Skip to main content

Introduction

Getting Started​

MONAILabel Server represents a significant advancement in the realm of medical image analysis, not only expanding the capabilities of the MONAILabel platform but also integrating seamlessly with QuPath, a prominent image analysis software. This extension now equips researchers and medical professionals with a comprehensive solution for accurate and automated segmentation of critical structures, including blood vessels, inflammation regions, and endocardium, within Whole Slide Imaging (WSI) data obtained after cardiac biopsy, stained with H&E or SRel. With the combined power of MONAILabel and QuPath, users can leverage cutting-edge AI models to streamline their analysis workflow, expedite cardiac rejection detection, and drive further insights in the field of digital pathology.

What you'll need​

To utilize MONAILabel Server effectively, you'll need the following prerequisites:

  • Python or Docker installed,

  • Cuda Toolkit at least v11.2,

  • QuPath at least v0.4.0,

  • GPU with CUDA support (This prerequisite is not supported on computers witn macOS).

Minimal requirements​

Minimum hardware requirements:

  • 8GB CPU RAM

  • Graphics card supporting CUDA - list

  • Graphics card with a minimum of 2GB memory

MONAILabel Server​

MONAILabel Server is a project born from the collaboration between Ivan Vykopal's Master thesis at the Faculty of Informatics and Information Technology (FIIT), Slovak University of Technology (STU), and Tomáš Matejov's Bachelor thesis. The server serves the specific purpose of aiding in the identification of cardiac allograft rejection. Additionally, this initiative was undertaken in close collaboration between medical professionals from the Institute for Clinical and Experimental Medicine (IKEM) and students and researchers from FIIT STU. This project was implemented based on an existing project from MONAI, specifically MONAILabel.