Help with moving objects

Previous Topic Next Topic
classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view

Help with moving objects

Pedro J CamelloDr Pedro J Camello
Hi all,

I´m measuring fluorescence from living cells while they are perfused in
a microscope. The fluid and the applied treatment causes frequent
movement of the cells (smooth muscle). I need to make plot of the
fluorescence in the "cellula" ROI along hundreds of images per
experiment, so that re-drawing the rois when the cells "escape" from the
ROI would be too slow and tedious.

I would like to know how to perform whether the following approach (and
if you think it could work!):

1- I draw a rectangular ROI around each cell, big enough that the cell
doesn´t "escapes" when moves or contracts
2- I define a minimum threshold to create a mask. This mask would be
choosen as the area where measurement (area and average value) would be
3- This must be done with several ROIs/cells in a stack, and within each
ROI the pixels above the threshold would give a Z proyection

Even small help would be very useful :)
Thanks in advance
Reply | Threaded
Open this post in threaded view

Re: Help with moving objects

Pedro J CamelloDr Pedro J Camello
Many thanks for your reply

at the moment what I´m doing is:
- using ROI manager or Multi measure plugin, I draw a rectangular ROI
around each cell.
 I don´t need the clever procedure you describe because I can play the
stack with the ROI to check that the cell never "escapes"

- threshold (I find this step a little subjective) the stack

- set measurements limited to threshold and perform Multi measurement.
This gives me the average values along the stack for the thresholded
pixels within each ROI (i.e., the "cellular" pixels)

The remaining problem is that the final values are no sense for me
because an image  format problem (see my thread "Information regarding

Steve Nanchy wrote:

>Depending upon how 'clean' the background is (i.e. not a lot of variation)
>would it be possible to:
>draw a ROI (rectangular) around a cell
>periodically 'check' each of the four sides (essentially single-line
>selections) to see if a cell starts to touch it (i.e. an increase in the
>average or the sum or the standard deviation of the intensities along that
>and, when that occurs, shift the ROI in the direction to still contain the
>cell (provided, of course, that apparent size of the cell does not
>Just a thought,