Measurements of fish larvae

Hello everyone,

I am quite new to ImageJ (and image analysis in general) and have been trying to do some measurements with the help of its documentation, but I am stuck at the segmentation stage. It would be great if you could give me some guidance or advice.

Here is an example of my pictures:

I am primarily interested in finding (1) the yolk-area and (2) the notochord length of the fish larvae. My goal is to be able to do this via Macro (possibly with minor manual modifications later).

So far I have tried the following for the yolk-sac area:

a) Subtract Background -> 8 bit -> Threshold (I fail to enclose the area of interest.)
b) Find Edge -> Binary -> (manual Erode -> Wand tracing tool) (I fail to select only the yolk-sac area as in the image it is not clearly enclosed.)
c) Manual Selection Brush Tool (It takes a lot of time to do so, but if the Batch process can initiate some process so that I only need to make minor adjustmnets to the selected region, it would also be very helpful.)

Any suggestions are highly appreciated. Thanks in advance!

We have recently found that the yolk sac is autofluorescent. If you are imaging on a system that has access to fluorescent channels, you might be able to use that to assist with your measurement.

As for the chord length, that seems more difficult due to the variety of straight lines you have in your image, and the likely variations between images. This tends to impact the yolk sac measurements as well, even using fluorescence, as the orientations of the larvae generally aren’t perfect.

Thank you for your reply! It is interesting to know that the yolk-sac is autofluorescent. Unfortunately the stereomicroscope I am using doesn’t have that function, but I will keep this in mind.
The orientations of the larvae are indeed troublesome. Had to practice a lot to get them on their side - but it is still an estimate by sight (I hope to find a better solution, too).

I played with the image in ImageJ again and found that I had forgotten there’s a function opposite to Erode -Dilate! Having applied both of these several times in process b) did get me a selection on the yolk area. It is a somewhat better method than c) which only involves freehand marking. My idea with the notochord length was to first Find Edge and Fill the notochord and then measure this shape. Still a lot to learn.

1 Like