PARADIM

PARADIM supports research at the interface of data science and medical imaging through a digital infrastructure built on open-source components with strict compliance to the DICOM standard and FAIR principles.

Key Features:

• Data Curation and Governance: The platform provides functionalities for data curation, annotation, and governance with rich metadata generation throughout the research lifecycle.
• Privacy and Security: PARADIM includes user identity and access management and data de-identification capabilities.
• Model Training and Evaluation: The platform supports model training, evaluation, and automation of large-scale computationally intensive pipelines including automatic segmentation, dose calculations, and AI model evaluation.
• Scalable Resource Management: PARADIM offers storage and scalable resource management for computationally intensive workflows.
• Interoperability: The platform ensures interoperability through strict compliance with the DICOM standard and adherence to FAIR principles.
• Traceability and Reproducibility: Rich metadata are generated all along the research lifecycle to ensure traceability and reproducibility of results.
• Medical Image Collections: PARADIM hosts several medical image collections for research use.

Scientific Applications:

• Automatic Segmentation: PARADIM enables automation of large-scale automatic segmentation pipelines on medical imaging data.
• Dose Calculations: The platform supports automated dose calculation workflows in medical imaging research.
• AI Model Development and Evaluation: PARADIM facilitates the development, evaluation, and deployment of AI models for medical imaging applications.
• Research Data Management: The platform addresses research data management best practices for data science and medical imaging research.

Methodology:

PARADIM is built from open-source components with strict compliance to the DICOM standard and rooted in FAIR principles; it generates rich metadata throughout the research lifecycle and automates large-scale, computationally intensive pipelines.