Blood-Vessel Synthesizer

This repository contains the implementation of the blood-vessel tree generator of our paper:
📚 Rauch N., Harders M., Interactive Synthesis of 3D Geometries of Blood Vessels, Eurographics 2021
📚 Rauch N., Harders M., Methods for User-Controlled Synthesis of Blood Vessel Trees in Medical Applications, IEEE ACCESS

The synthesizer emulates the abstract behavior of angiogenesis - a growth process in which blood vessels develop by elongating and branching from pre-existing vasculature to reach tissue devoid of any vasculature. The overall mechanism resembles the competiton for space between individual branches, which we modeled with the space colonization algorithm. During development the geometry of branches and bifurcations are further constrained to replicate commonly observed vascular patterns.

Build

The library depends on glm, eigen, and pybind11 which are already contained.

$ mkdir build && cd build/
$ cmake -DCMAKE_BUILD_TYPE=Release ..
$ cmake --build .

The synthesizer is implemented as a shared C++ library, with optional python-bindings (module).
The following compile options are available:

option	description
VS_PYTHON_BINDINGS	build Python Bindings (module)
VS_PROFILER	build with Profiler Functionality (performance measurements)
VS_COMPILE_NATIVE	compile for micro-architecture and ISA extensions of the host
VS_COMPILE_FASTMATH	compile with fastmath optimization

and can be enabled during configuration:

cmake -DCMAKE_BUILD_TYPE=Release -DVS_PYTHON_BINDINGS=ON -DVS_PROFILER=ON ..

⚠️ library is not statically link against the c++ libraries; on windows you need to move the necessary .dll to the lib folder:

• e.g. mingw (pthread) you need to add libgcc_s_seh-1.dll, libstdc++-6.dll, libwinpthread-1.dll

Python Example

Check the notebook notebook/api_showcase.ipynb for python-module examples.

# set path to compiled vessel module / library
import sys
sys.path.insert(0, '../build/lib/')
import vessel_module as vs

# define domain in which vessels should grow; can be = DomainCircle, DomainSphere, DomainLines, DomainVoxels
sphere = vs.DomainSphere([0.0, 0, 0], 0.5)

# create synthesizer object for domain
synth = vs.Synthesizer(sphere)

# modify settings (these are highly dependent on the dimensions of the domain, use scale to find an initial setup where things work)
synth.settings.steps = 100
synth.settings.samples = 1000
synth.settings.scale(1.5)

# set root note of tree (or use set_forest(...)) to grow from intial trees
synth.create_root(vs.System.ARTERIAL, [0.5, 0.0, 0.0])
synth.run()

# check if library was compiled with performance monitor
if vs.PerfMonitor.Enabled:
    import pandas as pd
    import matplotlib.pyplot as plt

    # retrieve time measurements as dictonary
    times = synth.get_arterial_perftimes()

    # create pandas dataframe
    df = pd.DataFrame.from_dict(times)
    df = df.astype("timedelta64[ns]").astype("int64")
    df = df / 1e+6;

    # plot
    columns = ["total_arterial", "step", "step_closest", "step_growth", "step_kill"]
    fig, ax = plt.subplots(figsize=(10,5))
    df[columns].plot(ax = ax, kind="line", lw=2.5)
    df[df.columns.difference(columns)].plot(ax = ax, kind="area", stacked=True, alpha=0.75)

    ax.set_title("Performance Measurements Arterial Development")
    ax.set_xlabel("steps")
    ax.set_ylabel("time [ms]")
    ax.grid(which='major', color='#DDDDDD', linewidth=0.5)

Citation

If you use this code in your research, please cite one of our papers:

@inproceedings{10.2312:egs.20211012,
    booktitle = {Eurographics 2021 - Short Papers},
    title = {{Interactive Synthesis of 3D Geometries of Blood Vessels}},
    author = {Rauch, Nikolaus and Harders, Matthias},
    year = {2021},
    month = {May},
    publisher = {The Eurographics Association},
    ISSN = {1017-4656},
    ISBN = {978-3-03868-133-5},
    DOI = {10.2312/egs.20211012}
}