Hi fellows, first of all thanks so much for your contribution with this great software!
In my case, I use CP to obtain the trajectories of moving objects and then get their speed (attached pipeline). The input data for CP is a stack of 1200 TIFF images.
The quotation marks in the title of the post mean that I don’t intend to say that tracking is simple, actually I am aware that it is a really complex matter. But this is a very basic, simple use of the tracking module.
I have two questions:
Most important is about how long it takes for this “simple” tracking use. Is it OK that it takes several hours (at least 3.5, when I hide the windows of all modules). Or maybe I am doing something wrong or redundant? Maybe this is normal for the number of files. I run it on a Toshiba Tecra, Centrino Duo, WinXP, 2GHz, 3GB of RAM.
Less important but related: I use ImageJ for thresholding the image sequence, therefore the input images are already binary. Is there a way to directly skip the thresholding step in IdentifyPrimaryObjects module? I tried to tell CP that it was a Binary Image (by selecting “Binary Image” in the “Thresholding method” field) but CP tells me that it did not find a binarized image… but most likely I am doing/thinking wrongly in this. I want to feed CP with a binarized image and would like to simply skip the thresholding by CP (this is not only to save time, it is that by now I am more comfortable by adjusting the threshold with ImageJ).
Again, thanks so much for CellProfiler!!!
Update: by looking into the DefaultOUT_Image.csv file (GREAT idea by the way, especially for finding actual or potential errors!!) I can see that the most time-consuming module is the tracking one (approx. 10 seconds per image). So I realize now that the threshold does probably not a big difference (as it belongs to the Identify module).
S5_track.cp (6.33 KB)