Special normalization

[Pedro J CamelloDr Pedro J Camello]

Hi all,

I have 2D spatiotemporal images where:
- x dimension represents space along a linear structure
- y dimension is time
- the pixel value is the variable of interest (contraction along a segment of intestine)

My problem is that there are places along the line (x dimension) with bright spots that add to the changes of interest takimg place with time, producing artifacts in some measurements.

I want to use the first row (or a few initial rows) to normalize the rest of the image. I think there are two ways:

- to normalize using as reference image a synthetic image made by repeating the first row: is it possible to do that in ImageJ?
- access, with a plugin or a macro, each row of the image and normalize its values to the values of the first row: has anybody a plugin/solution for this?

Thanks in advance

Pedro

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[Pedro J CamelloDr Pedro J Camello]

Hi Michael,

many thanks, I´ll try it.
What I don´t understand is why do you add all the pixels of a row.
What I was thinking of was:
- I select a single row (a X line) to be used as reference
- then I normalize each column of pixels respect to that line by simple division. So if I select row 1 as reference:

                           Y(x,y)/Y(x,1)

However, the problem with the procedure I describe in my post is that the changes taking place with time (along Y columns) will evolve always around 1, so that I cannot compare absolute values between different images, and that´s also an important point to analyze in my data (coming from different experiments and individuals).

That is why I´m thinking of removing, in an initial line, the spots by simple thresholding.

For example, in the next "image" I would remove the 30 pixel which is introducing a + 20 value along its column, distorting the wave-like change that runs from left to right

10 10 10 10 10 30 10 10 10
40 30 20 10 10 30 10 10 10
20 30 40 30 20 30 10 10 10
10 10 20 30 40 50 20 10 10
10 10 10 20 30 60 30 20 10

Though easy to make a crop and thresholding in the initial rows to "delete" the bright spots, what I find more difficult is to substract its value to the rest of the image. That´s because I was thinkin of creating a synthetic image (easy to operate with the experimental image)

Anyway I think your code will be very helpfull for me

Thanks and regards

Pedro

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html