Skip to content

Latest commit

 

History

History
172 lines (163 loc) · 6.08 KB

README.md

File metadata and controls

172 lines (163 loc) · 6.08 KB
Raw

Data Preparation

Vimeo-90K Septuplet Dataset

1. Download the original training + test set of Vimeo-septuplet (82 GB).

wget http://data.csail.mit.edu/tofu/dataset/vimeo_septuplet.zip
apt-get install unzip
unzip vimeo_septuplet.zip

This will create train and test folders in the directory of vimeo_septuplet/sequences. The folder structure is as follows:

vimeo_septuplet
├── sequences
    ├── 00001
        ├── 0266
            ├── im1.png
            ├── ...
            └── im7.png
        ├── 0268...
    ├── 00002...
├── readme.txt
├── sep_trainlist.txt
└── sep_testlist.txt

2. Split the Vimeo-septuplet into a training set and a test set (remember to configure the input and output path first!).

Please make sure you change the dataset's path to your download path in script, also you need to run for the training set and test set separately. The sep_*.txt (sep_trainlist.txt, sep_testlist.txt, sep_fast_testset.txt, sep_medium_testset.txt, sep_slow_testset.txt) can be downloaded from Link.

cd $ROOT/data
python sep_vimeo_list_Vimeo90K.py

You also need to split the test folder into fast, medium and slow subsets using sep_fast_testset.txt, sep_medium_testset.txt and sep_slow_testset.txt.

3. Generate low resolution (LR) images (remember to configure the input and output path first!):

cd $ROOT/data
python generate_LR_Vimeo90K.py

4. Create the LMDB files for faster I/O speed. Note that you need to configure your input and output path in the following script (remember to configure the input and output path first!):

cd $ROOT/data
python create_lmdb_mp_Vimeo90K_LR.py
python create_lmdb_mp_Vimeo90K_HR.py

The structure of generated lmdb folder is as follows:

Vimeo7_train.lmdb
├── data.mdb
├── lock.mdb
└── Vimeo7_train_keys.pkl

Please copy the Vimeo7_train_keys.pkl into the folder with higher level. After the aboved operations, we assume that you can get a folder with the following structure:

├── fast_of_test # For testing
│   ├── HR
│   └── LR
├── medium_of_test # For testing
│   ├── HR
│   └── LR
├── slow_of_test # For testing
│   ├── HR
│   └── LR
├── vimeo7_train_GT.lmdb # For training
│   ├── data.mdb
│   ├── lock.mdb
│   └── Vimeo7_train_keys.pkl
├── vimeo7_train_LR7.lmdb # For training
│   ├── data.mdb
│   ├── lock.mdb
│   └── Vimeo7_train_keys.pkl
└── Vimeo7_train_keys.pkl

Vid4 Dataset

You can download Vid4 dataset (vid4.tar) via Link. If you download the Vid4 dataset, you can extract the image from vid4.tar and put them into $ROOT/datasets/vid4. We assume that you can get a folder with the following structure:

vid4
├── HR
│   ├── calendar
│   │   ├── 00000000.png
│   │   ├── ...
│   │   └── ***.png
│   ├── city
│   ├── foliage
│   └── walk
└── LR
    ├── calendar
    │   ├── 00000000.png
    │   ├── ...
    │   └── ***.png
    ├── city
    ├── foliage
    └── walk

We only use this dataset for evaluation.

Adobe240fps Dataset

We recommend that you use the opencv-python with the version 4.4.0.46.

1. Download Dataset.

You can download Adobe240fps via Link(If nothing happens, you can insert this link to the address bar of the browser and press the Enter key.) and extract the video files in $ROOT/data/adobe240fps/video folder.

Then, You should download the record files: adobe240fps_folder_*.txt (adobe240fps_folder_train.txt, adobe240fps_folder_valid.txt, adobe240fps_folder_test.txt), which can be downloaded from Link for following processing operations.

2. Extract frames from the video files (remember to configure the input and output path first!).

cd $ROOT/data
python generate_frames_from_adobe240fps.py

3. Create the HR and LR image pairs (remember to configure the input and output path first!).

cd $ROOT/data
python generate_data_adobe240fps.py

4. Create LMDB files.

You should create file for following lmdb processing operation first.

cd $ROOT/data
python create_lmdb_list_adobe240fps.py

Create LMDB file.

cd $ROOT/data
python create_lmdb_mp_adobe240fps_LR.py
python create_lmdb_mp_adobe240fps_HR.py

Get pkl file.

cd $ROOT/data
sh get_adobe240fps_pkl.sh

5. Organize the files for validation and test.

We recommend that you organize the files of valid and test into the following structure:

adobe240fps
├── valid
│   ├── HR
│   │   ├── IMG_0030
│   │   │   ├── 0.png
│   │   │   ├── ...
│   │   │   └── ***.png
│   │   ├── GOPR9654a
│   │   ├── IMG_0002
│   │   └── IMG_0153
│   └── LR
│       ├── IMG_0030
│       │   ├── 0.png
│       │   ├── ...
│       │   └── ***.png
│       ├── GOPR9654a
│       ├── IMG_0002
│       └── IMG_0153
└── test
    ├── HR
    │   ├── GOPR9653
    │   │   ├── 0.png
    │   │   ├── ...
    │   │   └── ***.png
    │   ├── IMG_0001
    │   ├── IMG_0003
    │   └── IMG_0004a
    └── LR
        ├── GOPR9653
        │   ├── 0.png
        │   ├── ...
        │   └── ***.png
        ├── IMG_0001
        ├── IMG_0003
        └── IMG_0004a