This is the code of Edge-Based Image Restoration, implemented by lhaof.
If you use this code, you need to cite the following paper
@article{rares2005edge-based,
author = {Rares, Andrei and Reinders, Marcel and Biemond, Jan},
year = {2005},
month = {01},
pages = {1454-1468},
title = {Edge-Based Image Restoration.},
volume = {14},
journal = {IEEE Transactions on Image Processing}
}
I implemented this code when I was submitting my paper 'CASI: Context-Aware Semantic Inpainting'. Because I was required to compare my work with this algorithm 'Edge-Based Image Restoration'. Anyway, you may kindly cite my work below
@article{li2018context-aware,
title={Context-Aware Semantic Inpainting},
author={Li, Haofeng and Li, Guanbin and Lin, Liang and Yu, Hongchuan and Yu, Yizhou},
journal={IEEE Transactions on Systems, Man, and Cybernetics},
pages={1--14},
year={2018}
}
As I remember, 'imrestore.py' works for gray-scale images while 'rgbrestore.py' deals with RGB images. You need to install
Python
Matlab
matlab.engine for python (then you can call matlab function within a python script.)
Notice that the code can only recover a central square region in an image. If you want to process arbitrary corrupt regions, you should better re-implement the algorithm. Good luck and have fun.