Apply background correction to a sequence

tlwash123

Hello, so I am trying to do a background subtraction to a sequence of deconvoluted images using the original sequence of images but when I do this, only the first frame from the original sequence is used to subtract as the background from the entire deconvoluted sequence.

What I am trying to do is process it so as to subtract (frame 1.orig) from (frame 1.deconv) then subtract (frame 2.orig) from (frame 2.deconv) and so on.

Is this possible, do I need to write a macro?

KevinR

The "Background Correction" function specifically uses a single background or flat field image to correct the target, so that's acting as it should. If you have an entire sequence of background images to subtract, use the "Process | Operations" dialog to subtract your background sequence image from your deconvolved sequence as an "image" operation. 

I am, however, wondering a bit about your desired results. A deconvolved sequence will (a) not have much background and (b) I would expect the same background (due to camera issues) to apply to each frame. In fact, to get the best deconvolution results, background subtraction and flat fielding should be performed _before_ deconvolving, not after. 

tlwash123

Maybe it would be better to just explain my desired result and you could enlighten me on a more efficient way of obtaining these results, because it seems that I may be going about this the wrong way (I have been learning this program for only about two weeks now).

Please see attached image.

I have a set of images that have been CT scanned at ~2.5mm per slice. I am trying to enhance the photo so that I can see only the cracking (traced by blue) and get rid of all the small dark speckers (circled in green) while retaining the fine detail of the cracks as they originally were (need to measure the volume of cracks after constructing the 3D image). I've tried using filters and combinations of the filters but they either blur the image too much or I lose too much detail by the time I've gotten rid of most of the specs. Any advice?

KevinR

I would suggest filtering to the extent that your cracks can be thresholded, along with the specks and some other stuff, then using Count/Size to filter based on length, area, aspect, and/or other measurements where the cracks will have different statistics than the specks. 

As a general rule, thresholding should be quite aggressive to get as many of your desired objects as possible, identifying more than you want, with detailed measurements and filters then cutting back the unwanted items. Being too conservative with the threshold identification means that you cannot recover missed objects. Oversegment then reduce.