Problem identifying object in phase contrast images

Hi there,

I am a new user to cellprofiler and am having difficulties creat a pipeline that can measue bands in meat fibres.
I have attached an image below so that you can see what I am talking about. The consist of several meat myofibrils which are thread like structures that have two different types of band. I want to measure the width of both types of band.I have attached the pipeline which I had tried- not sure why its doesnt work.
I am not sure how to set an accurate threshold, as when I’m measuring after I cliack on the image a line doesnt appear between the two endpoints.
I have also treid to use the enhance edges module but find it turn my images totally black.

Any help you could give me would be greatly appreciated.
Many thanks
Ruth
2010_11_22.cp (7 KB)


Hello,

Your images are unfortunately difficult on two counts: First, brightfield images are notoriously difficult to analyze by automated image analysis algorithms - not just for CellProfiler but in general. Is it possible to use a fluorescence stain of some kind, where the contrast will be high between the cells and the background? Second, the algorithms in CellProfiler have been geared for two basic types of shapes (not that they cannot work on other shapes, but they have certainly been designed for particular shapes): (a) blobs/circles/round-ish objects like most cell types and colonies, (b) C elegans worms, which are linear. If you could generate fluorescence images, the Worm Toolbox modules in CellProfiler might help you; they do use machine learning where you train the computer about what to look for. This is sensible for very large experiments but spending time on the training makes less sense for experiments that are small, where it might be quicker to measure by hand!

Best wishes,
Anne