super_resolution.md

Super Resolution from a Single Image

DIV2K dataset

DIV2K dataset: DIVerse 2K resolution high quality images as used for the NTIRE challenge on super-resolution @ CVPR 2017

Download and organize data like:

tf_estimator_barebone/data/DIV2K/
├── DIV2K_train_HR
├── DIV2K_train_LR_bicubic
│   └── X2
│   └── X3
│   └── X4
├── DIV2K_valid_HR
└── DIV2K_valid_LR_bicubic
    └── X2
    └── X3
    └── X4

Dependencies

conda install tensorflow-gpu pillow

Inference

python -m datasets.div2k --model-dir MODEL_DIR --input-dir INPUT_DIR --output-dir OUTPUT_DIR

Performance

Compare with WDSR (PyTorch-based)

Networks	Parameters	DIV2K (val) PSNR	Pre-trained models	Training command
EDSR[1] Baseline	1,191,324	34.63	Download	details python trainer.py --dataset div2k --model edsr --job-dir ./div2k_edsr
WDSR[2] Baseline	1,190,100	34.78	Download	details python trainer.py --dataset div2k --model wdsr --job-dir ./div2k_wdsr

References

[1] Bee Lim, Sanghyun Son, Heewon Kim, Seungjun Nah, and Kyoung Mu Lee, "Enhanced Deep Residual Networks for Single Image Super-Resolution," 2nd NTIRE: New Trends in Image Restoration and Enhancement workshop and challenge on image super-resolution in conjunction with CVPR 2017. [PDF] [arXiv] [Slide]

[2] Jiahui Yu, Yuchen Fan, Jianchao Yang, Ning Xu, Zhaowen Wang, Xinchao Wang, Thomas Huang, "Wide Activation for Efficient and Accurate Image Super-Resolution", arXiv preprint arXiv:1808.08718. [arXiv] [Code]