用遗传算法重构生物组织的光学特性参数  

RECONSTRUCTING OPTICAL PARAMETERS OF TISSUE WITH GENETIC ALGORITHM

作　　者：张航[1] 应金品[1] 周斌和[1] 刘正东[1] 孙威[1] 

机构地区：[1]浙江大学物理系,杭州玉泉310027

出　　处：《中国生物医学工程学报》2001年第5期430-435,共6页Chinese Journal of Biomedical Engineering

基　　金：国家自然科学基金No.39970 2 2 0;国家橡塑中心基金的资助

摘　　要：本文利用遗传算法 (GeneticAlgorithm)和蒙特卡罗 (MonteCarlo)方法研究由准直激光束通过均匀生物组织时所形成的多重散射光分布重构生物组织的光学特性参数。由于蒙特卡罗法模拟结果的离散性和统计性 ,在逆问题的求解时使传统的优化方法存在难以克服的困难 ,遗传算法可以有效地解决这个困难。我们用遗传算法成功地实现了对均匀生物组织单个光学特性参数重构和多个特性参数的联合重构 ,并取得了较好的重构效果。Optical parameters of tissue was reconstructed by Genetic Algorithm and Monte Carlo method according to the multiple scattering photon distribution, which emerged when a collimated laser beam normally incident upon and propagated through the tissue. Since the result of Monte Carlo simulation was discrete and statistic, the traditional optimize algorithm encountered irresolvable difficulty, but it could be overcome effectively by Genetic Algorithm. The reconstuction of single optical parameter and simultaneous reconstruction of multiple parameters were realized well. We provided theoretical guidance for nondestructive examination of biological tissue.

关 键 词：遗传算法 蒙特卡罗方法 重构 光学特性参数 光分布 

分 类 号：R318.03[医药卫生—生物医学工程] R312[医药卫生—基础医学]