how do you distinguish between an artery and vein in the mice retina from looking at the fundus image and OCT image?
that’s hard to say without an example image. Would you mind sharing an image with us and tell us a bit what’s in the image? We may then be able to point you in the right direction for getting started with some image analysis.
I have attached an image here. This image is a fundus image of the mice retina after injected with a fluorescent dye. We are trying to distinguish between the arteries and veins in this image. Do you know a method of how could we distinguish it? And using image J how do we find the diameter of the arteries and the veins?
- Segment your image. Thresholding is the first thing to try, but it wont work for this image. You can look into local thresholds which should perform better than a single global threshold. There are many methods for segmentation that you can dig into.
- Analyze -> Local Thickness -> Local Thickness. There are a few options, you can look into the documentation.
Why wouldn’t I able to do thresholding with this image?
What I mean is that you won’t get very good results with a global threshold in this type of image because the smaller vessels have a different signal/intensity than the larger vessels. A local threshold will help with that, but I’m skeptical that it will do the job. Segmentation of retinal fundus images has been an open challenge problem in some recent machine learning competitions.
Hey Rujul, are you looking to do the measurements manually with the GUI, or automatically for a large number of images, which would be a major challenge?
Distinguishing between arterioles and venules will be challenging, even manually, if all you have is images like the one you gave as sample, which has poor focus btw. Do you have images/video earlier after the administration of the fluorescein? If you do it would make life a lot easier differentiating between the two type of vessels.
And I don’t think you gave a sample of the OCT images you have.