Need a pseudocode for image-analysis

Question

[murmublog]

In confocal imaging we get many stacks of images which can be reconstructed in a 3D format.but most imaging analysis takes into account a 2D projection of all the frames, which makes our analysis faulty (because many cells would upregulate and yet may not be visible because it is in some other frames). So my strategy is to do pixel based analysis across all the frames, determine which regions get up-regulated and then construct the 2D projection.

To sum up I'm attaching my idea in a picture.. I need help in creating a the pseudocode for my problem

[Pierre Lindenbaum]

"I have discovered a truly marvellous proof of this, which this margin is too narrow to contain"

[murmublog]

Hi Pierre can u elaborate me bit more?So that I can truly appreciate what u are saying

[Devon Ryan]

Pierre was being a bit snarky :)

[Devon Ryan]

Isn't there an imageJ plugin that does something like this? If so, try to see if you can find its code.

[murmublog]

Well, my experiment will be like this. My steps:

1. Take 2 frames of the image.... compare the pixel intensities and take std deviation
2. Then follow the same for the 'n' frames for the total image files
3. Then perform a PCA or ICA for the same..

So suggest me an pseudo code as I'm new to coding not much idea...so I need a way to start.

[george.ry]

I'm not seeing an image? (Unless it's on on dropbox, perhaps, and blocked by the proxy here?)

Answer 1

[Charles Warden]

I feel like there should be some sort of plug-in for ImageJ/Fiji, but I haven't done this particular analysis before. This is what I found with a quick Google search:

http://imagej.1557.x6.nabble.com/2D-image-to-3D-in-ImageJ-or-Fiji-td5000113.html

If you need to automate the process, you can write ImageJ macros:

http://rsb.info.nih.gov/ij/docs/macro_reference_guide.pdf

Answer 2

[abhay.krishna]

Hello,

Try seeing SD cube image rail, if this has been implemented here, you might get code for it as they have released API

http://www.semanticbiology.com/software/sdcube

Answer 3

[Martombo]

You can try to use opencv for this kind of analysis: http://opencv.org/ it is available in C++, python, java.

but I fear that the code is going to be too specific for your problem to get any general help here.