Microvessel density

Hello,
I am posting this topic, as I was not sure if the method and modules I was using was correct, despite going through the manual and other posts.
What I need to measure: Microvessel density in the image below. The vessels are stained brown.
Method I am using: Manually outline the vessels (using identify objects manually), measure area of the outlined vessels (using measure image area occupied). I have an output as indicated below:
Objects or Image Area Occupied Perimeter Total Area
Vessels 18573 4252 4915200

The total area is the area of the image, right?
I understood from one of your discussions that the measurements are in (pixel)2. I require the area in mm2 or µm2. Therefore, I multiplied the area in pixels by 102.01 (1 pixel 2= 102.1 um2).
Queries:

  1. I would like to know if the method described above is correct.
  2. Outlining each vessel is difficult when there are several vessels. Is there a module to select all vessels together (based on the staining intensity) ?
    Please help me with these.

Thanks,
Deepti

Hi Deepti,

  1. Yes, total area is the area of your image. I can’t tell if your if the value in your formula is correct without knowing either a) a measured pixel calibration from an image you’ve taken with that microscope, which would probably be in pixels (not pixels^2) or b) a lot of details about the configuration of your microscope, but in theory you have the right idea.
  2. You can definitely try to use IdentifyPrimaryObjects to automatically find the vessels in your image. You’ll probably want to use UnmixColors to get a greyscale image that’s easy to segment. Once the vessels are identified, you can use MeasureObjectArea to get their individual areas or keep using MeasureImageAreaOccupied to get a total area.

Good luck!

Deepti, I would add that it’s tough to say whether your approach is appropriate without including the pipeline you’ve set up so far, so we can see what the output looks like. Feel free to upload the pipeline or project file.

Thank you. As the image has a scale bar in microm
I will see if it would be possible to get the conversion value from pixel to microm from that.

Thank you for your reply.
Please find the attached pipeline.

pipeline.cppipe (3.9 KB)

Hello!! I want to do a similar measurment of vessel density in a given area. I am using the same output settings but I was wondering if the software calculates the total area of the vessel and not just the stained area. For example in this image there are big round cross sectioned vessels, so will the software just calculate the brown stained area or the whole area occupied by the vessel?

Hi,

It depends on your settings! Generally speaking, if you want the empty space so that you can calculate cross-sectional area, when you’re identifying the vessels you can turn on the “Fill holes” setting in IdentifyPrimaryObjects; likewise if you don’t you can turn that setting off. If the vessels are too big to find as whole objects merely with “Fill holes”, you may have to resort to some fancier tricks with smoothing or possibly even pixel classification upstream.

The easiest way to tell if you’re getting the answer you want is to load your images into the pipeline and just check! It should be fairly obvious from the output if IdentifyPrimaryObjects (open the “eye” next to the module to be able to see the output) is the actual boundary of what you wanted or not.