Our analysis of an article named "A non-local algorithm for image denoising".
- Authors : Erwan Gautier, Alex Foulon
- Article : https://www.iro.umontreal.ca/~mignotte/IFT6150/Articles/Buades-NonLocal.pdf
This repository contains experiments realted to the implementation of the NL-Means algorithm. For more details, read "nlmeansC/README.txt".
- Implementation Author : Antoni Buades toni.buades@gmail.com
To test the NLmeans algorithm, simply put your image of choice into the "images" folder and name it "reference.png". Then, execute our script using ./denoise.sh 20 for a noise standard deviation of 20. You can put any value between 0 and 99 as a parameter.
When finished, check the "images" folder and see the results in "noise.png" and "denoise.png".
To build the project, you first need to go into ./nlmeansC and execute the "make" command.
When trying to create noise you must use 1 as a "add_noise" parameter When trying to remove noise you must use 0 instead The first number is the noise level added or estimated
The first param must be the reference image when noising and the noisy image when denoising Template if you want to denoise noise.png: ./nlmeansC/nlmeans_ipol ./images/noise.png 60 0 ./images/denoise.png ./images/noise.png
Template if you want to noise reference.png: ./nlmeansC/nlmeans_ipol ./images/reference.png 60 1 ./images/denoise.png ./images/noise.png