We have provided pre-trained RDN model [1] and RDN-PNB model for image denoising tasks with noise level 50. The test dataset Set12, BSD68 and Urban100 are also included in this demo.

• RDN architecture [1]

• Our proposed pyramid non-local blocks(PNB) can be easily inserted into RDN to enhance the ability to capture long-range contextual information.

[1] Y. Zhang, Y. Tian, Y. Kong, B. Zhong, and Y. Fu, “Residual dense network for image restoration,” IEEE Trans. Pattern Anal. Mach. Intell. 2020.

The pre-trained checkpoint can be downloaded at Google Drive. Please download them and unzip to the corresponding directory.

.
├── checkpoints
│   ├── RDN
│   ├── RDN_PNB
├── TestData
│   ├── Set12
│   ├── BSD68
│   ├── Urban100
├── test.sh
├── test.py
├── model.py
├── module.py
├── README.md
├── figures

• Our code is based on Tensorflow v1.14 in Python 3.6.
• Our code has been tested on Ubuntu 16.04 and Mac.
• CPU is supported. But GPU is preferred for training.

sh test.sh

If you find this benchmark helpful for your research, please consider citing:

@article{
}