支持向量机测定组织光学参数的仿真实验研究
|
摘要
对新近发展的支持向量回归机确定组织光学参数的特性和规律进行了仿真实验研究。首先通过蒙特卡洛模拟获得与336组人体组织的光学特性参数相对应的组织表面漫反射光分布作为样本数据集,并将其分为训练集和测试集两部分。通过建立不同条件下的组织表面漫射光分布与光学参数之间的支持向量回归模型,研究了支持向量机用于测定组织光学参数时,训练集样本的个数、训练集与测试样本的关系以及数据预处理方法等因素对测量精度的影响。结果表明,在小样本条件下μt和μeff的平均相对误差分别为0.98%和4.34%;支持向量回归机用于组织光学参数测定不仅具有较高的精度,而且对样本数具有很好的适应性。
Characteristics and laws of optical parameters measuring by recent development of support vector regression algorithm are simulated and studied. Firstly 336 optical parameter samples and the corresponding diffuse reflectance distribution are obtained by Monte Carlo simulation, and the samples are divided into training set and testing set. By establishment of support vector machines(SVM) models between optical parameters and diffuse reflectance distribution in different conditions, various factors on prediction accuracy of SVM are researched, including the numbers of training set, the relationship between training set and testing set and the data preprocessing methods. Results show that in small sample condition the maximum average relative error ofμtand μeffare 0.98% and 4.34% respectively. The SVM used for determination of tissue optical parameters has high accuracy and good adaptability to the number of samples.
Characteristics and laws of optical parameters measuring by recent development of support vector regression algorithm are simulated and studied. Firstly 336 optical parameter samples and the corresponding diffuse reflectance distribution are obtained by Monte Carlo simulation, and the samples are divided into training set and testing set. By establishment of support vector machines(SVM) models between optical parameters and diffuse reflectance distribution in different conditions, various factors on prediction accuracy of SVM are researched, including the numbers of training set, the relationship between training set and testing set and the data preprocessing methods. Results show that in small sample condition the maximum average relative error ofμtand μeffare 0.98% and 4.34% respectively. The SVM used for determination of tissue optical parameters has high accuracy and good adaptability to the number of samples.
引文
[1]GHOSH N,MOHANTY S K,MAJUMDER S K,et al.Measurement of optical transport properties of normal and malignant human breast tissue[J].Appl Opt,2001,40:176-184.
[2]AMELINK A,STERENBORG H J.Measurement of the local optical properties of turbid media by differential path-length spectroscopy[J].Appl Opt,2004,43(15):3048-3054.
[3]KIENLA A,LILGE L,PATTERSON M S,et al.Spatially resolved absolute diffuse reflectance measurements for noninvasive determination of the optical scattering and absorption coefficients of biological tissue[J].Appl Opt,1996,35(13):2304-2314.
[4]ISHAN BARMAN,NARAHARA CHARI DINGARI,NARASIMHAN RAJARAM,et al.Rapid and accurate determination of tissue optical properties using least-squares support vector machines[J].Biomedical Optics Express,2011,2(3):593-599.
[5]FARRELL T J,WILSON B C,PATTERSON M S.A diffusion theory model of spatially resolved steady state diffuse reflectance for the noninvasive determination of tissue optical properties in-vivo[J].Med Phys,1992,19:879-888.
[6]MARTELLI F,SASSAROLI A,BIANCO S D,et al.Solution of the time-dependent diffusion equation for layered diffusive media by the eigenfunction method[J].Physical Review E,2003,67(5):1-14.
[7]LI C-X.Artificial neural network method for determining optical properties from double-integrating-spheres measurements[J].C Opt Letters,2010,10:174-176.
[8]FARRELL T J,WILSON B C,PATTERSON M S.The use of a neural network to determine tissue optical properties from spatially resolved diffuse reflectance measurements[J].Physics in Medicine and Biology,1992,37(12):2281-2286.
[9]CORTES C,VAPNIK V N.Support vector networks[J].Machine Learning,1995,20(3):273-297.
