Our in-painting paper for the 2011 Computational Intelligence Lab course.
The paper is found here: https://github.com/leobuettiker/CIL-2011-Inpainting-by-Sensitive-Mean-Filtering/blob/master/report.pdf?raw=true
The paper is by no means extraordinary or especially good. But we think, our method is a good baseline that might help others to get started.
The code for our main approach is found in https://github.com/leobuettiker/CIL-2011-Inpainting-by-Sensitive-Mean-Filtering/tree/master/code. While the final code is in the file notSoBaselineAnymoreFast.m, the same result is provided in inPainting.m with matrix multiplication, which is slightly slower but uses no library that might be not available on octave.
If you use our paper, please cite it:
@UNPUBLISHED{Buttiker2011,
author = {B{\"u}ttiker, Leo and F{\"a}ssler, Urs and R{\"u}egg, Nicolas},
title = {Inpainting by Sensitive Mean Filtering},
note = {A Paper produced during the 2011 Computational Intelligenc Lab course,
Group moebs, Department of Computer Science, ETH Zurich, Switzerland},
month = {June},
year = {2011},
keywords = {image restoration, inpainting},
url = {https://github.com/leobuettiker/CIL-2011-Inpainting-by-Sensitive-Mean-Filtering/blob/master/report.pdf?raw=true}
}
For other approaches see: