Image Reconstruction Algorithms for Optical Tomography

光学断层扫描图像重建算法

• 批准号: 7465432
• 负责人: JOHN C SCHOTLAND
• 金额: $ 34.85万
• 依托单位: UNIVERSITY OF PENNSYLVANIA
• 依托单位国家: 美国
• 财政年份: 2007
• 资助国家: 美国
• 起止时间: 2007-07-05 至 2010-04-30
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/7465432
• 关键词: Algorithms; Biological Models; Clinical Research; Collection; Complex; Computer Simulation; Data; Data Set; Development; Diffuse; Electrons; Fiber; Frequencies; Functional Imaging; Goals; Image; Lasers; Light; Lighting; Mammography; Measurement; Methods; Multimodal Imaging; Noise; Numbers; Optical Instrument; Optical Tomography; Optics; Performance; Physiologic pulse; Problem Solving; Property; Pulse taking; Research; Research Personnel; Sapphire; Signal Transduction; Source; Structure; Surface; System; Tail; Testing; Three-Dimensional Image; Time; Tissues; Validation; absorption; base; bioimaging; charge coupled device camera; data acquisition; design; detector; experience; image reconstruction; improved; instrument; instrumentation; optical fiber; programs; reconstruction; research study; simulation; success; tomography; tool; transmission process

DESCRIPTION (provided by applicant): The goal of the proposed research is to solve the problem of developing computationally efficient image reconstruction algorithms for noncontact optical tomography (OT) with large data sets. Specifically, we intend to devise fast algorithms for three-dimensional image reconstruction at the megavoxel level. To achieve this goal, we will exploit recent advances obtained by the Investigators on analytic methods for image reconstruction in OT. The algorithms will be systematically studied by computer simulations and validated by experiments in model systems. This approach allows the separation of effects due to algorithmic errors from those due to measurement errors. The planned "benchtop" experiments are the next logical step after algorithm development but before clinical studies and are critical to the success of this proposal. The specific aims are to (1) Construct instrumentation for multispectral noncontact optical tomography. This will entail upgrading an existing single-wavelength continuous-wave OT system. The planned enhancements include an electron multiplying CCD camera for faster data acquisition and improved signal to noise; illumination at three wavelengths using optically switched diode lasers; and time-resolved transmission measurements using a pulsed Ti-sapphire laser for spectroscopic measurements of absorption and scattering; (2) Implement and optimize fast image reconstruction algorithms for OT with large data sets. The planned studies will first focus on nonlinear reconstruction of the absorption. The focus will then shift to simultaneous linear and nonlinear reconstruction of absorption and scattering using multispectral and multifrequency data. The algorithms will be tested and optimized using numerical simulations prior to experimental validation in phantoms; (3) Evaluate the performance of the above reconstruction algorithms using phantoms. The reconstruction algorithms will be tested in a set of experiments which parallel the numerical simulations carried out under specific aim 2. Image quality will be assessed in tissue-equivalent phantoms. We will utilize the parallel plate imaging geometry which is often employed in optical mammography. The results of this research are expected to impact the design of clinically useful instruments for noncontact optical tomography.

描述(申请人提供)：本研究的目标是解决为大数据集的非接触式光学层析成像(OT)开发计算高效的图像重建算法的问题。具体地说，我们打算设计出用于兆像素级别的三维图像重建的快速算法。为了实现这一目标，我们将利用研究人员在OT图像重建的分析方法方面取得的最新进展。这些算法将通过计算机仿真进行系统的研究，并在模型系统中进行实验验证。这种方法允许将由算法误差引起的影响与由测量误差引起的影响分开。计划中的“台式”实验是在算法开发之后但在临床研究之前的下一个合乎逻辑的步骤，对这项提议的成功至关重要。具体目标是：(1)构建多光谱非接触光学层析成像仪器。这将需要升级现有的单波长连续波OT系统。计划中的改进措施包括：(1)安装电子倍增CCD相机，以加快数据采集速度，提高信噪比；(2)利用光学开关二极管激光器在三个波长进行照明；(2)利用脉冲钛蓝宝石激光器进行时间分辨的透射率测量，以便对吸收和散射进行光谱测量；(2)对大数据量的OT实施和优化快速图像重建算法。计划中的研究将首先集中在吸收的非线性重建上。然后，重点将转移到利用多光谱和多频率数据同时重建吸收和散射的线性和非线性。在体模实验验证之前，将使用数值模拟对算法进行测试和优化；(3)使用体模对上述重建算法的性能进行评估。重建算法将在一系列实验中进行测试，这些实验与在特定目标2下进行的数值模拟平行。图像质量将在组织等效模型中进行评估。我们将利用光学乳房摄影术中经常使用的平行板成像几何结构。这项研究的结果有望对临床上有用的非接触式光学断层扫描仪器的设计产生影响。