[10]OSUNAS E,FREUND R,GUOSI F.Training Support Vector Machines:An Application to Face Detection[C]//Proceedings of IEEE Conference on CVPRE.Puerto Ricl,1997:130-136.
[11]PLATT J.Fast Training of Support Vector Machines using Sequential Minimal Optimization[C]//Advances in Kernel Methods.MIT Press,1999:185-208.
[12]MACIEJ ZIEBA,JAKUB M TOMCZAK,MAREK LUBICZ.Boosted SVM for extracting rules from imbalanced data in application to prediction of the post-operative life expectancy in the lung cancer patients[J].Appl Soft Comput J,2013,7:99-108.
[13]Support Vector Machines Software[EB/OL].[2014-04-10].http://www.csie.ntu.edu.tw/:cjlin/libsvm.
[14]Monte Carlo Simulation Software[EB/OL].[2014-04-10].http://oilab.tamu.edu/mc.html(accessed Feb 5,2005).
[15]XU HEPING,PATTERSON MICHAEL S.Determination of the optical properties of tissue-simulating phantoms from interstitial frequency domain measurements of relative fluence and phase difference[J].Optics Express,2006,14(14):6485-6501.
[2]AMELINK A,STERENBORG H J.Measurement of the local optical properties of turbid media by differential path-length spectroscopy[J].Appl Opt,2004,43(15):3048-3054.
[3]KIENLA A,LILGE L,PATTERSON M S,et al.Spatially resolved absolute diffuse reflectance measurements for noninvasive determination of the optical scattering and absorption coefficients of biological tissue[J].Appl Opt,1996,35(13):2304-2314.
[4]ISHAN BARMAN,NARAHARA CHARI DINGARI,NARASIMHAN RAJARAM,et al.Rapid and accurate determination of tissue optical properties using least-squares support vector machines[J].Biomedical Optics Express,2011,2(3):593-599.
[5]FARRELL T J,WILSON B C,PATTERSON M S.A diffusion theory model of spatially resolved steady state diffuse reflectance for the noninvasive determination of tissue optical properties in-vivo[J].Med Phys,1992,19:879-888.
[6]MARTELLI F,SASSAROLI A,BIANCO S D,et al.Solution of the time-dependent diffusion equation for layered diffusive media by the eigenfunction method[J].Physical Review E,2003,67(5):1-14.
[7]LI C-X.Artificial neural network method for determining optical properties from double-integrating-spheres measurements[J].C Opt Letters,2010,10:174-176.
[8]FARRELL T J,WILSON B C,PATTERSON M S.The use of a neural network to determine tissue optical properties from spatially resolved diffuse reflectance measurements[J].Physics in Medicine and Biology,1992,37(12):2281-2286.
[9]CORTES C,VAPNIK V N.Support vector networks[J].Machine Learning,1995,20(3):273-297.
[10]OSUNAS E,FREUND R,GUOSI F.Training Support Vector Machines:An Application to Face Detection[C]//Proceedings of IEEE Conference on CVPRE.Puerto Ricl,1997:130-136.
[11]PLATT J.Fast Training of Support Vector Machines using Sequential Minimal Optimization[C]//Advances in Kernel Methods.MIT Press,1999:185-208.
[12]MACIEJ ZIEBA,JAKUB M TOMCZAK,MAREK LUBICZ.Boosted SVM for extracting rules from imbalanced data in application to prediction of the post-operative life expectancy in the lung cancer patients[J].Appl Soft Comput J,2013,7:99-108.
[13]Support Vector Machines Software[EB/OL].[2014-04-10].http://www.csie.ntu.edu.tw/:cjlin/libsvm.
[14]Monte Carlo Simulation Software[EB/OL].[2014-04-10].http://oilab.tamu.edu/mc.html(accessed Feb 5,2005).
[15]XU HEPING,PATTERSON MICHAEL S.Determination of the optical properties of tissue-simulating phantoms from interstitial frequency domain measurements of relative fluence and phase difference[J].Optics Express,2006,14(14):6485-6501.