MedAP Contour Editor

MedAP Contour Editor is annotation platform for medical image segmentation. The platform utilizes deep learning segmentation models for pre-annotation process that reduces the annotator efforts. MedAP performs pre-annotation and stored the original image, segmentation mask and annotated image with created outer contour.

Main MedAP components:

• GUI - Tkinter
• Deep Learning model - MUCSNet (modified MicroSegNet model for segmentation of whole prostate sweep)

---

Requirements

• Docker installed (Install Docker)
• NVIDIA GPU with drivers installed
• NVIDIA Container Toolkit (Install Guide)
• X11 (on Linux) to forward GUI from container
• Download 'networks' folder (run setup.sh script)
• Download model from link and name the file 'MUCSNet.pth'

---

Build Docker Image

docker build --network=host -t docker-image:tag .

xhost +local:docker

docker run -it \
  -e DISPLAY=$DISPLAY \
  -v /tmp/.X11-unix:/tmp/.X11-unix \
  --gpus all \
  docker-image:tag