COMPUTER-ASSISTED CHEST RADIOGRAPH READER

计算机辅助胸部X线读片机

• 批准号: 6445973
• 负责人: Peter N Soliz
• 金额: $ 35.4万
• 依托单位: KESTREL CORPORATION
• 依托单位国家: 美国
• 财政年份: 2000
• 资助国家: 美国
• 起止时间: 2000-09-30 至 2002-09-29
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/6445973
• 关键词: artificial intelligence; bioimaging /biomedical imaging; biomedical equipment development; clinical biomedical equipment; computer assisted diagnosis; diagnosis quality /standard; digital imaging; image processing; mass screening; thoracic radiography

The long term objective of this phase is to provide a means for reducing inter- and intra- reader variability in diagnosing interstitial lung diseases in chest radiographs through a computer-based system for analyzing digital images. The Computer-assisted Chest Radiograph Reader System (CARRS) applies recognized principles in the psychophysics of human vision, incorporates neural network-based image analysis and integrates these with a graphical user interface. Advances in digital image processing, and classification techniques have made CARRS feasible for meeting screening, research arid development, and clinical requirements. CARRS will implement the International Labor Organization (ILO) classification procedures. The specific aims of this project are to implement enhancements to the CARRS prototype developed in Phase I and to validate the advanced version with several hundred chest radiographs. PROPOSED COMMERCIAL APPLICATION: Today, there exists a need for an automated chest radiograph diagnostic system to screen the thousands of images collected daily at radiological service centers and hospitals worldwide and throughout the United States. A computer-based system that eliminates or reduces inter- and intra-reader variability significantly is required to improve the management and early diagnosis of the disease. CARRS would be marketed and sold to most radiological services centers and hospitals worldwide and throughout the U.S.

该阶段的长期目标是提供一种方法，通过基于计算机的数字图像分析系统，减少胸片中诊断间质性肺疾病时读者间和读者内的差异。计算机辅助胸片阅读器系统 (CARRS) 应用人类视觉心理物理学中公认的原理，结合基于神经网络的图像分析，并将其与图形用户界面集成。数字图像处理和分类技术的进步使 CARRS 能够满足筛选、研究、开发和临床要求。 CARRS 将执行国际劳工组织 (ILO) 的分类程序。该项目的具体目标是对第一阶段开发的 CARRS 原型进行增强，并通过数百张胸部 X 光照片验证高级版本。拟议的商业应用：如今，需要一种自动胸部放射线诊断系统来筛选每天在世界各地和美国各地的放射服务中心和医院收集的数千张图像。需要一种基于计算机的系统来显着消除或减少读者之间和读者内部的变异性，以改善疾病的管理和早期诊断。 CARRS 将向全球和美国各地的大多数放射服务中心和医院进行营销和销售。