Automatic Quantification of Atherosclerosis in Contrast-Enhanced MicroCT Scans of Mouse Aortas Ex Vivo

Figure 1

Representation of the segmentation algorithm. (a) Cropping to ROI. (b) High threshold to create aortic wall draft. (c) Manual contour converted into a mask. (d) Inward dilation by two voxels and juxtaposition to wall draft. (e) Inversion and inner volume extracted as lumen draft. (f) Dilation/erosion (5 voxels) to remove plaque from lumen draft. (g) Addition of lumen mask and wall draft. (h) Subtraction of lumen generates wall mask. (i) Plaque segmentation: Gaussian filter (; ) then low threshold ( cm^-1). (j) Lumen mask applied. (k) Refinement with opening (; ) followed by (l) closing () filters. (m) Aortic lesions defined as contacts between plaque and wall. (n) Aortic wall, plaque, and lesion concatenated. indicates manual interventions.