前交叉韧带断裂后足底压力特征的聚类分析

李晓理; 黄红拾; 王杰; 于媛媛; 敖英芳

doi:10.16383/j.aas.2017.c160197

前交叉韧带断裂后足底压力特征的聚类分析

doi: 10.16383/j.aas.2017.c160197

李晓理^1,,
黄红拾^2,,
王杰^3, ,,
于媛媛^2,,
敖英芳^2, ,

1.
北京工业大学电子信息与控制工程学院北京 100124
2.
北京大学第三医院运动医学研究所北京 100191
3.
北京科技大学自动化学院北京 100083

基金项目:

国家自然科学基金 61473034

低温重点实验室开放基金 CRYO 201316

国家自然科学基金 61673053

北京大学医-信交叉建设孵化基金 BMU2016-12

北京市科技新星交叉学科项目 Z161100004916041

高等学校博士学科点专项科研基金 2013000611008

低温重点实验室开放基金 TIPC

低温重点实验室开放基金 CAS

详细信息

作者简介:
李晓理北京工业大学电子信息与控制工程学院教授.主要研究方向为多模型控制和自适应控制.E-mail:lixiaolibjut@bjut.edu.cn

黄红拾北京大学第三医院运动医学研究所副主任医师.主要研究方向为运动医学/康复, 运动能力的训练和提高, 运动损伤防护的中西医结合康复技术, 运动生物力学、康复技术和辅具功能研发.本文共同第一作者.E-mail:huanghs@bjmu.edu.cn

于媛媛北京大学第三医院运动医学研究所康复治疗师.主要研究方向为运动康复, 运动能力的训练和提高, 运动生物力学测试、康复技术和辅具功能研发.E-mail:yyy1091012@126.com

通讯作者:
王杰北京科技大学自动化学院硕士研究生.主要研究方向为模糊控制与聚类算法.本文共同通信作者.E-mail:yizhe9288@163.com

敖英芳北京大学运动医学研究所教授、主任医师.主要研究方向为运动创伤学, 为膝关节运动损伤、交叉韧带与软骨损伤的修复与重建.本文通信作者.E-mail:yingfang.ao@vip.sina.com

计量
- 文章访问数: 2454
- HTML全文浏览量: 209
- PDF下载量: 704
- 被引次数: 0
出版历程
- 收稿日期: 2016-02-29
- 录用日期: 2016-08-15
- 刊出日期: 2017-03-20

Cluster Analysis of Plantar Pressure Characteristics after Anterior Cruciate Ligament Deficiency

LI Xiao-Li^1
,,
HUANG Hong-Shi^2
,,
WANG Jie^{3
, ,},
YU Yuan-Yuan^2
,,
AO Ying-Fang^{2
, ,}

1.
College of Electronic Information and Control Engineering, Beijing University of Technology, Beijing 100124
2.
Institute of Sports Medicine, Peking University Third Hospital, Beijing 100191
3.
School of Automation and Electronic Engineering, University of Science and Technology Beijing, Beijing 100083

Funds:

National Natural Science Foundation of China 61473034

Opening Foundation of Key Laboratory of Cryogenics CRYO 201316

National Natural Science Foundation of China 61673053

the Incubation Grant for Medicine and Information Sciences of Peking University BMU2016-12

Beijing Nova Programme Interdisciplinary Cooperation Project Z161100004916041

Specialized Research Fund for the Doctoral Program of Higher Education 2013000611008

Opening Foundation of Key Laboratory of Cryogenics TIPC

Opening Foundation of Key Laboratory of Cryogenics CAS

More Information

Author Bio:
Professor at the College of Electronic Information and Control Engineering, Beijing University of Technology. His research interest covers multiple model control and adaptive control.

Associate chief physician at Peking University Institute of Sports Medicine. His research interest covers sports medicine/rehabilitation, elite athletes sports training and performance enhancement methods, East-meets-West in rehabilitation techniques in sports injuries and prevention programs, biomechanics of sports and lower extremity injury, isokinetic measurement and training, brace and shoe functions. Co-first author of this paper.

Therapist at Peking University Institute of Sports Medicine. Her research interest covers sports medicine/reh- abilitation, elite athletes sports training and performance enhancement methods, biomechanics of sports and lower extremity injury, isokinetic measurement and training.

Corresponding author: WANG Jie Master student at the School of Automation and Electrical Engineering, University of Science and Technology Beijing. His research interest covers fuzzy control and clustering algorithm. Co-corresponding author of this paper.; AO Ying-Fang Professor, chief physician at Peking University Institute of Sports Medicine. His research interest covers sports injury, knee injury and cartilage repair, clinical studies and basic research on sports injuries, clinical studies on motor function following knee cruciate ligament injuries, repair and reconstruction of knee cruciate ligament. Corresponding author of this paper.

