Awesome
Implementation of "L0TV: a new method for image restoration in the presence of impulse noise"
This repository is reproduced code of "L0TV: a new method for image restoration in the presence of impulse noise" presented in CVPR2015. This method is for image restoration in the presence of impulse noise.
The paper link is here
Build and run
$ mkdir build
$ cmake .. && make
$ cd ..
$ ./L0TV Lena.png
Example
After run the script, the original image is shown in your screen.
<img src="https://user-images.githubusercontent.com/14243883/31289510-ec2ef160-ab03-11e7-8246-83ecf02d2311.jpg" width="320px">Then, push '+' key and add noise, or push '-' key an reduce noise. The sample image added noise, is shown as below.
<img src="https://user-images.githubusercontent.com/14243883/31289507-ec0a6afc-ab03-11e7-8c32-3c831e9890d6.jpg" width="320px">To run the algorithm, push 'q' key and ENTER. The noise image is here.
<img src="https://user-images.githubusercontent.com/14243883/31289508-ec10ca50-ab03-11e7-9f7b-76ee655e781a.jpg" width="320px">The result image is shown as below. Almost noise was erased.
<img src="https://user-images.githubusercontent.com/14243883/31289509-ec1638c8-ab03-11e7-981e-cfd04fbd6373.jpg" width="320px">