IKR-Net

IKR-Net Blind Image Superresolution

This is the repository for the following paper. Please cite this journal paper, if you use this code in your research:

Ates, H. F., Yildirim, S., & Gunturk, B. K. (2023). Deep learning-based blind image super-resolution with iterative kernel reconstruction and noise estimation. Computer Vision and Image Understanding, 233, 103718.

Required libraries

Required libraries are provided in requirements.txt file.

Test code

You can test the model (for X4 scaling and noise-free images) as follows:

./codes/python test_BLDKernet.py

Test results (PSNR, SSIM, Visuals of SR images and estimated kernels) are provided under "experiments" folder.

Trained Models

Trained model should be placed in "models" folder