图像质量评价模型引导的CT迭代重建方法研究

CT是一类重要的临床诊断设备，CT图像在病人医疗诊治全过程都发挥作用，而图像质量评价（IQA）是其中的重要问题。本研究针对其中的图像重建算法和医生诊断等两个环节中的IQA问题开展研究工作，分析这两个环节的IQA标准的不同之处，指出采用IQA模型引导图像重建计算时存在的问题；由于迭代重建是目前低剂量和其它不完备投影数据CT重建的主要计算方法，其中的正则化参数选择是决定图像质量的一个关键因子，同时其自适应选择仍然是一个难题；本项目研究基于IQA模型引导的CT迭代重建算法，可以自适应地选择正则化参数而不增加计算量，同时提出感知特征增强的迭代重建算法，利用通用IQA模型引导CT重建时，可以获得满足医生诊断质量评价优化的重建结果。本项目的意义在于首次研究以图像诊断价值为优化目标的迭代重建算法，提出可自动调整重建参数的高效计算模型，提升CT设备重建算法的效能。

As an important clinical diagnostic equipment, Computed Tomography (CT) plays a role in the whole process of patient medical diagnosis and treatment, among which image quality assessment (IQA) is an important issue. This project focuses on the IQA issues in the aspects of image reconstruction algorithm and the process of clinical diagnosis. In this study, we will analysis the differences between the IQA standards in these two aspects and explore the problems in using IQA model to guide image reconstruction calculation. .Since the iterative reconstruction algorithm dominates the scenarios of low-dose and incomplete projection data in CT reconstruction, in which the regularization parameter selection is a key ingredient to determine the image quality. However, it still has difficulty in selecting the optimal parameter adaptively. This project plans to study an IQA model-based CT iterative reconstruction algorithm, which can adaptively select the regularization parameters without increasing the computational amount. Moreover, by using the general IQA model to guide the CT reconstruction process, we will propose a feature-aware iterative reconstruction framework to make the results satisfy the doctors' diagnostic quality. The significance of this project could be concluded that we will research a diagnostic quality-oriented CT iterative reconstruction algorithm for the first time. Based on this, an efficient model will be proposed that can automatically adjust reconstruction parameters, improving the efficiency of CT reconstruction algorithm.