Commit
·
826870d
1
Parent(s):
0692298
Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,77 @@
|
|
1 |
---
|
2 |
license: apache-2.0
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
license: apache-2.0
|
3 |
---
|
4 |
+
# MFQEv2 Dataset
|
5 |
+
|
6 |
+
For some video enhancement/restoration tasks, lossless reference videos are necessary.
|
7 |
+
|
8 |
+
We open-source the dataset used in our [MFQEv2 paper](https://arxiv.org/abs/1902.09707), which includes 108 lossless YUV videos for training and 18 test videos recommended by [ITU-T](https://ieeexplore.ieee.org/document/6317156).
|
9 |
+
|
10 |
+
## 1. Content
|
11 |
+
|
12 |
+
- 108 lossless YUV videos for training.
|
13 |
+
- 18 lossless YUV videos for test, recommended by ITU-T.
|
14 |
+
- An HEVC compression tool box.
|
15 |
+
|
16 |
+
43.1 GB in total.
|
17 |
+
|
18 |
+
## 2. Download Raw Videos
|
19 |
+
|
20 |
+
[[Dropbox]](https://www.dropbox.com/sh/tphdy1lmlpz7zq3/AABR4Qim-P-3xGtouWk6ohi5a?dl=0)
|
21 |
+
|
22 |
+
or [[百度网盘 (key: mfqe)]](https://pan.baidu.com/s/1oBZf75bFGRanLmQQLAg4Ew)
|
23 |
+
|
24 |
+
## 3. Compress Videos
|
25 |
+
|
26 |
+
We compress both training and test videos by [HM](https://hevc.hhi.fraunhofer.de/) 16.5 at low delay P (LDP) mode with QP=37. The video compression toolbox is provided at the dataset folder.
|
27 |
+
|
28 |
+
We will get:
|
29 |
+
|
30 |
+
```tex
|
31 |
+
MFQEv2_dataset/
|
32 |
+
├── train_108/
|
33 |
+
│ ├── raw/
|
34 |
+
│ └── HM16.5_LDP/
|
35 |
+
│ └── QP37/
|
36 |
+
├── test_18/
|
37 |
+
│ ├── raw/
|
38 |
+
│ └── HM16.5_LDP/
|
39 |
+
│ └── QP37/
|
40 |
+
├── video_compression/
|
41 |
+
│ └── ...
|
42 |
+
└── README.md
|
43 |
+
```
|
44 |
+
|
45 |
+
### Ubuntu
|
46 |
+
|
47 |
+
1. `cd video_compression/`
|
48 |
+
2. Edit `option.yml`.
|
49 |
+
3. `chmod +x TAppEncoderStatic`
|
50 |
+
4. `python unzip_n_compress.py`
|
51 |
+
|
52 |
+
### Windows
|
53 |
+
|
54 |
+
1. Unzip `train_108.zip` and `test_18.zip` manually!
|
55 |
+
2. `cd video_compression\`
|
56 |
+
3. Edit `option.yml` (e.g., `system: windows`).
|
57 |
+
4. `python unzip_n_compress.py`
|
58 |
+
|
59 |
+
## 4. Citation
|
60 |
+
|
61 |
+
If you find this helpful, please star and cite:
|
62 |
+
|
63 |
+
```tex
|
64 |
+
@article{2019xing,
|
65 |
+
doi = {10.1109/tpami.2019.2944806},
|
66 |
+
url = {https://doi.org/10.1109%2Ftpami.2019.2944806},
|
67 |
+
year = 2021,
|
68 |
+
month = {mar},
|
69 |
+
publisher = {Institute of Electrical and Electronics Engineers ({IEEE})},
|
70 |
+
volume = {43},
|
71 |
+
number = {3},
|
72 |
+
pages = {949--963},
|
73 |
+
author = {Zhenyu Guan and Qunliang Xing and Mai Xu and Ren Yang and Tie Liu and Zulin Wang},
|
74 |
+
title = {{MFQE} 2.0: A New Approach for Multi-Frame Quality Enhancement on Compressed Video},
|
75 |
+
journal = {{IEEE} Transactions on Pattern Analysis and Machine Intelligence}
|
76 |
+
}
|
77 |
+
```
|