plug-in Imagej

For who is interested in digital processing, IMAGEJ is an excellent system, available both in open source and mutli-platforms. A very active community proposes plenty of extra functions, coming as plug-ins. The hereunder link shows how it works. Without any intention of superseding Helicon focus nor CombineZM, I have tried to test the method and "process my own pixels".... It is to be noted that in the plug-in ImageJ, StackFocuser (wich works very well) is already installed. Some other elements have to be installed: FileOpener for a multi-file downloading, TurboReg and StackReg in order to align the images stack.http://bigwww.epfl.ch/thevenaz/turboreg ImageJ allows also to record macros, which make easier the running through the different steps.

Fusion algorithm:

Considering that one pixel is sharper than another if the standard deviation (calculated for his adjacent population) is higher : for one point (x,y), the standard deviation is calculated for each color, and then for each color in the images’ stack....and the "best-rated" colors are saved in the final image. Then a threshold is added in order to shunt the too low values of the system and consequently avoid shimmering in the backgrounds. The last step is to define filters adapted to soften and better the image...

It takes quite few time, and I found it a good template for those who want to have fun by realizing their own digital filters. I have tried to document it; I found great help in this book Digital Image Processing de Wilhelm Burger and Mark J. Burge .

How to install the plug-in :

 Download the file zip (see annexe) and put it in ./imagej/plugins directory. -
 Run imagej , and open the menu plugins/Compile and Run…
 Select the directory plugins and search fusion2_.java in the list. Normally a message should be displayed: "Stack required". The file fusion2_.class having been created will be proposed in the menu when running again ImageJ. I also discovered recently another program used for compiling the sharpness zone : zerene stacker it works very well,, uses maybe less parameters, but is quite simple to use. Moreover, it is written with Java and should be proposed with Linux.