OBJECT DEFINITION IN TOMOGRAPHIC RADIOLOGY

断层放射学中的对象定义

• 批准号: 6092174
• 负责人: JAYARAM K UDUPA
• 金额: $ 5万
• 依托单位: UNIVERSITY OF PENNSYLVANIA
• 依托单位国家: 美国
• 财政年份: 1997
• 资助国家: 美国
• 起止时间: 1997-12-01 至 2000-11-30
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/6092174
• 关键词: bioimaging /biomedical imaging; computational neuroscience; computer assisted diagnosis; computer system design /evaluation; human data; image enhancement; magnetic resonance imaging; multiple sclerosis; tomography

DESCRIPTION (Adapted from Applicant's Abstract): The broad long-term goal of this project was to develop theories, algorithms, and their portable computer implementations for quantifying object information captured in multidimensional images and to apply them to solve specific biomedical problems. The lack of cost-effective methods with proven precision and accuracy for extracting object information from images remains one of the major impediments in many radiological applications. With this in mind, the applicants proposed: (1) to advance the theory, algorithms, and their efficient computer implementations for detecting and delineating objects in multidimensional, multiparametric images in general, and (2) to apply these methods to the problem of quantifying MS lesions of the brain via MR imagery. The methods proposed in this application are based on a theory relating to definition of objects in images. Since images are by nature fuzzy, the theory considers objects as a set of image elements that "hang-together" fuzzily. A fuzzy topological concept called fuzzy connectedness is introduced that captures the idea of "hanging-togetherness." Although the concept is computationally impractical, key theoretical results lead to practical computer algorithms for detecting fuzzy objects in images. MS is an intensively studied disease of the nervous system. Its detection and quantification via MR images has proved critical to the monitoring of this disease and of its response to therapy. However, a practical solution to this problem is still not available. Supported by strong preliminary results, this application proposes practical methods based on fuzzy connectedness for the detection and quantification of MS lesions. The overall hypothesis underlying this research is that the methods resulting from this investigation will be practical, cost-effective, and reliable for quantifying MS lesions with a precision and accuracy that is acceptable for conducting clinical trials.

描述（改编自申请人的摘要）：广泛的长期目标 该项目的目的是开发理论、算法及其可移植性 用于量化捕获的对象信息的计算机实现 多维图像并应用它们来解决特定的生物医学问题 问题。 缺乏具有成本效益且精度经过验证的方法 从图像中提取物体信息的准确性仍然是影响因素之一 许多放射学应用中的主要障碍。 考虑到这一点， 申请人提出：（1）推进理论、算法及其应用 用于检测和描绘物体的高效计算机实现 一般而言，多维、多参数图像，以及 (2) 应用这些图像 通过 MR 量化脑部多发性硬化症病变问题的方法 图像。 本申请中提出的方法基于理论 与图像中对象的定义有关。 由于图像本质上是 模糊，该理论将对象视为一组图像元素， 模糊地“挂在一起”。 模糊拓扑概念称为fuzzy 引入了连通性，捕捉了以下想法： “挂在一起”。 虽然这个概念是计算出来的 不切实际的关键理论结果导致实用的计算机算法 用于检测图像中的模糊对象。 MS 是一种经过深入研究的疾病 神经系统的。 通过 MR 图像对其进行检测和量化 事实证明对于监测这种疾病及其对疾病的反应至关重要 治疗。 然而这个问题目前还没有切实可行的解决方案 可用的。 该应用程序得到了强有力的初步结果的支持 提出了基于模糊连通性的实用检测方法 和 MS 病变的量化。 这背后的总体假设 研究表明，本次调查得出的方法将是 实用、经济高效且可靠，可用于量化 MS 病变 进行临床试验可接受的精密度和准确度。