结合颅骨形态特征与神经网络的民族判别

孙慧杰; 赵俊莉; 郑鑫; 热孜万古丽·夏米西丁; 李奕; 周明全

doi:10.13700/j.bh.1001-5965.2020.0446

结合颅骨形态特征与神经网络的民族判别

doi: 10.13700/j.bh.1001-5965.2020.0446

1.
青岛大学数据科学与软件工程学院, 青岛 266071
2.
新疆师范大学计算机科学技术学院, 乌鲁木齐 830054
3.
教育部虚拟现实应用工程研究中心, 北京 100875

基金项目:

国家自然科学基金 61702293

全国统计科学研究项目 2020355

教育部虚拟现实应用工程研究中心基金项目 MEOBNUEVRA201601

中国博士后科学基金 2017M622137

山东省重点研发计划重大科技创新工程 2019JZZY020101

详细信息

作者简介:
孙慧杰女，硕士研究生。主要研究方向：计算机图形学

赵俊莉女，博士, 副教授, 硕士生导师。主要研究方向：计算机图形学、计算机视觉、虚拟现实

通讯作者:
赵俊莉，E-mail: zhaojl@yeah.net

中图分类号: TP183
计量
- 文章访问数: 502
- HTML全文浏览量: 76
- PDF下载量: 186
- 被引次数: 0
出版历程
- 收稿日期: 2020-08-24
- 录用日期: 2020-09-05
- 网络出版日期: 2021-03-20

Ethnic identification by combining features of skull morphology with neural network

1.
School of Data Science and Software Engineering, Qingdao University, Qingdao 266071, China
2.
School of Computer Science and Technology, Xinjiang Normal University, Urumqi 830054, China
3.
Virtual Reality Research Center of Ministry of Education, Beijing 100875, China

Funds:

National Natural Science Foundation of China 61702293

National Statistical Science Research Project 2020355

the Open Research Fund of the Ministry of Education for Engineering Research Center of Virtual Reality Application MEOBNUEVRA201601

China Postdoctoral Science Foundation 2017M622137

Key Research and Development Plan-Major Scientific and Technological Innovation Projects of ShanDong Province 2019JZZY020101

More Information

Corresponding author: ZHAO Junli, E-mail: zhaojl@yeah.net

摘要

摘要:
针对颅骨民族判别问题，提出结合颅骨形态特征与神经网络的判别方法，可以推进法医人类学的发展，加快探索民族发展历程。首先，根据颅骨形态学相关研究，提取36个维吾尔族和汉族颅骨数据的几何特征；其次，采用反向传播神经网络（BPNN）对特征向量进行民族判别，并通过Adam算法对网络进行优化，避免陷入局部最优值，添加正则化项保证算法稳定性；最后，分别采用2种网络结构进行对比实验，输入层、隐藏层和输出层的神经元个数分别为36、6、2和36、12、2，并设置不同初始学习率进行对比实验。结果表明：隐藏层神经元个数为12、学习率为0.000 1时，分类精度最高，测试阶段平均准确率最高为97.5%。为了验证所提方法的普适性，生成116例国外颅骨数据进行实验，测试阶段平均准确率为90.96%。相比较于支持向量机（SVM）、决策树、KNN、Fisher等机器学习方法，所提方法学习能力更强且分类精度有明显提升。
- 颅骨形态特征 /
- 反向传播神经网络(BPNN) /
- 颅骨民族判别 /
- 机器学习 /
- Adam算法
Abstract:
China is a multi-ethnic country. It is of great significance for the skull identification to realize the skull ethnic identification through computers, which can promote the development of forensic anthropology and exploration of national development. Firstly, according to the skull morphology studies, 36 Uighur and Han geometric features of the skull data are extracted, and the Back-Propagation Neural Network (BPNN) of feature vectors is used for ethnic identification. In order to optimize the network, Adam algorithm is adopted to avoid falling into local minimum, and to ensure the stability of the algorithm with regularization terms. Two network structures are used for comparative experiments. The number of neurons in the input layer, hidden layer and output layer are 36, 6, 2 and 36, 12, 2, respectively, and different initial learning rates are set for comparative experiments. The results show that, when the number of hidden-layer neurons is 12 and the learning rate is 0.000 1, the classification accuracy is the highest and the highest accuracy rate in the test stage is 97.5%. In order to verify the universality of the method in this paper, 116 foreign skull data are generated for experiments, and the accuracy rate of the test stage is 90.96%. Compared with machine learning methods such as Support Vector Machine (SVM), decision-making tree, KNN, and Fisher, the proposed method has stronger learning ability and significantly improved classification accuracy.
- features of skull morphology /
- Back-Propagation Neural Network (BPNN) /
- skull ethnic identification /
- machine learning /
- Adam algorithm

