基于扰动算法的组织光学断层图像重建  

Reconstruction of tissue diffusion optical tomography using perturbation method

作　　者：奉华成[1] 白净[1] 宋小磊[1] 

机构地区：[1]清华大学生物医学工程系,北京100084

出　　处：《清华大学学报（自然科学版）》2007年第12期2196-2199,共4页Journal of Tsinghua University(Science and Technology)

基　　金：国家自然科学基金资助项目(30670577;60571013;60331010);国家"九七三"重点基础研究基金项目(2006CB705700);清华-裕元基金资助项目

摘　　要：为了克服扰动算法初始值选择的局限性,提出了定性重建组织的吸收系数图像的预定标技术。根据由初始猜测值求出的边界上的正问题的解与实际边界测量值的比例关系,对吸收系数的初始值进行修正,使之达到一个合适的值之后再用于求逆运算,并在迭代过程中保持漫射系数的值不变。仿真实验中对组织吸收系数图像在采用和不采用预定标技术的情况下分别进行重建。结果表明,在初始值偏离真实解较远时,只有采用预定标技术才能得到预期结果。可见,预定标技术是对传统扰动算法的一种有效的改进。A prescaling technique was developed to qualitatively reconstruct the tissue absorption maps which can overcome the limitation that the initial estimate must be close enough to the ideal solution when the perturbation method is employed. Before beginning the iterations, the estimated absorption coefficient was scaled to an appropriate value which was then used in the reverse process, with the diffusion coefficient kept constant throughout the iterations. Reconstructions of the tissue absorption map with and without the prescaling technique show that the prescaling technique is needed when the initial estimate is far from the true solution.

关 键 词：组织光学 漫射光学断层 扰动算法 图像重建 

分 类 号：R318[医药卫生—生物医学工程]