Datasets:
Update README.md
Browse files
README.md
CHANGED
|
@@ -1,4 +1,6 @@
|
|
| 1 |
-
|
|
|
|
|
|
|
| 2 |
license: cc-by-nc-sa-4.0
|
| 3 |
task_categories:
|
| 4 |
- object-detection
|
|
@@ -13,32 +15,25 @@ size_categories:
|
|
| 13 |
- 10K<n<100K
|
| 14 |
-----------------------------------------------------------------------
|
| 15 |
|
| 16 |
-
|
| 17 |
-
<u>**Accurate Eyes Detection Dataset**</u>
|
| 18 |
-
|
| 19 |
-
|
| 20 |
-
**Dataset Description**
|
| 21 |
This dataset is designed for real-time "precise eye detection" under various conditions. It contains highly accurate bounding box annotations, manually retraced to ensure maximum precision.
|
| 22 |
|
| 23 |
-
|
| 24 |
-
**Dataset Annotations Format**
|
| 25 |
- High-quality "COCO format annotations"
|
| 26 |
|
| 27 |
-
|
| 28 |
-
**Key features**
|
| 29 |
- 72,317 images of close-up eyes, people's faces, and real-life scenes.
|
| 30 |
- <u>Various Image acquisition conditions</u>:
|
| 31 |
- Single person (1 or 2 eyes), whole people, groups, ...
|
| 32 |
- Open and closed eyes.
|
| 33 |
- Diversity of lighting and poses.
|
| 34 |
|
| 35 |
-
|
| 36 |
-
|
| 37 |
-
1. <u>Preprocessing</u>:
|
| 38 |
- Auto-orientation (with EXIF orientation stripping)
|
| 39 |
- Resize to 640x640 (Stretch)
|
| 40 |
|
| 41 |
-
2. <u>Data Augmentation</u> (create 7 versions for each image):
|
| 42 |
- Random crop: 0-30% of the image
|
| 43 |
- Random rotation: ±15 degrees
|
| 44 |
- Random shear: ±15° horizontal/vertical
|
|
@@ -48,12 +43,9 @@ This dataset is designed for real-time "precise eye detection" under various con
|
|
| 48 |
- Salt and pepper noise: 0.3% of pixels
|
| 49 |
- Cutout: 7 boxes, each with a size of 2%
|
| 50 |
|
| 51 |
-
|
| 52 |
**Nota importante**: Tutti i bounding box sono stati **completamente riannotati manualmente** su Roboflow per garantire la massima accuratezza.
|
| 53 |
|
| 54 |
-
|
| 55 |
-
**Dataset Structures**
|
| 56 |
-
|
| 57 |
dataset/
|
| 58 |
├── train/
|
| 59 |
│ ├── images/
|
|
@@ -63,62 +55,51 @@ dataset/
|
|
| 63 |
│ └── annotations.json
|
| 64 |
|
| 65 |
|
|
|
|
| 66 |
|
| 67 |
-
|
| 68 |
-
|
| 69 |
-
1. <u>What you can do</u>:
|
| 70 |
- Share - copy and redistribute the material
|
| 71 |
- Adapt - remix, transform, and create derivative works
|
| 72 |
- Use for research, study, and non-commercial projects
|
| 73 |
|
| 74 |
-
2. <u>Mandatory conditions</u>:
|
| 75 |
- Attribution - You must give appropriate credit, provide a link to the license, and indicate if changes were made.
|
| 76 |
- NonCommercial - You may not use the material for commercial purposes.
|
| 77 |
- ShareAlike - If you remix, transform, or create derivative works, you must distribute your contributions under the **same license** as the original.
|
| 78 |
|
| 79 |
-
3. <u>Example of attribution</u>
|
| 80 |
-
"Accurate Eyes Detection Dataset" by AndreaPorri is licensed under CC BY-NC-SA 4.0
|
| 81 |
Dataset source: https://huggingface.co/datasets/AndreaPorri/Eyes-Detection
|
| 82 |
|
| 83 |
-
|
| 84 |
-
**Expected Uses**
|
| 85 |
- Real-time eye detection.