HTML全文

图 1 颅骨民族鉴定流程

Figure 1. Flowchart of skull ethnic identification

下载: 全尺寸图片幻灯片

图 2 颅骨头部轮廓、鼻子及眼睛环状边界信息

Figure 2. Skull contour, nose and eye ring boundary information

下载: 全尺寸图片幻灯片

图 3 颅骨头部轮廓边界特征示意图

Figure 3. Schematic diagram of skull contour boundary feature

下载: 全尺寸图片幻灯片

图 4 BPNN网络结构

Figure 4. BPNN network structure

下载: 全尺寸图片幻灯片

图 5 网络结构[36∶6∶2]下不同学习率的训练损失变化

Figure 5. Training loss change of network structure [36∶6∶2] under different learning rates

下载: 全尺寸图片幻灯片

图 6 网络结构[36∶12∶2]下不同学习率的训练损失变化

Figure 6. Training loss change of network structure [36∶12∶2] under different learning rates

下载: 全尺寸图片幻灯片

图 7 学习率为0.000 1时2种网络结构的训练平均准确率

Figure 7. Training accuracy rate of two network structures when learning rate is 0.000 1

下载: 全尺寸图片幻灯片

图 8 不同分类方法的ROC曲线

Figure 8. ROC curves of different classification methods

下载: 全尺寸图片幻灯片

图 9 汉族、维吾尔族和国外颅骨分类的混淆矩阵

Figure 9. Confusion matrix of Han, Uighur and foreign skull ethnic identification

下载: 全尺寸图片幻灯片

表 1 维吾尔族与汉族面部及头部主要形态差异

Table 1. Main morphological differences of face and head between Uygur and Han

主要特征	维吾尔族	汉族
眼睛	眼裂开度中等、内外平行、眼为楔形结构	眼裂开度中等、倾斜程度水平
鼻子	鼻根中等偏高、鼻尖大多向前、鼻长明显长于其他民族、属狭鼻型	鼻根中等、鼻基部水平、属中鼻型
肤色	肤色呈黄、白或棕褐	黄色居多
头型	特圆头型和超圆头型	特圆头型和圆头型

下载: 导出CSV

表 2 不同网络结构及参数下的测试平均准确率

Table 2. Test accuracy rate under different network structures and parameters

网络结构	学习率	迭代次数/10⁴	平均准确率/%
	0.01	10	95
[36∶6∶2]	0.001	10	94.16
	0.000 1	10	92.5
	0.01	10	93.3
[36∶12∶2]	0.001	10	94
	0.000 1	10	97.5

下载: 导出CSV

表 3 不同分类方法的平均准确率

Table 3. Identification rate of different classification methods

方法	平均准确率/%
SVM	77.5
KNN	84.1
Fisher	57.5
决策树	91.6
本文	97.5

下载: 导出CSV

参考文献(41)

