Figure 1
From: Statistical physics approach to quantifying differences in myelinated nerve fibers
Processing steps for characterizing myelinated axons and their structure.
The characterization process is done on a total of 67 electron micrographs (EM images) of the fornix, belonging to 6 different subjects. (a) We show, as an example, one of the EM images of subject AM77 (6 years old). The scale bar (in black) measures 2 μm. (b) Our recognition algorithm (see Methods for details) is used to segment the areas of myelinated axons (in red). (c) Schematic diagram of a small set of myelinated axon contours and the respective Voronoi tessellation of the embedding space. For one of the axons we show the relevant properties used to characterize each axon: they are the centroid position Ri, axon area Ai, Voronoi cell area 
and axon shape parameters, collectively referred to as Γi. These properties are then used as a basis to calculate the 45 features used to describe each EM image (e.g. axon density, hexagonality index, etc.).
