Back to All Open Science Grantees
Enhancing the Open Health Imaging Foundation Web Medical Imaging Framework
Proposal Summary
To develop training materials, perform software maintenance, expand outreach, and provide community support for the Open Health Imaging Foundation (OHIF) web-based medical imaging framework including its underlying libraries (e.g., Cornerstone).
Project
The OHIF Viewer is an extensible web viewer framework for medical imaging. It offers a professionally-designed user interface that makes it easy to operate by end users, and provides basic image review functionality (e.g. image manipulation and measurement) as well as advanced visualization (e.g. multiplanar reformatting). It is written as a client-only, single-page web application which can easily be embedded into third-party applications or hosted as a standalone website. The platform provides extension points for software developers to include custom tools or visualizations and adapt the system for their workflows. It is standards compliant and relies on DICOMweb for data exchange and OpenID Connect for authentication, but can be configured to use alternatives. Additionally, the user interface (UI) components are provided in a standalone library so that developers can create custom extensions. Extensions have been built to provide support for 3D image segmentation, radiation therapy image annotations, and visualization of whole-slide microscopy images using Google Maps-style tiling (i.e. digital pathology). The core business logic, as well as community-contributed internationalization files for the UI components, can be used independently through separate NPM packages. OHIF relies on Cornerstone by default for DICOM image retrieval, decoding, rendering, and interactive tool support.
Key Personnel
Project
Cornerstone is a set of JavaScript libraries that can be used to build web-based medical imaging applications. It is domain-specific software designed for radiology tasks (e.g. windowing high bit depth images, decoding compressed DICOM images, applying non-linear value of interest lookup tables, supporting standard metadata and pixel data retrieval APIs). It supports CPU and GPU-based rendering, and a pluggable image loader framework to allow developers to support custom formats. It provides support for DICOMweb out of the box, including the ability to retrieve individual image frames, which is critical for performance when viewing large multiframe images (e.g. for breast tomosynthesis). All DICOM transfer syntaxes are supported through cross-compiled codec libraries such as OpenJPEG, and images are decoded inside background (WebWorker) threads to ensure that the user interface continues to function fluidly. Cornerstone tools provide a framework for developing measurement and annotation tools which can be used for image analysis on desktop and mobile devices. It includes tools such as interactive window/leveling, measurements such as length and angle, ROI measurements (e.g. rectangle, ellipse), segmentation brush and scissors tools, as well as a magnifying glass and CINE tools, gesture-controls (e.g. pinch-to-zoom), among others.