Computational tools and algorithms for impedance imaging

电阻抗断层成像(EIT)使用体表电极的测量来计算内部电特性的图像。这些图像在医学(即血液和空气流动)、地球物理学(即地下水和矿物定位)和工业过程监测中提供了有用的信息。申请人是国际公认的EIT成像和功能图像分析算法专家。在过去的赠款期间，捐款通过a)领导关于算法和数据分析的共识项目，b)已成为该领域使用最广泛的开放源码软件工具，c)在科学文献(53篇期刊论文)中传播，以及d)指导21名研究生、4份PDF和60名本科生。研究团队开发的软件是三家公司产品的核心组件。 总体而言，研究将集中在两个方面：第一，一个健壮的算法和验证主题(由这笔赠款资助)，它将开发一个图像评估和“调试”框架，并利用开发的核心技术，一个应用主题(由合作者和工业合作伙伴资助)。 尽管取得了成功，但EIT一直被批评缺乏稳健性，因为图像重建固有的逆问题是病态的，因此小的数据或模型的不准确可能导致大的图像伪影。这项研究计划将通过开发一套工具来分析和纠正逆问题的解决方案来解决这个问题。虽然已经开发了几个单独的工具(由申请人和其他人开发)，但该项目的核心创新是一个集成的评估/调试软件框架。将开发三个新组件：模型和数据验证工具，电子硬件和电极的非理想特性模型(如放大器输入阻抗和串扰)，以及用于查询导致可疑伪像的数据的图像检查工具。 与此同时，将继续开展EIT技术在医疗和地球物理应用方面的应用工作。医疗应用将侧重于动态肺成像(用于新生儿、重症监护和阻塞性肺病患者)和血流动力学监测(无创血压和心输出量测量)。地球物理应用将侧重于监测不稳定(易发生滑坡)土壤和永久冻土区的地下水和地表移动。

Electrical impedance tomography (EIT) uses measurements at body surface electrodes to calculate an image of the internal electrical properties. Such images provide useful information in medicine (i.e. blood and air flow), geophysics (i.e. ground water and mineral localization), and industrial process monitoring. The applicant is an internationally recognized expert in algorithms for EIT imaging and functional image analysis. In the past grant period, contributions have been disseminated by a) leading of consensus projects on algorithms and data analysis, b) an open-source software tool which has become the most widely used in the field, c) in the scientific literature (53 journal papers), and d) supervising 21 graduate students, 4 PDFs and 60 undergraduate students. Software developed by the research team is a core component of three companies' products. Overall, research will focus on two streams: first, a Robust Algorithms and Validation theme (funded by this grant), which will develop an image evaluation and “debugging” framework, and, using the core technology developed, an Applications theme (funded by collaborators and industrial partners). Despite its successes, EIT has been criticized for a lack of robustness, as the inverse problem inherent to image reconstruction is ill-conditioned, so that small data or model inaccuracies can result in large image artefacts. This research program will address this issue by developing a set of tools to analyse and correct inverse problem solutions. While several individual tools have been developed (by the applicant and others), the core innovation of this project is an integrated evaluation/debugging software framework. Three new components will be developed: a model and data validation tool, models of non-ideal characteristics of electronics hardware and electrodes (such as amplifier input impedance and crosstalk), and an image inspection tool to query the data contributing to suspected artefacts. At the same time, work will be pursued on applications of EIT technology for medical and geophysical applications. Medical applications will focus on dynamic lung imaging (for use with neonatal, intensive care, and obstructive lung disease patients) and hemodynamic monitoring (non-invasive blood pressure and cardiac output measures). Geophysical applications will focus on monitoring of groundwater and surface movement in unstable (landslide-prone) soils and permafrost regions.