Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,101 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: apache-2.0
|
| 3 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
---
|
| 4 |
+
# AuthentiVision π
|
| 5 |
+
|
| 6 |
+
<div align="center">
|
| 7 |
+
|
| 8 |
+
<img src="assets/img_1.jpg" alt="Logo" width="300"/>
|
| 9 |
+
|
| 10 |
+
|
| 11 |
+
**State-of-the-art Face Authentication for Detecting AI-Generated Images**
|
| 12 |
+
|
| 13 |
+
[Huggingface](https://medium.com/@haijian06/authentivision-detecting-ai-generated-faces) | [Data](https://huggingface.co/datasets/haijian06/face-auth-dataset) | [Demo](https://huggingface.co/spaces/haijian06/AuthentiVision) | [Technical Paper](https://arxiv.org/abs/xxx)
|
| 14 |
+
|
| 15 |
+
</div>
|
| 16 |
+
</div>
|
| 17 |
+
|
| 18 |
+
## π― Real vs. AI-Generated Face Comparison
|
| 19 |
+
|
| 20 |
+
<div align="center">
|
| 21 |
+
<table>
|
| 22 |
+
<tr>
|
| 23 |
+
<td><b>Real Face</b></td>
|
| 24 |
+
<td><b>AI-Generated Face</b></td>
|
| 25 |
+
</tr>
|
| 26 |
+
<tr>
|
| 27 |
+
<td>
|
| 28 |
+
<img src="assets/real_face_example.jpg" alt="Real Face" width="200"/>
|
| 29 |
+
</td>
|
| 30 |
+
<td>
|
| 31 |
+
<img src="assets/ai_face.jpg" alt="AI-Generated Face" width="200"/>
|
| 32 |
+
</td>
|
| 33 |
+
</tr>
|
| 34 |
+
<tr>
|
| 35 |
+
<td>
|
| 36 |
+
<img src="assets/real_face_example.jpg" alt="Real Face" width="200"/>
|
| 37 |
+
</td>
|
| 38 |
+
<td>
|
| 39 |
+
<img src="assets/ai_face_2.jpg" alt="AI-Generated Face" width="200"/>
|
| 40 |
+
</td>
|
| 41 |
+
</tr>
|
| 42 |
+
</table>
|
| 43 |
+
</div>
|
| 44 |
+
## π Features
|
| 45 |
+
|
| 46 |
+
- High accuracy in distinguishing real faces from AI-generated ones
|
| 47 |
+
- Multiple feature extraction techniques for robust detection
|
| 48 |
+
- Easy-to-use API for quick integration
|
| 49 |
+
- Lightweight and efficient inference
|
| 50 |
+
- Comprehensive documentation and examples
|
| 51 |
+
|
| 52 |
+
## π Quick Start
|
| 53 |
+
|
| 54 |
+
```bash
|
| 55 |
+
git clone https://github.com/haijian06/AuthentiVision
|
| 56 |
+
cd AuthentiVision
|
| 57 |
+
pip install -e .
|
| 58 |
+
```
|
| 59 |
+
|
| 60 |
+
```python
|
| 61 |
+
from authentivision import AuthentiVision
|
| 62 |
+
|
| 63 |
+
# Initialize detector
|
| 64 |
+
detector = AuthentiVision()
|
| 65 |
+
|
| 66 |
+
# Make prediction
|
| 67 |
+
label, confidence = detector.predict("path_to_image.jpg")
|
| 68 |
+
print(f"Prediction: {label} (Confidence: {confidence:.2f})")
|
| 69 |
+
```
|
| 70 |
+
|
| 71 |
+
|
| 72 |
+
## π Documentation
|
| 73 |
+
|
| 74 |
+
For detailed documentation, please visit our [tech blog](https://timelabhub.github.io/).
|
| 75 |
+
|
| 76 |
+
## π― Use Cases(Coming soon)
|
| 77 |
+
|
| 78 |
+
- Identity verification systems
|
| 79 |
+
- Social media content moderation
|
| 80 |
+
- Digital forensics
|
| 81 |
+
- Security applications
|
| 82 |
+
## π License
|
| 83 |
+
|
| 84 |
+
This project is licensed under the Apache License 2.0 - see the [LICENSE](LICENSE) file for details.
|
| 85 |
+
## π Acknowledgments
|
| 86 |
+
|
| 87 |
+
- Thanks to all contributors and researchers in the field
|
| 88 |
+
- Special thanks to the open-source community
|
| 89 |
+
|
| 90 |
+
## π Citation
|
| 91 |
+
|
| 92 |
+
If you use AuthentiVision in your research or project, please cite our technical blog:
|
| 93 |
+
|
| 94 |
+
```bibtex
|
| 95 |
+
@online{authentivision2024,
|
| 96 |
+
title={AuthentiVision: Finding Yourself in the Real World},
|
| 97 |
+
author={Haijian Wang and Zhangbei Ding and Yefan Niu and Xiaoming Zhang},
|
| 98 |
+
year={2024},
|
| 99 |
+
url={https://timelabhub.github.io/},
|
| 100 |
+
note={Medium blog post}
|
| 101 |
+
}
|