I have a bunch of stacks of images. About 1500x1500, 16 bit grayscale. Each image in the stack represents a small slice of the visible light spectrum (about 25 images per stack, more or less).
Actually, there are TWO stacks covering the spectrum - leading to two “composite” images - but I think we can ignore that for now.
My question is: has anyone used one of the “stack registration” tools to register this type of image. Most of the demos I see imply that the stack registration tools are intended for stacks which move in Z through a volume. My stacks are all of the same field of view (well…almost…and that’s the problem) and move through the spectrum. These two cases seem to me to be just different enough that it’s not clear how techniques that work for one case will work for the other
I’m about to run some tests - but I thought I would ask to see if someone has been here before and can give me some advice. I’m *especially* interested in answers from folk who can report that “no, the standard volume stack registration methods don’t work well - but here is a technique that *does* work”. I have grown tired of re-inventing wheels.
The issue is that the imaging process was a big clunky, and the images in the stacks are not perfectly aligned. The range of motion is on the order of 10 pixels in X, or Y, or both. I just want to remove that jitter. Given the causes of the jitter, I doubt that there is a consistent drift as you go through the stack - I suspect that there are occasional jumps, in random directions - perhaps as often as every slice, perhaps only 2 or 3 “bumps” in the entire 25-image stack.
--
Kenneth Sloan
[hidden email]
--
ImageJ mailing list:
http://imagej.nih.gov/ij/list.html
Locked
