dyna-imaging-mod

Introduction

This repository contains all necessary code to: generate realistic vascular geometry with or without tumor, solve direct problem of advection (in plasma) - diffusion (in tissue), use this information to solve an inverse problem through Brix2 or Tofts models. It is composed of the following folders:

• sources: contains all code sources: libnix (helper library) Vascularization (generate geometry direct problem) and VascularizationInverse (inverse problem)
• docker: contains Dockerfile that automatically builds the code in a controled environement.

Build the code

It is recommended to use the provided Dockerfile that will create a Ubuntu 22.04 environment with the necessary dependencies installed. The Dockerfile will also compile all sources and produce executables for each of the code mentioned on previous section.

To build the docker image (from the same directory than this README.md):

docker build -f ./docker/Dockerfile -t dce-mri .

This will create a docker image with tag dce-mri. You can check that the image is indeed created by running docker image ls -a.

To run the created docker image:

docker run -i -t --network host dce-mri

The Dockerfile will automatically download the latest sources on the gitlab repository. The different code executable and libraries will be built in their dedicated folders: libnix, Vascularization, VascularizationInverse.

Simulation examples

Direct problem

Inside the docker container: Create the vascularization

mkdir /Output/no_tumor
mkdir /Output/tumor
cd /Output/no_tumor
/build/Vascularization.out -x100 -y100 -z1 -V. -Irandom0 -R100 -X/Output/no_tumor/no_tumor.xml

cd /Output/tumor
/build/Vascularization.out -V/Output/no_tumor/no_tumor.xml -R10 -r7 -Tr15 -Tx25 -Ty25 -Tc10 -N -R100 -X/Output/tumor/tumor.xml

Compute the transport in the created vascularization

cd /Output/tumor
mkdir conc && cd conc
/build/Vascularization.out -V../tumor.xml -Cupwind

Inverse problem

Inside the docker image

cd /Output/tumor/
mkdir inverse && cd inverse
COMMAND="/build/VascularizationInverse.out"
PREF="/Output/tumor/conc/markerMRI00"
SUFF=".C.binary"

for i in $(seq -f "%03g" 0 180)
do
  COMMAND="$COMMAND $PREF$i$SUFF"
done

$COMMAND

Develop in a container

To build the docker development image (from the same directory than this README.md):

docker build -f ./docker/Dockerfile.dev -t dce-mri_dev .

This will create a docker image with tag dce-mri_dev. You can check that the image is indeed created by running docker image ls -a.

To run the created docker image:

docker run -i -t --network host dce-mri_dev --mount type=bind,source="($pwd)"/sources,target=/source

You can then develop, compile and run your changes with the required packages.