|
| 86 |
- Part of an Eye-Tracking pipeline for computer vision applications.
|
| 87 |
- Academic research.
|
| 88 |
- Research for medical applications.
|
| 89 |
|
| 90 |
-
|
| 91 |
-
**Limitations and Ethical Considerations**
|
| 92 |
- The dataset contains <u>images of faces</u>
|
| 93 |
- Use permitted <u>only for non-commercial purposes</u>
|
| 94 |
- It is the user's responsibility to ensure ethical use in compliance with privacy laws.
|
| 95 |
|
| 96 |
-
|
| 97 |
-
**Original Datasets of Origin**
|
| 98 |
This dataset was created based on:
|
| 99 |
1. **EyeCon Dataset** - https://app.roboflow.com/andreap/eyecon-eaux0-oykgj/1
|
| 100 |
2. **Eyes Detection Dataset** - https://app.roboflow.com/andreap/eyes_detection-bupne/1
|
| 101 |
|
|
|
|
|
|
|
| 102 |
|
| 103 |
-
|
| 104 |
-
**Citazione**
|
| 105 |
If you use this dataset in your research, please cite it as:
|
| 106 |
|
| 107 |
```bibtex
|
| 108 |
-
|
| 109 |
-
|
| 110 |
-
|
| 111 |
-
{
|
| 112 |
-
|
| 113 |
-
|
| 114 |
-
|
| 115 |
-
|
| 116 |
-
|
| 117 |
-
license = {CC BY-NC-SA 4.0},
|
| 118 |
-
url = {https://huggingface.co/datasets/AndreaPorri/Eyes-Detection}
|
| 119 |
-
}
|
| 120 |
-
|
| 121 |
-
|
| 122 |
-
|
| 123 |
-
**Disclaimer**
|
| 124 |
-
The user is responsible for the ethical and legal use of this dataset. The creator assumes no responsibility for any improper use.
|
|
|
|
| 1 |
+
<u>**Accurate Eyes Detection Dataset**</u>
|
| 2 |
+
|
| 3 |
+
------------------------------------------------------------------------
|
| 4 |
license: cc-by-nc-sa-4.0
|
| 5 |
task_categories:
|
| 6 |
- object-detection
|
|
|
|
| 15 |
- 10K<n<100K
|
| 16 |
-----------------------------------------------------------------------
|
| 17 |
|
| 18 |
+
## **Dataset Description**
|
|
|
|
|
|
|
|
|
|
|
|
|
| 19 |
This dataset is designed for real-time "precise eye detection" under various conditions. It contains highly accurate bounding box annotations, manually retraced to ensure maximum precision.
|
| 20 |
|
| 21 |
+
## **Dataset Annotations Format**
|
|
|
|
| 22 |
- High-quality "COCO format annotations"
|
| 23 |
|
| 24 |
+
## **Key features**
|
|
|
|
| 25 |
- 72,317 images of close-up eyes, people's faces, and real-life scenes.
|
| 26 |
- <u>Various Image acquisition conditions</u>:
|
| 27 |
- Single person (1 or 2 eyes), whole people, groups, ...
|
| 28 |
- Open and closed eyes.
|
| 29 |
- Diversity of lighting and poses.
|
| 30 |
|
| 31 |
+
## **Preprocessing e Augmentation applied**
|
| 32 |
+
### 1. <u>Preprocessing</u>:
|
|
|
|
| 33 |
- Auto-orientation (with EXIF orientation stripping)
|
| 34 |
- Resize to 640x640 (Stretch)
|
| 35 |
|
| 36 |
+
### 2. <u>Data Augmentation</u> (create 7 versions for each image):
|
| 37 |
- Random crop: 0-30% of the image
|
| 38 |
- Random rotation: ±15 degrees
|
| 39 |
- Random shear: ±15° horizontal/vertical
|
|
|
|
| 43 |
- Salt and pepper noise: 0.3% of pixels
|
| 44 |
- Cutout: 7 boxes, each with a size of 2%
|
| 45 |
|
|
|
|
| 46 |
**Nota importante**: Tutti i bounding box sono stati **completamente riannotati manualmente** su Roboflow per garantire la massima accuratezza.