[1]	张振标. 藏族的体质特征[J]. 人类学学报, 1985, 4(3): 250-258. https://www.cnki.com.cn/Article/CJFDTOTAL-RLXB198503006.htm ZHANG Z B. The physical characters of Zang(Tibetan) nationality[J]. Acta Anthropologica Sinica, 1985, 4(3): 250-258(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-RLXB198503006.htm
[2]	张振标, 张建军. 广西壮族体质特征[J]. 人类学学报, 1983, 2(3): 260-271. https://www.cnki.com.cn/Article/CJFDTOTAL-RLXB198303006.htm ZHANG Z B, ZHANG J J. Physical characters of Zhuang nationality in Guangxi[J]. Acta Anthropologica Sinica, 1983, 2(3): 260-271(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-RLXB198303006.htm
[3]	张振标, 张建军. 海南岛黎族体质特征之研究[J]. 人类学学报, 1982, 1(1): 53-71. https://www.cnki.com.cn/Article/CJFDTOTAL-ZSDS198603015.htm ZHANG Z B, ZHANG J J. Anthropological studies on Li nationality in Hainan island[J]. Acta Anthropologica Sinica, 1982, 1(1): 53-71(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-ZSDS198603015.htm
[4]	艾琼华, 肖辉, 赵建新, 等. 维吾尔族的体质特征研究[J]. 人类学学报, 1993, 12(4): 63-71. https://www.cnki.com.cn/Article/CJFDTOTAL-RLXB199304006.htm AI Q H, XIAO H, ZHAO J X, et al. A survey on pyhsical characteristics of Uigur nationality[J]. Acta Anthropologica Sinica, 1993, 12(4): 63-71(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-RLXB199304006.htm
[5]	郑连斌, 朱钦, 王巧玲, 等. 宁夏回族体质特征研究[J]. 人类学学报, 1997, 16(1): 11-21. https://www.cnki.com.cn/Article/CJFDTOTAL-RLXB701.001.htm ZHENG L B, ZHU Q, WANG Q L, et al. The physical characteristic of Hui nationality in Ningxia[J]. Acta Anthropologica Sinica, 1997, 16(1): 11-21(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-RLXB701.001.htm
[6]	张兴华, 郑连斌, 宇克莉, 等. 山东寿光汉族体质特征[J]. 人类学学报, 2011, 30(2): 95-106. https://www.cnki.com.cn/Article/CJFDTOTAL-RLXB201102008.htm ZHANG X H, ZHENG L B, YU K L, et al. A study of the physical characteristics of the Han people from Shouguang in Shangdong[J]. Acta Anthropologica Sinica, 2011, 30(2): 95-106(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-RLXB201102008.htm
[7]	李咏兰, 陆舜华, 郑连斌, 等. 江西汉族体质特征[J]. 解剖学报, 2012, 43(1): 114-122. doi: 10.3969/j.issn.0529-1356.2012.01.022 LI Y L, LU S H, ZHENG L B, et al. Physical characteristics of the Han people in Jiangxi[J]. Acta Anatomica Sinica, 2012, 43(1): 114-122(in Chinese). doi: 10.3969/j.issn.0529-1356.2012.01.022
[8]	郑连斌, 武亚文, 张兴华, 等. 四川汉族体质特征[J]. 解剖学报, 2011, 42(5): 695-702. doi: 10.3969/j.issn.0529-1356.2011.05.025 ZHENG L B, WU Y W, ZHANG X H, et al. Physical characteristics of Han in Sichuan[J]. Acta Anatomica Sinica, 2011, 42(5): 695-702(in Chinese). doi: 10.3969/j.issn.0529-1356.2011.05.025
[9]	李咏兰, 陆舜华, 郑连斌, 等. 浙江汉族的体质特征[J]. 解剖学报, 2013, 44(5): 707-716. doi: 10.3969/j.issn.0529-1356.2013.05.024 LI Y L, LU S H, ZHENG L B, et al. Physical characteristics of Zhejiang Han[J]. Acta Anatomica Sinica, 2013, 44(5): 707-716(in Chinese). doi: 10.3969/j.issn.0529-1356.2013.05.024
[10]	肖明. 民族形象与解剖结构-西北地区几个少数民族外貌特征的探索[J]. 西北民族大学学报(哲学社会科学版), 1980(1): 86-93. https://www.cnki.com.cn/Article/CJFDTOTAL-XBMZ198001010.htm XIAO M. National image and anatomical structure-A study on the physical features of several ethnic minorities in Northwest China[J]. Journal of Northwest University for Nationalities(Philosophy and Social Science), 1980(1): 86-93(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-XBMZ198001010.htm
[11]	李志洁, 段晓东, 王存睿. 中国6个民族的面部几何特征聚类分析[J]. 大连民族学院学报, 2015, 17(1): 73-76. doi: 10.3969/j.issn.1009-315X.2015.01.019 LI Z J, DUAN X D, WANG C R. Cluster analysis of facial geometric features for six Chinese nationalitie[J]. Journal of Dalian Nationalities University, 2015, 17(1): 73-76(in Chinese). doi: 10.3969/j.issn.1009-315X.2015.01.019
