README.md

CVAT MONAILabel extension

Requirement

Install CVAT and enable Semi-Automatic and Automatic Annotation

• https://opencv.github.io/cvat/docs/administration/basics/installation/
• https://opencv.github.io/cvat/docs/administration/advanced/installation_automatic_annotation/

Reference Guide for installing CVAT

git clone https://github.com/opencv/cvat
cd cvat
git checkout v2.1.0

# use real-ip instead of localhost if you want to share it on network
export CVAT_HOST=127.0.0.1
export CVAT_VERSION=v2.1.0

docker-compose -f docker-compose.yml -f components/serverless/docker-compose.serverless.yml up -d
docker exec -it cvat bash -ic 'python3 ~/manage.py createsuperuser'

wget https://github.com/nuclio/nuclio/releases/download/1.5.16/nuctl-1.5.16-linux-amd64
chmod +x nuctl-1.5.16-linux-amd64
sudo mv nuctl-1.5.16-linux-amd64 /usr/bin/nuctl

Installation

Run ./deploy.sh to install all available models from MONAI Label into CVAT.

# all functions (endoscopy, pathology)
./deploy.sh

# to deploy specific function
./deploy.sh endoscopy

# to deploy specific function and model
./deploy.sh endoscopy tooltracking

Currently, following sample models are available for CVAT.

Endoscopy

• ToolTracking (Detector)
• DeepEdit (Interactor)

Pathology

• Segmentation Nuclei (Detector)
• Deepedit Nuclei (Detector)
• NuClick (Interactor)

Using Plugin

Currently we can use MONAI Label models only for annotation. Other features like ActiveLearning, Finetuning/Training models is not supported.

Annotation functions are verified only on basic png/jpeg images in CVAT.

Models

Detector

Interactor

Currently CVAT supports single polygon as result for Interactor. Hence, NuClick model in CVAT will return only one polygon mask.