Image Analysis - Measuring change in roi for an image array

We are trying to find the best way to measure change in channels of sponge tissue during a contraction.

The roi's can be found in fig. 2 D R1 - R4. http://jeb.biologists.org/content/210/21/3736.full.pdf+html

There is also a supplemental video showing the contraction. Movie 1 http://jeb.biologists.org/content/suppl/2007/10/12/210.21.3736.DC1/JEB003392-movie1.mov

As is often the case we have hit a few stumbling blocks:

1) I have tried calculating the mean, std2 and ratio of black to white pixels, for the roi of my grey scale image array. However, these measurements are not very informitive (no clear trend; furthermore, the changes don't match what we see in the video). Im wondering if this could be because we need to threshold the ROI? Any thoughts on what would be a more appropriate type(s) of analysis that could measure the event?

2. A minor thing is that it would be nice to have the 'picking of the roi' in the script so that we could run it, it would prompt us to select one and then a second (and maybe a third) roi, and then it would continue to run the script.

Any help is much appreciated.

Matt

Answer 1

[David Young]

I'm not exactly sure what you mean by "roi" here, but the starting point would normally be to try thresholding followed by measurements using regionprops.