Accelerating Community-Driven Medical Innovation with VTK

Kitware, Inc., Clifton Park NY

Investigators

Linked publications & trials

Abstract

Abstract Tens of thousands of medical researchers and practitioners around the world use VTK—the Visualization Toolkit—an open source, freely available software development toolkit providing advanced 3D interactive visualization, image processing and data analytics algorithms. They either use VTK directly in their in-house research applications or indirectly via one of the multitude of medical image analysis and bioinformatics applications that is built using VTK: 3D Slicer, Osirix, BioImageXD, MedINRIA, SCIRun, ParaView, and others. Furthermore, VTK also provides 3D visualizations for clinical applications such as BrainLAB’s VectorVision surgical guidance system and Zimmer’s prosthesis design and evaluation platform. VTK is currently accessed at a rate of more than 2.5 million/downloads per year, and remains a vital system even thirty years after its initial release in 1993. Considering its broad distribution and prevalent use, it can be argued that VTK has had a greater impact on medical research, and patient care, than any other open-source medical computing package. This proposal is in response to the multitude of requests we have been receiving from the VTK medical and biomedical computing communities. The aims of this work are as follows: Aim 1: Ubiquitous Visual Analytics: Establish an easy-to-use workflow so that users and developers can write VTK-based data analytics tools that are automatically transformed into portable, cross-platform executable modules. These modules–written just once–can then be used in a portable manner on any computing platform (desktop, cloud and web-based) and imported into applications. Aim 2: AI-Ready Visualization: Extend VTK’s current support of AR/VR to integrate AI into immersive workflows such as surgical and point-of-care applications. Provide visualization techniques, including saliency maps and Grad-CAM, to build trust in and validate AI workflows. Employ AI to drive the data transformation and rendering processes such as adaptively configuring transfer functions to highlight and efficiently convey clinically important information. Aim 3: Enhanced Community Support: Enable users and developers to more rapidly learn about, develop code for, and deliver VTK-based visual analytics applications. Using open-source LLM AI /ML models, train these models on the decades of community resources and millions of lines of VTK source code to provide rapid collation of relevant information, and enable the generation of programming extracts to accelerate the creation of specialized workflows.