TOPIC #411 - PHASE I SBIR CONTRACT - INTELLIGENT IMAGE ANONYMIZATION WITH XNAT

话题

• 批准号: 10274066
• 负责人: TIMOTHY OLSEN
• 金额: $ 39.97万
• 依托单位: RADIOLOGICS, INC.
• 依托单位国家: 美国
• 财政年份: 2020
• 资助国家: 美国
• 起止时间: 2020-09-16 至 2021-06-15
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10274066
• 关键词: Clinical Research; Clinical Trials; Computer software; Contracts; Data; Database Management Systems; Databases; Detection; Healthcare; Image; Industry Collaboration; Intelligence; Natural Language Processing; Phase; Radiology Specialty; Research; Resources; Risk; Secure; Services; Small Business Innovation Research Grant; Vision; base; clinical imaging; deep learning; design; image archival system; imaging informatics; neural network; prototype; research and development; tool

This Fast Track SBIR aims to implement comprehensive image anonymization within an enterprise imaging informatics platform built on XNAT. Our vision is for this platform to provide large healthcare enterprises with tools to generate secure research databases at scale that mirror their clinical image archives. These databases would then provide local academic and industry collaborators with a rich resource for clinical research and development of AI-powered applications. Thus, our proposed anonymization services are designed to be scalable, risk-based, and verifiable. The platform's AI-powered image anonymization will include automated detection of PHI using a deep learning based natural language processing engine and automated detection of PHI in image content using a convolutational neural network. The anonymization services will be integrated into Radiologics enterprise and clinical trial XNAT products.

此Fast Track SBIR旨在在基于XNAT构建的企业成像信息学平台内实施全面的图像匿名化。 我们的愿景是，该平台为大型医疗保健企业提供工具，以生成大规模的安全研究数据库，反映其临床图像档案。 这些数据库将为当地学术和行业合作者提供丰富的资源，用于临床研究和人工智能应用程序的开发。因此，我们提出的匿名化服务被设计为可扩展的，基于风险的和可验证的。该平台的人工智能图像匿名化将包括使用基于深度学习的自然语言处理引擎自动检测PHI，以及使用卷积神经网络自动检测图像内容中的PHI。 匿名化服务将集成到放射学企业和临床试验XNAT产品中。