Datasets:
Commit
·
bdbc530
1
Parent(s):
aa28160
Update README.md
Browse files
README.md
CHANGED
|
@@ -1,14 +1,75 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
```
|
| 2 |
---
|
| 3 |
license: apache-2.0
|
| 4 |
---
|
| 5 |
```
|
| 6 |
|
| 7 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 8 |
|
| 9 |
-
|
| 10 |
-
|
| 11 |
-
|
| 12 |
-
|
| 13 |
|
| 14 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: openrail
|
| 3 |
+
task_categories:
|
| 4 |
+
- image-to-image
|
| 5 |
+
language:
|
| 6 |
+
- en
|
| 7 |
+
tags:
|
| 8 |
+
- deepfake
|
| 9 |
+
- diffusion model
|
| 10 |
+
pretty_name: DeepFakeFace'
|
| 11 |
+
---
|
| 12 |
```
|
| 13 |
---
|
| 14 |
license: apache-2.0
|
| 15 |
---
|
| 16 |
```
|
| 17 |
|
| 18 |
+
The dataset accompanying the paper
|
| 19 |
+
"Robustness and Generalizability of Deepfake Detection: A Study with Diffusion Models".
|
| 20 |
+
|
| 21 |
+
[[Website](https://sites.google.com/view/deepfakeface/home)] [[paper](https://arxiv.org/abs/2309.02218)] [[GitHub](https://github.com/OpenRL-Lab/DeepFakeFace)].
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
### Introduction
|
| 25 |
+
|
| 26 |
+
Welcome to the **DeepFakeFace (DFF)** dataset! Here we present a meticulously curated collection of artificial celebrity faces, crafted using cutting-edge diffusion models.
|
| 27 |
+
Our aim is to tackle the rising challenge posed by deepfakes in today's digital landscape.
|
| 28 |
+
|
| 29 |
+
Here are some example images in our dataset:
|
| 30 |
+
<div align="center">
|
| 31 |
+
<img width="100%" height="auto" src="docs/images/deepfake_examples.jpg">
|
| 32 |
+
</div>
|
| 33 |
+
|
| 34 |
+
Our proposed DeepFakeFace(DFF) dataset is generated by various diffusion models, aiming to protect the privacy of celebrities.
|
| 35 |
+
There are four zip files in our dataset and each file contains 30,000 images.
|
| 36 |
+
We maintain the same directory structure as the IMDB-WIKI dataset where real images are selected.
|
| 37 |
+
|
| 38 |
+
- inpainting.zip is generated by the Stable Diffusion Inpainting model.
|
| 39 |
+
- insight.zip is generated by the InsightFace toolbox.
|
| 40 |
+
- text2img.zip is generated by Stable Diffusion V1.5
|
| 41 |
+
- wiki.zip contains original real images selected from the IMDB-WIKI dataset.
|
| 42 |
+
|
| 43 |
+
### DeepFake Dataset Compare
|
| 44 |
|
| 45 |
+
We compare our dataset with previous datasets here:
|
| 46 |
+
<div align="center">
|
| 47 |
+
<img width="100%" height="auto" src="docs/images/compare.jpg">
|
| 48 |
+
</div>
|
| 49 |
|
| 50 |
+
### Experimental Results
|
| 51 |
+
|
| 52 |
+
Performance of RECCE across different generators, measured in terms of Acc (%), AUC (%), and EER (%):
|
| 53 |
+
<div align="center">
|
| 54 |
+
<img width="78%" height="auto" src="docs/images/table1.jpg">
|
| 55 |
+
</div>
|
| 56 |
+
|
| 57 |
+
|
| 58 |
+
Robustness evaluation in terms of ACC(%), AUC (%) and EER(%):
|
| 59 |
+
<div align="center">
|
| 60 |
+
<img width="100%" height="auto" src="docs/images/table2.jpg">
|
| 61 |
+
</div>
|
| 62 |
+
|
| 63 |
+
### Cite
|
| 64 |
+
|
| 65 |
+
Please cite our paper if you use our codes or our dataset in your own work:
|
| 66 |
+
|
| 67 |
+
|
| 68 |
+
```
|
| 69 |
+
@article{song2023deepfake,
|
| 70 |
+
title={Robustness and Generalizability of Deepfake Detection: A Study with Diffusion Models},
|
| 71 |
+
author={Haixu Song, Shiyu Huang, Yinpeng Dong, Wei-Wei Tu},
|
| 72 |
+
journal={arXiv preprint arXiv:2309.02218},
|
| 73 |
+
year={2023}
|
| 74 |
+
}
|
| 75 |
+
```
|