Tools.
Last updated October, 2024
MIDRC has developed various tools to assist researchers in tasks such as cohort building, inter-repository indexing, automated interpretation of radiology reports, and selecting appropriate metrics for evaluating AI performance, among others. Tools are continuously being developed, so please check back often!
This page is dedicated to stand-alone tools. No programming skills required! Please see our ‘algorithms, code, methods’ page for other tools such as open-source material downloadable from GitHub.
The MIDRC Data Portal is a comprehensive platform designed to provide researchers with seamless access to a vast repository of medical imaging data. It features user-friendly search and filtering tools that enable efficient exploration of diverse datasets, facilitating advanced research and discovery in medical imaging. The portal supports various data formats and integrates with analytical tools to enhance research capabilities and streamline data management.
Please also check out the MIDRC seminar on using the data portal!
The MIDRC BDF Imaging Hub is an extensive platform that provides access to a wide array of imaging datasets and resources. It offers a centralized repository for medical imaging data, tools, and documentation, facilitating collaboration and data sharing among researchers. The platform supports advanced data search, visualization, and analysis, making it easier to conduct comprehensive research and develop insights from diverse imaging datasets.
RadGPT is an AI-powered tool designed to generate detailed radiology reports from medical images. Leveraging advanced natural language processing models, RadGPT assists radiologists by automatically interpreting imaging data and producing coherent, clinically relevant reports. This tool aims to enhance efficiency and accuracy in radiology, enabling healthcare professionals to focus on critical diagnostic tasks. RadGPT is available on the VILMEDIC platform, where users can explore its features and integrate it into their workflow.
The MIDRC-MetricTree is an interactive decision tree tool designed to assist researchers in evaluating and selecting appropriate performance metrics for medical imaging models. It guides users through a structured decision-making process, helping to identify the most relevant metrics based on specific research goals and requirements. This tool enhances the accuracy and effectiveness of model evaluation by providing tailored recommendations and insights into metric selection.
Read more in the MIDRC-MetricTree peer-reviewed publication.
The MIDRC AI/ML Bias Identification and Mitigation tool is a resource designed to help researchers detect and address biases in artificial intelligence and machine learning models used in medical imaging. It offers a structured approach to identifying potential sources of bias and provides strategies for mitigating them, ensuring more equitable and accurate outcomes. This tool supports the development of fairer and more reliable AI/ML systems by promoting best practices in bias management and model evaluation.
Read more in the peer-reviewed publication on AI fairness and bias.
Under development, available soon!
The MIDRC Helper AI tool is an advanced resource designed to assist researchers in navigating the extensive Medical Imaging and Data Resource Center (MIDRC) datasets. Through AI-driven search capabilities, it allows users to efficiently locate relevant imaging data, streamline the research process, and accelerate discoveries in medical imaging. The tool supports both basic and complex queries, making it accessible for a wide range of research needs, from initial data exploration to in-depth analysis.
Learn more in this video on MIDRC helper AI!
Under development, available soon!
The MIDRC DART (Data Analysis, Review, and Transformation) tool facilitates the integration and sharing of imaging data across different institutions within the MIDRC ecosystem. It provides a streamlined process for data curation, harmonization, and secure transfer, enhancing collaboration and research efficiency. The video highlights its features, including data quality checks, customizable workflows, and automated processes that ensure the integrity and consistency of shared imaging datasets.
Learn more in this video on MIDRC BIH indexing for restricted clinical trials dataI!
Under development, available soon!
The MIDRC Interoperability Tool between Imaging & Non-Imaging Data Commons is designed to facilitate seamless data integration and exchange between imaging and non-imaging data repositories. This tool ensures interoperability across different data commons, enhancing the ability to securely share, process, and analyze diverse datasets in healthcare and research environments, contributing to more comprehensive and collaborative research efforts.
Learn more in this video on MIDRC interoperabilityI!
Questions? Check out our answers to frequently asked questions!
How to acknowledge 1) MIDRC funded research and 2) use of data downloaded from the MIDRC Data Commons