Supplementary Figure 1: Validation of FibrilTool with artificial images.

a-d. Examples of artificial images (372x372 pixels) obtained by generating 50 random segments (black segments, width 2 pixels) with (a,b) or with no (c,d) directional bias; the images in (a,c) were “degraded” by applying a Gaussian blur of radius 5 pixels (b,d); the theoretical orientation and anisotropy score can be computed directly from the distribution of segments, while the output of FibrilTool is shown in blue. e,f. Results of the analysis of a set of 10x3 such images with FibrilTool: raw images (continuous lines), moderately degraded images (blur of radius 3, dashed lines), and degraded images (blur of radius 5, dot-dashed lines). Data are plotted as a function of the theoretical anisotropy score: anisotropy score measured with FibrilTool (e); difference in degrees between orientation measured with FibrilTool and theoretical orientation (f).