flexible ROI option with trackmate or other tracking package

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

flexible ROI option with trackmate or other tracking package

Ziqiang Huang
Dear ImageJ users,

Does anyone tried before to get flexible ROI that tracing the outline of a cell in a time-lapse image datasets?

I have DAPI stainings of some cells (around several hundreds, not overpopulated but do overlap with each other some time. They mainly undergo Brownian motions), and imaged them over a time of 60 hours, and acquired an image every 15 minutes.
Those DAPI signal will potentially change intensity or StdDev over time, and in different treatment groups, this temporal dynamic might be different.
There I would like to delineate each cell individually, and this I can easily done by shareholding, play around with binary masks and particle analysis. A simple macro could do this in a few minutes iterating through every time frame. To track the cells, I am satisfied with Trackmate which returns reasonably good tracks and capable of dealing with gap (missing signal in a few frames) or merge events etc. But currently I found it only provide circle ROIs. I checked the imagej website and also Jean-Yves's webpage about trackmate, as well as Jaqaman's MATLAB implementation and software (u-track 2.0). It seems a certain part of the algorithm they applied rely on the circle shaped ROIs (to compute cutoff radius, and to estimate intensity change due to merge and so on). For my dataset, the circle could not faithfully represent the whole cell, and therefore I would like to keep the tracking result, while be able to measure intensity or StdDev of the whole cell.

It seems to me I probably have to code the tracking part myself (the LAP algorithms, to link cell and segments from ROIs) . However I feel this is a quite common requirement in cell biology and image analysis, that to track and measure whole cells over time.
I would expect there should be already some implementations or tools that can do this. It will be really helpful if anyone knows about this that can share with me some information.

Best Regards,
Ziqiang Huang

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html
Reply | Threaded
Open this post in threaded view
|

Re: flexible ROI option with trackmate or other tracking package

Straatman, Kees (Dr.)
Dear Ziqiang Huang,

Without really looking into this would it be possible to use the macro function Roi.contains(x, y) where the x,y are from the circle ROIs?

You might be able to link the circles found with the tracker with the cells you have identified.

Best wishes

Kees

-----Original Message-----
From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of Ziqiang Huang
Sent: 03 January 2018 15:05
To: [hidden email]
Subject: flexible ROI option with trackmate or other tracking package

Dear ImageJ users,

Does anyone tried before to get flexible ROI that tracing the outline of a cell in a time-lapse image datasets?

I have DAPI stainings of some cells (around several hundreds, not overpopulated but do overlap with each other some time. They mainly undergo Brownian motions), and imaged them over a time of 60 hours, and acquired an image every 15 minutes.
Those DAPI signal will potentially change intensity or StdDev over time, and in different treatment groups, this temporal dynamic might be different.
There I would like to delineate each cell individually, and this I can easily done by shareholding, play around with binary masks and particle analysis. A simple macro could do this in a few minutes iterating through every time frame. To track the cells, I am satisfied with Trackmate which returns reasonably good tracks and capable of dealing with gap (missing signal in a few frames) or merge events etc. But currently I found it only provide circle ROIs. I checked the imagej website and also Jean-Yves's webpage about trackmate, as well as Jaqaman's MATLAB implementation and software (u-track 2.0). It seems a certain part of the algorithm they applied rely on the circle shaped ROIs (to compute cutoff radius, and to estimate intensity change due to merge and so on). For my dataset, the circle could not faithfully represent the whole cell, and therefore I would like to keep the tracking result, while be able to measure intensity or StdDev of the whole cell.

It seems to me I probably have to code the tracking part myself (the LAP algorithms, to link cell and segments from ROIs) . However I feel this is a quite common requirement in cell biology and image analysis, that to track and measure whole cells over time.
I would expect there should be already some implementations or tools that can do this. It will be really helpful if anyone knows about this that can share with me some information.

Best Regards,
Ziqiang Huang

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

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