This folder contains a reference implementation of the algorithm proposed in the following paper: Y. Li, M. Tofighi, J. Geng, V. Monga and Y. C. Eldar, "An Algorithm Unrolling Approach to Deep Blind Image Deblurring," IEEE Transactions on Image Processing, under review.
If you would like to use the code for any publications, please kindly cite the above reference.
Requirements: Python (3.7.2) PyTorch (1.0.1) Numpy (1.16.2) Scipy (1.2.1) Scikit-Image (0.14.2)
Descriptions: loader.py Data loading module, including functions for data loading, augmentations, pre-processing, etc. networks.py Defining network architecture test.py Inference module, which performs the actual work of blind deblurring parameters.py Parameter configurations, including network architectures, data paths, etc. operations.py Helper module containing miscellaneous functions, such as complex operations, scaling, padding, etc.
Instructions:
Enjoy!
Contact: If there are any questions, please contact Yuelong Li (liyuelongee@gmail.com)