Update testing dataset descriptions
Browse filesAdd the description of testing datasets.
README.md
CHANGED
@@ -1,3 +1,38 @@
|
|
1 |
---
|
2 |
license: apache-2.0
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
license: apache-2.0
|
3 |
---
|
4 |
+
|
5 |
+
We evaluate DAEFR on one synthetic dataset **CelebA-Test**, and two real-world datasets **LFW-Test** and **WIDER-Test**.
|
6 |
+
|
7 |
+
<table>
|
8 |
+
<tr>
|
9 |
+
<th>Datasets</th>
|
10 |
+
<th>Filename</th>
|
11 |
+
<th>Short Description</th>
|
12 |
+
<th>Source</th>
|
13 |
+
</tr>
|
14 |
+
<tr>
|
15 |
+
<td>CelebA-Test (HQ)</td>
|
16 |
+
<td>celeba_512_validation.zip</td>
|
17 |
+
<td>3000 (HQ) ground truth images for evaluation</td>
|
18 |
+
<td><a href="https://github.com/wzhouxiff/RestoreFormer">RestoreFormer</a></td>
|
19 |
+
</tr>
|
20 |
+
<tr>
|
21 |
+
<td>CelebA-Test (LQ)</td>
|
22 |
+
<td>self_celeba_512_v2.zip</td>
|
23 |
+
<td>3000 (LQ) synthetic images for testing</td>
|
24 |
+
<td>Ourselves</td>
|
25 |
+
</tr>
|
26 |
+
<tr>
|
27 |
+
<td>LFW-Test (LQ)</td>
|
28 |
+
<td>lfw_cropped_faces.zip</td>
|
29 |
+
<td>1711 real-world images for testing</td>
|
30 |
+
<td><a href="https://github.com/TencentARC/VQFR">VQFR</a></td>
|
31 |
+
</tr>
|
32 |
+
<tr>
|
33 |
+
<td>WIDER-Test (LQ)</td>
|
34 |
+
<td>Wider-Test.zip</td>
|
35 |
+
<td>970 real-world images for testing</td>
|
36 |
+
<td><a href="https://shangchenzhou.com/projects/CodeFormer/">CodeFormer</a></td>
|
37 |
+
</tr>
|
38 |
+
</table>
|