https://huggingface.co/datasets/choosealicense/licenses/blob/main/markdown/apache-2.0.md) file.
|
126 |
+
|
127 |
+
## Acknowledgments
|
128 |
+
|
129 |
+
We would like to express our gratitude to **Dr. Sreeja S R** for her valuable contributions to the [Imgs folder](https://huggingface.co/datasets/NeuroVizv0yaZ3R/IITKGP_Fence_dataset/tree/main/Unlabeled/RGB/Imgs) within the RGB directory of the Unlabeled dataset, which greatly enriched its diversity and quality.
|
130 |
+
|
131 |
+
|
132 |
+
## Contact
|
133 |
+
[[email protected]](mailto:[email protected])
|
134 |
+
|