[12]	周文莲, 吴新智. 现代人头骨面部几项非测量性状的观察[J]. 人类学学报, 2001, 20(4): 42-48. https://www.cnki.com.cn/Article/CJFDTOTAL-RLXB200104004.htm ZHOU W L, WU X Z. Observations of some non-metrical Traits in the mordern human skulls[J]. Acta Anthropologica Sinica, 2001, 20(4): 42-48(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-RLXB200104004.htm
[13]	WEINBERG S M, PUTZ D A, MOONEY M P, et al. Evaluation of non-metric variation in the crania of black and white perinates[J]. Forensic Science International, 2005, 151(2-3): 177-185. doi: 10.1016/j.forsciint.2005.02.009
[14]	HOLLAND T D. Race determination of fragmentary crania by analysis of the cranial base[J]. Journal of Forensic Science, 1986, 31(2): 719-725. doi: 10.1520/jfs12305j
[15]	I·SCAN M Y, STEYN M. Craniometric determination of population affinity in South Africans[J]. International Journal of Legal Medicine, 1999, 112(2): 91-97. doi: 10.1007/s004140050208
[16]	HOLIDAY T W, FALSETTI A B. A new method for discriminating African-American from European-American skeletons using postcranial osteometrics reflective of body shape[J]. Journal of Forensic Science, 1999, 44(5): 926-930. http://www.ncbi.nlm.nih.gov/pubmed/10486943/
[17]	GILL G W, HUGHES S S, BENNETT S M, et al. Racial identification from the midfacial skeleton with special reference to American Indians and whites[J]. Journal of Forensic Science, 1988, 33(1): 92-99. http://www.ncbi.nlm.nih.gov/pubmed/3351476
[18]	刘玉勇. 华北地区汉族成人面颅骨X线片性别判定的研究[J]. 中国司法鉴定, 2016, 84(1): 26-31. doi: 10.3969/j.issn.1671-2072.2016.01.004 LIU Y Y. The sex determination of Han nationality in North China by adult facial skull X-ray[J]. Chinese Journal of Forensic Sciences, 2016, 84(1): 26-31(in Chinese). doi: 10.3969/j.issn.1671-2072.2016.01.004
[19]	KHAITAN T, KABIRAJ A, GINJUPALLY U, et al. Cephalometric analysis for gender determination using maxillary sinus index: A novel dimension in personal identification[J]. International Journal of Dentistry, 2017: 1-4. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC5360944/
[20]	FRANKLIN D, FREEDMAN L, MILNE N. Sexual dimorphism and discriminant function sexing in indigenous South African crania[J]. Homo-Journal of Comparative Human Biology, 2005, 55(3): 213-228. doi: 10.1016/j.jchb.2004.08.001
[21]	税午阳, 殷荣超, 周明全, 等. 中国汉族人颅骨数字模型的性别判别方法[J]. 中国法医学杂志, 2013, 28(6): 19-21. https://www.cnki.com.cn/Article/CJFDTOTAL-FUAN201306004.htm SHUI W Y, YIN R C, ZHOU M Q, et al. Sex determination from digital skull model for the Han people in China[J]. Chinese Journal of Forensic Medicine, 2013, 28(6): 19-21(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-FUAN201306004.htm
[22]	AMORES-AMPUERO A. Sexual dimorphism in base of skull[J]. Anthropologischer Anzeiger, 2017, 74(1): 9-14. doi: 10.1127/anthranz/2017/0603
[23]	SMALL C, SCHEPARTZ L, HEMINGWAY J, et al. Three-dimensionally derived interlandmark distances for sex estimation in intact and fragmentary crania[J]. Forensic Science International, 2018, 287(2): 127-135.
[24]	MAASS P, FRIEDING L J. Morphometric analysis of the neurocranium in an adult South African cadaveric sample[J]. Journal of Forensic Sciences, 2019, 64(2): 367-374. doi: 10.1111/1556-4029.13878
[25]	LUO L, WANG M Y, DUAN F Q, et al. Automatic sex determination of skulls based on a statistical shape model[J]. Computational and Mathematical Method in Medicine, 2013: 1-6.
