|
ImageJ Plug-In for Fast Denoising of Noisy Biological Images
|
Ted
Choueiri
Section Microtechnique, EPFL
|
Semester
Project
November 2006
|
Abstract
|
This project objective is to implement a simple, easy-to-use ImageJ
plug-in for fast denoising of noisy images using a blind Wiener Filter. The
noise is a white Gaussian noise, and the noisy image’s Power Spectral Density
follows :
|
|
Example
![goldhill](projet_files/image014.jpg)
Image Goldhill
|
![Noisy goldhill](projet_files/image015.jpg)
Noisy Goldhill, SNR 15
|
![PGdenoisednoisy](projet_files/image016.jpg)
Projected Gradient denoised, SNR 21.24
|
![](projet_files/image017.jpg)
Projected Newton denoised, SNR 21.27
|
![PSD (blue) and Linear Fit (black)](projet_files/image011.jpg)
PSD of noisy input image (blue) and linear
fit (black)
|
![PG PSD input PSD model](projet_files/image012.jpg)
Projected Gradient Method : PSD of input
(blue) and PSD of model (black)
|
![PN PSD input model](projet_files/image013.jpg)
Projected Newton Method : PSD of input
(blue) and PSD of model (black)
|
|
|
|