Analyzing skin layer thickness

imagej

#1

Dear ImageJ Users,
I have been trying to measure skin layer thickness mainly the stratum corneum using an OCT image. First, I set the scale using a meter rule image. Thn used “drag and drop” to upload the Oct image of the skin where I saw the skin layers but using a straight that ran across the the layer and hitting “ctrl M” to measure the thickness looked highly errorous due to human judgement of the edges of the layers. Can someone help with more accurate procedure. Thanks


#2


Here is the image I am analyzing
Thanks


#3

Hi,

Is it so that the bottom of the skin layer is always at the bottom of the image, as seems to be illustrated in your example image with the line drawn in it?

If so, you could try and process the image such, that you will end up with the bright upper layer (not very easy):


Then you could obtain the coordinates of the upper line via: Analyze > Tool > Save XY Coordinates. In the binarized image that l managed to create, you’ll see that there will me multiple y-values for each x-coordinate → you’ll need to extract the lowest y-value for each x-coordinate (0,0 is in the upper left corner of the image).

Is this what you are after?


#4

@engr:
This gives you a nice area, but as @Danielle_Z says: is this what you are after?

selectWindow("11fb545db9ffaa98c6eda17f64820a0505622c06.png");
run("Smooth");
run("North");
run("Fill Holes");
run("8-bit");
run("Despeckle");
setAutoThreshold("Default dark");
doWand(324, 270);
selectWindow("Recorder");

#5

Good day!

Here is the result from automatically contour tracing


which may give you the thickness as a function of the horizontal position, if you define a reference line, e.g. the bottom border of the image.
(Obtained with a commercially available ImageJ-plugin.)

Please have a look at the following posts/threads:
http://forum.image.sc/t/retinal-layer-segementation-of-raw-bitmap-oct-b-scans-for-export/10651/2?u=herbie
http://forum.image.sc/t/segmentation-in-oct-imaging/8152?u=herbie

Regards

Herbie


#6

It is close, however, thanks so much as I got several experiments that distracted my attention to this one and finally I got an alternative. Thanks for your concern Danielle


#7

Thanks so much eljonco


#8

Oh thanks, so much, Herbie.