A human retinal vasculature image (captured by a commercial fundus camera) was segmented to generate a digital geometric model of a 2D vascular network with varying vessel size and density. In order to convert this 2D planar map into a full 3D representation and incorporate circular vessel cross sections, several 2D maps were edited and stacked to vary in-plane vessel width, resulting in a rectangular volume with circular void channels.  Three versions are included here (a) the original semi-planar geometry, (b) a dual-network version in which the original geometry is duplicated at a larger depth, and rotated 90 degrees and (c) a single-network version in which the network is angled (3 Degrees) varying in depth from 0.5 mm to 2.8 mm.  The final designs were converted into the STL file format. Contrast agents (e.g., hemoglobin, fluorescence dye, etc.) may be injected into the empty channels for different biophotonic imaging purposes. For more details on the cleaning process please refer to the "finishing  instructions" section.