The algorithm is described here: http://dx.doi.org/10.1109/TMI.2013.2261536

Building

The source code is organized in following folders:

core - which contains the Makefile and the main source file core/lib - which contains the rest of the source code core/test - which contains tests for the code

To build the program, you will first need the following packages (with development headers for the libraries:

libvtk libeigen3 g++-4.7 cmake ninja (or make)

These can be installed on a Debian GNU/Linux machine by issuing (as root) apt-get install libvtk5-dev libeigen3-dev g++-4.7 ninja-build cmake cmake-qt-gui git

To build the program as a standalone algorithm:

• git clone git@g.nosp@m.ithu.nosp@m.b.com:Danielhiversen/AngleCorr.git angelCorr
• mkdir -p build; cd build
• cmake -G "Ninja" -DCMAKE_BUILD_TYPE=Release -DANGLECORRECTION_TEST_DATA_DIR=../testData/ ../angelCorr/core/
• ninja

Problems with the Eigen library? Try:

• git clone git@g.nosp@m.ithu.nosp@m.b.com:RLovelett/eigen.git
• cd eigen
• git checkout 3.2.10
• cmake -G "Ninja" -DANGLECORRECTION_TEST_DATA_DIR=../testData/ -DEIGEN_DIR=PATH_TO_EIGEN ../core/

The test data can be downloaded as: git clone https://github.com/Danielhiversen/angleCorr_data.git testData

Using

Running the program angle_correction without any arguments gives a short help text:

$ ./angle_correction

Usage: ./angle_correction centerline.vtk image_prefix Vnyq cutoff nConvolutions

As indicated, the program must be run with 6 parameters:

centerline.vtk - the center line

image_prefix - the prefix of the images. The suffix $NUMBER.mhd will be appended to it, where $NUMBER starts at 0 and progresses until no more files are found.

Vnyq - Nyquist velocity used when acquiring the images

cutoff - The lower cutoff value for abs(cos(theta)), used for the velocity estimation algorithms

nConvolutions - the number of times to run the smoothAll algorithm

Output

An output file are generated: output.vtk

output.vtk is a file that may be visualized by ParaView or CustuX, and contains the center line as a point set. Each point is annotated with the following data:

The flow direction (normalized gradient of the center line curve) The estimated flow velocity (simple estimate) The estimated acquisition direction vector