In my case, I want to measure the curvature of blue-marked particles in RGB color images, but the only possible way I've found to do it is by manually tracing the particles' perimeter and measuring the curvature with the Kappa plugin. As there are so many particles to measure, I think it would be much easier to just select them (i.e. with a threshold) and later measure them all automatically (if that is possible in anyway).
Thanks in advance!
I have no czi reader available... Could you repost in another format (bmp, jpg, tif, xdr or raw)?
OK, thanks Oliver.
Most but not all blue shapes are oval, i.e. the curvature varies per shape and in each shape. How many curvature numbers do you want?
(I'm ignorant of the capabilities of ImageJ)
You could:
- Find a good threshold in the image to segment the shapes. Easier said than done: the out-of-focus blue shapes are grainy and may overlap the shapes you want to analyze.
- Do a kind of marching cubes in 2D, resulting in vertices and line segments.
- At each vertex, calculate the angular defect. This, divided by half the length of each adjacent line segment gives you the curvature at each vertex.
(- Do a sanity check with the Gauss-Bonnet theorem.)
What do the images represent?
OK, Lambert. Thanks a lot for your help! These images represent chloroplasts (in blue) and the nuclear envelope (in red).
a = Image 1.czi (RGB).tif
b = detail of a, with blue overlapping red and a grainy out-of-focus shape
c = the blue channel (level=86, window=1); too noisy
d = same as c, but after applying a 5x5 smoothing filter
e = marching squares at value 86.4
f = detail of e. In red the vertices, in blue the line segments
At each vertex, the angular defect is 180 degrees minus the angle between the two adjacent line segments. Dividing the angular defect by half the length of those two line segments yields a curvature per micron, but there are more ways to express curvature.
Good luck!
Better download the attached image. Research Gate only displays it downsized...
- Badges
- Science topic
- Similar topics
- Biological Science
- Molecular Biology
- Biomolecules
- Nucleic Acids
- RNA