About

This project is an integration of several key open-source software required for annotation and segmentation of DICOM images.

Getting Started

There are a few prerequisites for this project:

1. Docker
2. Docker Machine
3. Virtual Box

Docker Swarm Visualization

More archive nodes can be added to the distributed service as illustrated above.

You can also scale the service vertically by adding another docker swarm as illustrated below.

Docker-machine creation:

A docker machine instance can be created using the following command :

docker-machine create --driver virtualbox {Node name}

To access the created docker-machine :

docker-machine ssh {Node name}

Docker swarm creation:

$ docker swarm init --advertise-addr <NODE-IP>

Eg:

$ docker swarm init --advertise-addr 192.168.99.100
Swarm initialized: current node (dxn1zf6l61qsb1josjja83ngz) is now a manager.

To add a worker to this swarm, run the following command:

    docker swarm join \
    --token SWMTKN-1-49nj1cmql0jkz5s954yi3oex3nedyz0fb0xx14ie39trti4wxv-8vxv8rssmk743ojnwacrr2e7c \
    192.168.99.100:2377

To add a manager to this swarm, run 'docker swarm join-token manager' and follow the instructions.

Follow the instructions in the code block to add worker nodes.

---

Docker Node Description

Node1

On node1, we run Lightweight Directory Access Protocol (LDAP). The LDAP image that is used in our docker compose file is dcm4che/slapd-dcm4chee:2.4.48-22.1.This image requires two mount points :

The default mount points are highlighted in orange. The mount points can be changed as per preference as long as they are present on node1. By default LDAP is available on port 389(highlighted in red in the image above), but it can be changed as per user requirement. In order to make these changes, you must directly edit the docker-compose file, and re-deploy the stack.

Node2

On node2, we run our database - postgres . The postgres image that is used in our docker compose file is dcm4che/postgres-dcm4chee:12.2-22.This image requires one mount point :

The default mount points are highlighted in orange. The mount point can be changed as per preference as long as it is present on node2. The database running on node2 is a centralised database for the whole docker swarm, so it must be present on node2 to ensure centralised access to the whole docker swarm. By default the database would be running on port 5232(highlighted in red in the image above), but it can be changed as per user requirement.

Node3

On node3, we run the dcm4chee archive along with wildfly. The dcm4chee archive image we use is dcm4che/dcm4chee-arc-psql:5.22.1.This image requires two point points: The default mount points are highlighted in orange. The mount points can be changed as per preference as long as they are present on node3. Our port configuration

Ports for the webserver:

HTTP on port 8080.
HTTPS on port 8443.

Ports for wildfly administration:.

HTTP on port 9990.
HTTPS on port 9993.

DICOM AE-title port : 11112.

These ports should not be changed.

NOTE: By default, the nodes in the swarm are named node1, node2 and node3. You can name these however you want, as long as you maintain the names in docker-compose files and during the creation of the corresponding docker-machines.

---

Deployment Instructions:

1. Create a Docker swarm with 3 nodes. Name them node1, node2 and node3.

2. On node1 , run the following commands :

   cd /opt
   mkdir dcm4chee-arc
   cd dcm4chee-arc
   mkdir ldap
   mkdir slapd.d
   cd ~
   

3. On node2 , run the following commands:

   cd /opt
   mkdir dcm4chee-arc
   cd dcm4chee-arc
   mkdir db
   cd ~

4. On node3 , run the following commands:

  cd /opt
  mkdir dcm4chee-arc
  cd dcm4chee-arc
  mkdir wildfly
  mkdir storage
  cd ~

5. On node1 , clone this repository.

git clone https://github.com/kaiser-team/dcm4che-docker.git

6. Go into the cloned reporsitory and run :

docker stack deploy -c docker-compose.yml dcm4chee

Upon succesful deployment, DCM4CHEE will be on http://[docker-machine-ip]:8080/dcm4chee-arc/ui2.

Shutting down and removing the services

To stop all the services running on the swarm

docker service scale ldap=0 db=0 arc=0

To remove all the services running on the swarm

docker stack rm dcm4chee

Uploading data

You can use common C-STORE methods such as storescu, storescp or the dicomWeb module to upload data into dcm4chee. Be sure to set up the correct AE Titles within dcm4chee. You can do this by going to the ui of your local dcm4chee and then navigating to Configuration in the hamburger menu icon.

Store-scu command :

storescu +sd -aec DCM4CHEE [docker-machine-ip] 11112 {absolute path to folder}

Utilities

We have created a DICOM Utils library, located at: https://github.com/kaiser-team/dicom-utils