HOME
OVERVIEW
APPLICATIONS
INTERACTIVE APPLETS
Build a robot
Extract anatomy
Traveling Salesmen
Evolving Curves
Image Denoising
HISTORY OF THE METHODS/FLOW CHART
PUBLICATIONS
EDUCATIONAL MATERIAL
ACKNOWLEDGEMENTS
ABOUT THE AUTHOR/CV
Copyright:
1996, 1999, 2006
J.A. Sethian
|
A Java Applet for the Min/Max Denoising Scheme
Coarse-grain version
Instructions:
- Using the mouse, click points on a curve in the box.
- Next, put the mouse in window that says "Noise", and type in a number
between 0 and 100.
- If you type in 0, no noise will be added
- If you type in 20, then 20/100 = 20% of the pixels will be thrown away,
and replaced with random grey-scale pixel values somewhere between white and
black.
- If you type 100, then all the pixels are thrown away, and
if you think any algorithm is going to find the original picture after
the entire thing has been trashed, you are nuts.
But I know you'll try it anyway.
- While the cursor is in the Noise window, be sure to hit Return.
- Now, click on "Go".
- The following things will happen:
- First, the points will be connected and you'll see the black shape on
a white background.
- Next, the amount of noise you requested will be added to the picture,
and that picture will be displayed.
- Then, the min/max de-noising algorithm will start, and
the image will be shown every once in awhile as it is cleaned.
- Java has its virtues, but it can also be slow as molasses in
January. You may find yourself waiting a while.
- NOTE: YOU MUST STOP THIS ALGORITHM BY CLICKING ON STOP!!!
If you don't do this, my Java applet will keep running and running and
running.
Follow these instructions carefully, or my java applet will take ahold of your machine and bring it to its knees!
Original |
|
|
Original with Noise |
|
|
Denoised |
Instructions:
Using your mouse, click several points on the white panel
to define the vertices of a polygon. When you are finished
entering the points, place the cursor in the Noise window,
and type an integer between 0 and 100, and then hit return
(while the cursor is still in the window). Then click on "Go".
Press "Stop" to end the program.
Some things to watch out for:
- We use a 100x100 grid of pixels, which is then
expanded to be three times as large.
- The algorithm keeps running, trying to find places to de-noise. Once the
image is cleaned up, the code keeps running, even though nothing changes.
This means you don't have to decide when to stop the process.
If You Want To Learn More:
- Malladi, R., and Sethian, J.A.,
Image Processing via Level Set Curvature Flow ,
Proceedings of the National Academy of Sciences,
Vol. 92(15), pp. 7046--7050, July 1995.
|