[26]	杨稳, 刘晓宁, 朱菲, 等. 基于统计形状模型的颅骨自动性别识别[J]. 计算机科学, 2019, 46(6): 282-287. https://www.cnki.com.cn/Article/CJFDTOTAL-JSJA201906042.htm YANG W, LIU X N, ZHU F, et al. Automatic sex determination of skulls based on statistical shape model[J]. Computer Science, 2019, 46(6): 282-287(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-JSJA201906042.htm
[27]	LUO L, CHANG L, LIU R, et al. Morphological investigations of skulls for sex determination based on sparse principal component analysis[C]//Proceedings of the Chinese Conference on Biometric Recognition. Berlin: Springer, 2013: 449-456.
[28]	杨稳, 刘晓宁, 刘雄乐, 等. 结合改进卷积神经网络和最小二乘法的颅骨性别鉴定[J]. 人类学学报, 2019, 38(2): 265-275. https://www.cnki.com.cn/Article/CJFDTOTAL-RLXB201902012.htm YANG W, LIU X N, LIU X L, et al. Skull sex identification using improved convolution neural network and least squares method[J]. Acta Anthropologica Sinica, 2019, 38(2): 265-275(in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-RLXB201902012.htm
[29]	YANG W, LIU X N, WANG K G, et al. Sex determination of three-dimensional skull based on improved back propagation neural network[J]. Computational and Mathematical Methods in Medicine, 2019: 1-8.
[30]	HEFNER J T, SPRADLEY M K, ANDERSON B. Ancestry assessment using random forest modeling[J]. Journal of Forensic Sciences, 2014, 59(3): 583-589. doi: 10.1111/1556-4029.12402
[31]	OAKLEY C, BAI L, LIAO I Y, et al. A novel method for race determination of human skulls[C]//International Conference on Pattern Recognition. Berlin: Springer, 2018: 89-102.
[32]	DUAN F, YANG Y, LI Y, et al. Skull identification via correlation measure between skull and face shape[J]. IEEE Transactions on Information Forensics and Security, 2014, 9(8): 1322-1332. doi: 10.1109/TIFS.2014.2332981
[33]	LORENSEN W E, CLINE H E. Marching cubes: A high resolution 3D surface construction algorithm[J]. ACM SIGGRAPH Computer Graphics, 1987, 21(4): 163-169. doi: 10.1145/37402.37422
[34]	HU Y, DUAN F, YIN B, et al. A hierarchical dense deformable model for 3D face reconstruction from skull[J]. Multimedia Tools and Applications, 2013, 64(2): 345-364. doi: 10.1007/s11042-012-1005-4
[35]	HUANG R, ZHAO J, DUAN F, et al. Automatic craniofacial registration based on radial curves[J]. Computers & Graphics, 2019, 82(5): 264-274. http://www.sciencedirect.com/science/article/pii/s0097849319300883
[36]	RUMELHART D E, HINTON G E, WILLIAMS R J. Learning representations by back-propagating errors[J]. Nature, 1986, 323(6088): 533-536. doi: 10.1038/323533a0
[37]	KINGMA D, BA J. Adam: A method for stochastic optimization[C]//Proceedings of the 3rd International Conference for Learning Representations, 2015: 1-15.
[38]	BERAR M, DESVIGNES M, BAILLY G, et al. 3D semi-landmarks based statistical face reconstruction[J]. Journal of Computing & Information Technology, 2006, 14(1): 31-43. http://www.zhangqiaokeyan.com/open-access_resources_thesis/0100014454929.html
[39]	SHUI W Y, ZHOU M Q, MADDOCK S, et al. A PCA-based method for determining craniofacial relationship and sexual dimorphism of facial shapes[J]. Computers in Biology and Medicine, 2017, 90(8): 33-49.
[40]	王琳, 赵俊莉, 段福庆, 等. 颅面复原方法综述[J]. 计算机工程, 2019, 45(12): 8-18. https://www.cnki.com.cn/Article/CJFDTOTAL-JSJC201912002.htm WANG L, ZHAO J L, DUAN F Q, et al. Survey on craniofacialreconstruction method[J]. Computer Engineering, 2019, 45(12): 8-18(in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-JSJC201912002.htm
[41]	GUPTA S, CASTLEMAN K R, MARKEY M K, et al. Texas 3D face recognition database[C]//Proceedings of the Image Analysis & Interpretation. Piscataway: IEEE Press, 2010: 97-100.