Skip to content

Latest commit

 

History

History
119 lines (103 loc) · 3.65 KB

datasets_restoration.md

File metadata and controls

119 lines (103 loc) · 3.65 KB

Prepare Datasets for Restoration

Image Super-Resolution

  • Training dataset: DIV2K dataset.
  • Validation dataset: Set5 and Set14.

It is recommended to symlink the dataset root to $MMEditing/data:

mmediting
├── mmedit
├── tools
├── configs
├── data
│   ├── DIV2K
│   │   ├── DIV2K_train_HR
│   │   ├── DIV2K_train_LR_bicubic
│   │   │   ├── X2
│   │   │   ├── X3
│   │   │   ├── X4
│   │   ├── DIV2K_valid_HR
│   │   ├── DIV2K_valid_LR_bicubic
│   │   │   ├── X2
│   │   │   ├── X3
│   │   │   ├── X4
│   ├── val_set5
│   │   ├── Set5_bicLRx2
│   │   ├── Set5_bicLRx3
│   │   ├── Set5_bicLRx4
│   ├── val_set14
│   │   ├── Set14_bicLRx2
│   │   ├── Set14_bicLRx3
│   │   ├── Set14_bicLRx4

If your folder structure is different, you may need to change the corresponding paths in config files.

Crop sub-images

For faster IO, we recommend to crop the DIV2K images to sub-images. We provide such a script:

python tools/preprocess_div2k_dataset.py --data-root ./data/DIV2K

The generated data is stored under DIV2K and the data structure is as follows, where _sub indicates the sub-images.

mmediting
├── mmedit
├── tools
├── configs
├── data
│   ├── DIV2K
│   │   ├── DIV2K_train_HR
│   │   ├── DIV2K_train_HR_sub
│   │   ├── DIV2K_train_LR_bicubic
│   │   │   ├── X2
│   │   │   ├── X3
│   │   │   ├── X4
│   │   │   ├── X2_sub
│   │   │   ├── X3_sub
│   │   │   ├── X4_sub
│   │   ├── DIV2K_valid_HR
│   │   ├── ...
...

Prepare annotation list

If you use the annotation mode for the dataset, you first need to prepare a specific txt file.

Each line in the annotation file contains the image names and image shape (usually for the ground-truth images), separated by a white space.

Example of an annotation file:

0001_s001.png (480,480,3)
0001_s002.png (480,480,3)

Video Super-Resolution

Note that we merge train and val datasets in REDS for easy switching between REDS4 partition (used in EDVR) and the official validation partition.
The original val dataset (clip names from 000 to 029) are modified to avoid conflicts with training dataset (total 240 clips). Specifically, the clip names are changed to 240, 241, ... 269.

You can prepare the REDS dataset by running:

python tools/preprocess_reds_dataset.py ./data/REDS

It is also recommended to symlink the dataset root to $MMEditing/data:

mmediting
├── mmedit
├── tools
├── configs
├── data
│   ├── REDS
│   │   ├── train_sharp
│   │   │   ├── 000
│   │   │   ├── 001
│   │   │   ├── ...
│   │   ├── train_sharp_bicubic
│   │   │   ├── 000
│   │   │   ├── 001
│   │   │   ├── ...
│   ├── REDS4
│   │   ├── GT
│   │   ├── sharp_bicubic

Prepare LMDB dataset

If you want to use LMDB datasets for faster IO speed, you can make LMDB files by:

# DIV2K
python tools/preprocess_div2k_dataset.py --data-root ./data/DIV2K --make-lmdb
# REDS
python tools/preprocess_reds_dataset.py --root-path ./data/REDS --make-lmdb