|
| 47 |
|
| 48 |
+
## **Dataset Structures**
|
|
|
|
|
|
|
| 49 |
dataset/
|
| 50 |
├── train/
|
| 51 |
│ ├── images/
|
|
|
|
| 55 |
│ └── annotations.json
|
| 56 |
|
| 57 |
|
| 58 |
+
## **License - CC BY-NC-SA 4.0**
|
| 59 |
|
| 60 |
+
### 1. <u>What you can do</u>:
|
|
|
|
|
|
|
| 61 |
- Share - copy and redistribute the material
|
| 62 |
- Adapt - remix, transform, and create derivative works
|
| 63 |
- Use for research, study, and non-commercial projects
|
| 64 |
|
| 65 |
+
### 2. <u>Mandatory conditions</u>:
|
| 66 |
- Attribution - You must give appropriate credit, provide a link to the license, and indicate if changes were made.
|
| 67 |
- NonCommercial - You may not use the material for commercial purposes.
|
| 68 |
- ShareAlike - If you remix, transform, or create derivative works, you must distribute your contributions under the **same license** as the original.
|
| 69 |
|
| 70 |
+
### 3. <u>Example of attribution</u>:
|
| 71 |
+
"Accurate Eyes Detection Dataset" by AndreaPorri is licensed under CC BY-NC-SA 4.0
|
| 72 |
Dataset source: https://huggingface.co/datasets/AndreaPorri/Eyes-Detection
|
| 73 |
|
| 74 |
+
## **Expected Uses**
|
|
|
|
| 75 |
- Real-time eye detection.
|
| 76 |
- Part of an Eye-Tracking pipeline for computer vision applications.
|
| 77 |
- Academic research.
|
| 78 |
- Research for medical applications.
|
| 79 |
|
| 80 |
+
## **Limitations and Ethical Considerations**
|
|
|
|
| 81 |
- The dataset contains <u>images of faces</u>
|
| 82 |
- Use permitted <u>only for non-commercial purposes</u>
|
| 83 |
- It is the user's responsibility to ensure ethical use in compliance with privacy laws.
|
| 84 |
|
| 85 |
+
## **Original Datasets of Origin**
|
|
|
|
| 86 |
This dataset was created based on:
|
| 87 |
1. **EyeCon Dataset** - https://app.roboflow.com/andreap/eyecon-eaux0-oykgj/1
|
| 88 |
2. **Eyes Detection Dataset** - https://app.roboflow.com/andreap/eyes_detection-bupne/1
|
| 89 |
|
| 90 |
+
## **Disclaimer**
|
| 91 |
+
The user is responsible for the ethical and legal use of this dataset. The creator assumes no responsibility for any improper use.
|
| 92 |
|
| 93 |
+
## **Citazione**
|
|
|
|
| 94 |
If you use this dataset in your research, please cite it as:
|
| 95 |
|
| 96 |
```bibtex
|
| 97 |
+
@dataset{accurate_eyes_detection_2025,
|
| 98 |
+
author = {Andrea Porri},
|
| 99 |
+
title = {Accurate Eyes Detection Dataset},
|
| 100 |
+
year = {2025},
|
| 101 |
+
publisher = {Hugging Face},
|
| 102 |
+
version = {2.0},
|
| 103 |
+
license = {CC BY-NC-SA 4.0},
|
| 104 |
+
url = {https://huggingface.co/datasets/AndreaPorri/Eyes-Detection}
|
| 105 |
+
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|