论文中文题名: | 基于虚拟源模型的荧光扩散断层成像算法研究 |
姓名: | |
学号: | 201407356 |
学科代码: | 085208 |
学科名称: | 电子与通信工程 |
学生类型: | 工程硕士 |
学位年度: | 2017 |
院系: | |
专业: | |
第一导师姓名: | |
论文外文题名: | Research on the algorithm of fluorescence diffuse optic tomography based on virtual source model |
论文中文关键词: | |
论文外文关键词: | |
论文中文摘要: |
荧光散射断层成像(FDOT)是光学层析扫描与荧光显影剂结合发展出的新兴三维医学成像技术,对癌症的早期诊断具有非常重要的意义。
本文提出了利用平面镜反射的原理对现有常规单角度FDOT系统进行优化的方法,有效解决了现有的多角度FDOT成像系统普遍存在特异性高、不易推广、造价高等问题。FDOT分子成像技术包括正向问题和逆向问题。本文对于FDOT的正向问题,介绍了扩散近似模型以及利用有限元方法对该模型的求解。FDOT的逆向问题即是重建问题,在正向的基础上,本文采用虚拟源模型的图像重建算法,利用小波变化方程对采集信息进行数据压缩,降低了重建的时间,并引入先验信息,重建中使用正则化约束,降低重建误差。通过仿真实验,验证了该优化方法的可行性,最后分别在单角度和优化后的多角度系统上进行了多组动物对比实验,同时选择不同平面镜角度进行FDOT多角度实验,并利用TOAST++工具箱进行图像重建。
仿真结果和动物实验数据表明,相较于常规单角度成像系统平面镜多角度FDOT系统重建出的肿瘤断层图像完整,能够很好地反映肿瘤的分布,重建的效果从成像时间、信噪比、误差几个方面都有所降低,提升了重建图像的清晰度和分辨率,多角度FDOT系统成象在平面镜为左右时成像效果最佳,这使得FDOT分子成像技术能更早的应用于肿瘤临床前研究和临床实践。
﹀
|
论文外文摘要: |
Fluorescence diffuse optic tomography (FDOT) is a new 3D medical imaging technology developed by combining optical tomography with fluorescence. It is important for the early diagnosis of cancer.
This paper describe the existing conventional single angle FDOT system optimization method based on the principle of the plane mirror, effectively solve the multi angle imaging system of FDOT current, high specificity and high cost problem is not easy to promote.
A method of FDOT system optimization is described in this paper which is using the principle of plane mirror reflection based on classical single angle FDOT system. Compared with the existing multi-angle FDOT system, it solved the problems of high specificity, difficult popularization and high cost. FDOT imaging includes forward problem and inverse problem. For the forward problem, this paper described a diffusion approximation equation, and introduced the finite element method to solve the equation. The inverse problem of FDOT is the image reconstruction problem. The wavelet transform is used to compress the data, which reduces the time of reconstruction. And a priori information is introduced in the reconstruction. The regularization constraint is used to reduce the reconstruction error. Through the simulation experiment, proved the feasibility of the optimization method. Finally, in terms of single angle and the optimized system were carried out animal experiments, and choose a different plane mirror angle FDOT multi angle experiment, the images were reconstructed using TOAST++ toolbox.
The results of simulation and animal experiment data show that the tumor reconstruction image of multi-angle FDOT system compared with the classical single angle FDOT system is complete which can better reflect the distribution of tumor. The effect of reconstruction is reduced from imaging time, noise-signal ratio and reconstruction error, and enhance the definition and resolution of the reconstructed image. The imaging effect is best when the plane mirror angle is about 40 degrees. This makes the FDOT molecular imaging technology earlier applied in preclinical and clinical studies.
﹀
|
中图分类号: | TP391.413 |
开放日期: | 2017-06-19 |