Elements: MVP: Open-Source AI-Powered MicroVessel Processor for Next-Generation Vascular Imaging Data

点击翻译按钮获取中文摘要

Effectively acquiring and processing complicated microstructures and their morphology become increasingly important in scientific and engineering discoveries. A growing body of evidence in animal and human studies shows that micro-cerebrovascular abnormalities are the real source of many neurologic disorders and vascular diseases. Therefore, there is an urgent need for better detection and understanding of vascular characteristics in vivo at the micro-level. However, the existing techniques and software solutions are developed based on general-scale medical images for processing macro-level organs and tissues, not applicable to the challenging next-generation vascular imaging data and micro-level vasculature. Furthermore, they are difficult to be customized to specific scientific microvascular imaging applications for domain developers. The overall objective of this project is to design and develop a rigorous, scalable, intelligent, and comprehensive data and software infrastructure, MicroVessel Processor (MVP), which supports and sustains advancements of understanding and analyzing the complicated 3D microvascular networks, provides services to a large community, and fosters continuous innovation in the multidisciplinary domains.This project integrates new paradigm of data-driven 3D microvascular discovery - generate new knowledge and understanding and accelerate discovery and innovation via the advanced software cyberinfrastructure, along with use-case driven functional applications and evaluations. The capabilities of the newly developed AI-based high-fidelity 3D image enhancement, 3D geometric segmentation / extraction, reconstruction, and analytics techniques are integrated in the MVP system as the core modules in the software infrastructure. The MVP database including a large number of subjects with varying anatomical shapes, morphological features, and different data modalities is collected and constructed as the infrastructure (i.e., knowledgebase), which provides the knowledge for 3D microvascular networks tracking and analysis. The MVP system is designed and built based on an open-source and drag-drop real-time application system, in order to reduce both human labors and computation costs, accelerate the speed and efficiency of re-development and extensibility. This unique framework supports a wide adoption in microvasculature analytics in biomedical engineering, big-data neuroscience, and AI-based clinical diagnosis. This award by the Office of Advanced Cyberinfrastructure is jointly supported by the Division of Chemical, Bioengineering, Environmental and Transport Systems within the Engineering Directorate.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.