基于压缩感知的新型碳纳米球管显微CT重建算法研究

项目名称： 基于压缩感知的新型碳纳米球管显微CT重建算法研究

项目编号： No.61201117

项目类型： 青年科学基金项目

立项/批准年度： 2013

项目学科： 电子学与信息系统

项目作者： 郑健

作者单位： 中国科学院苏州生物医学工程技术研究所

项目金额： 25万元

中文摘要： 近年来，采用碳纳米管作为阴极射线源的新型X射线球管（碳纳米球管）因其具有微型化和快点火等优点正受到越来越多的关注。但目前碳纳米球管CT存在因功率低导致投影数据噪声较大等不足，此外随着美国食品与药品管理局（FDA）对CT扫描辐射剂量安全性要求的不断提高，采用传统的滤波反投影算法及迭代算法对碳纳米球管CT进行重建的效果都不理想。基于此，本项目拟进行基于压缩感知的新型碳纳米球管显微CT(Micro-CT)系统的重建算法研究，主要对系统成像模型、投影数据的噪声特性、基于压缩感知的成像目标优化函数及优化算法等进行深入研究，探索出效果较好的高时间分辨率、高空间分辨率及低剂量的碳纳米球管显微CT重建算法，缩小与国外研究者在这一领域的差距并力争在算法理论上能有所突破，为下一代临床诊断CT的研制提供重要的研究基础。

中文关键词： 压缩感知；迭代重建；金属伪影校正；字典学习；

英文摘要： In recent years, a novel X-ray tube which uses the carbon nanotubes (CNT) as the cathode-ray source has attracted more and more attention. The CNT X-ray tube has great advantages of miniaturization and instant electron emission. Currently, the CNT X-ray tube CT still has some disadvantages such as greater noise in projection data since the low tube current as well as the scatter effect. Moreover, considered that the U.S. Food and Drug Administration (FDA) are continuously improving the safety standards of radiation dose for the CT scan, conventional filtered back projection (FBP) based reconstruction algorithms and iterative reconstruction algorithms are not effective for the CNT X-ray tube CT. To address these problems, we intend to peform a deep research about the reconstruction algorithm based on compressed sensing (CS) theory for a novel Micro-CT system using CNT X-ray tube. We will mainly focuse on the system imaging model, the noise characteristics of the projection data, reconsturction objective function based on CS theory as well as fast and accurate optimization strategies. The aim of this study is to explore an effective reconstruction algorithm for CNT X-ray tube Micro-CT, which will be capable of performing a CT scan with high temporal resolution, high spatial resolution and low radiation dose. Als

英文关键词： compressed sensing；iterative reconstruction；metal artifact reduction；dictionary learning；