摘要

摘要: 运动过程中，人体的步态特征可以在足底压力图像上有准确的记录，而这也就可以成为判断步态正常与否的一条有效依据.通过一组压力传感器阵列获取人体运动过程的足底压力分布数据，提取步态的运动学和动力学特性.在此基础上，采用极限学习机（Extreme learning machines，ELM）神经网络聚类算法对足底压力数据进行分析，完成正常与异常步态的分类辨识工作.本文从实际临床数据出发，对前交叉韧带断裂患者进行步态分析，并据医生的临床诊断结果进行校验.该方法在步态分析上取得了较为良好的效果，仿真结果表明了其有效性.
- 足底压力 /
- 步态特征 /
- 极限学习机神经网络 /
- 前交叉韧带断裂 /
- 聚类分析
Abstract: The gait characteristics of an actor can be recorded accurately on the plantar pressure map in a movement. It can be used to distinguish whether the gait of this actor in a movement is abnormal or not. Using a set of pressure sensors, the plantar pressure during dynamic motion is collected, and the kinetic and dynamic characteristics of gait are extracted. Then extreme learning machines (ELM) neural network cluster algorithm is used to the analyze of the plantar pressure data and identification of normal or abnormal gait is done. Based on actual clinical data, this method carries out an analysis of patients with anterior cruciate ligament deficiency, which is checked according to the doctor's clinical diagnosis results. Result shows that this method is effective.
- Plantar pressure /
- gait characteristics /
- extreme learning machines (ELM) neural network /
- anterior cruciate ligament deficiency /
- cluster analysis

HTML全文

图 1 FootScan 足底压力测量平板系统

Fig. 1 FootScan plantar pressure measurement

下载: 全尺寸图片幻灯片

图 2 整个着地周期的5个特征时刻

Fig. 2 Five key moments of the whole stance phase

下载: 全尺寸图片幻灯片

图 3 ELM神经网络结构图

Fig. 3 The structure of ELM neural network

下载: 全尺寸图片幻灯片

图 4 足底压力数据图

Fig. 4 Plantar pressure map

下载: 全尺寸图片幻灯片

图 5 正常测试者与ACL断裂测试者的左侧足底压力COP轨迹线

Fig. 5 Left feet's COP lines of normal and ACLD participators

下载: 全尺寸图片幻灯片

图 6 正常测试者与ACL断裂测试者的右侧足底压力COP轨迹线

Fig. 6 Right feet's COP lines of normal and ACLD participators

下载: 全尺寸图片幻灯片

图 7 COP轨迹线放缩效果图

Fig. 7 Zoom of COP line

下载: 全尺寸图片幻灯片

表 1 正常测试者左侧足底压力数据特征序列表

Table 1 COP feature vectors of left feet in the normal group

序号	组别	$C_1x$	$C_1y$	$C_2x$	$C_2y$	$C_3x$	$C_3y$	$C_4x$	$C_4y$	$C_5x$	$C_5y$
1	$A$	2	12	5	12	24	14	26	15	26	14
2		3	11	4	12	17	12	23	12	27	14
3		2	13	3	13	10	13	23	12	27	14
4		3	13	5	13	15	11	23	12	28	13
5		3	13	5	13	14	11	22	11	28	11

6	$B$	3	12	5	13	13	11	22	11	27	11
7		3	13	4	13	13	10	22	10	28	13
8		3	13	6	13	15	8	21	7	27	7
9		3	13	5	13	14	11	22	11	28	11
10		3	12	5	13	13	11	22	11	27	11

11	$C$	3	12	5	13	10	11	21	9	28	12
12		3	10	5	9	12	9	21	10	27	15
13		2	13	9	10	13	10	11	10	26	13
14		2	9	4	9	8	9	21	9	27	15
15		3	9	5	10	6	10	23	13	28	14

16	$D$	3	8	4	9	8	9	22	12	27	15
17		3	8	5	9	13	8	22	10	28	16
18		3	12	8	11	16	9	21	10	27	13
19		2	13	7	12	13	10	22	10	27	13
20		3	13	8	11	19	8	21	9	27	13

21	$E$	3	12	10	10	17	9	21	9	27	12
22		3	14	5	14	13	12	6	14	27	9
23		2	15	5	15	7	15	23	10	27	8
24		3	15	8	13	14	9	10	11	28	7
25		3	14	5	15	16	10	22	9	26	10

下载: 导出CSV

表 2 左膝ACL断裂测试者左侧足底压力数据特征序列表

Table 2 COP feature vectors of left feet in the group of left side ACLD

序号	组别	$C_1x$	$C_1y$	$C_2x$	$C_2y$	$C_3x$	$C_3y$	$C_4x$	$C_4y$	$C_5x$	$C_5y$
1	$A$	2	12	4	12	11	12	20	11	27	14
2		2	11	9	9	15	9	21	9	27	15
3		2	10	6	9	11	9	21	10	28	15
4		3	10	11	7	15	9	21	9	28	15
5		3	11	6	11	14	9	20	9	27	13

6	$B$	3	10	11	7	15	9	21	9	28	15
7		3	10	11	6	13	7	20	8	27	15
8		2	10	7	10	12	9	20	8	27	14
9		4	12	5	13	13	11	20	11	27	13
10		2	11	5	12	13	12	21	12	27	14

11	$C$	3	13	4	14	10	14	23	12	28	13
12		3	11	5	11	12	12	22	13	28	14
13		2	12	4	13	10	12	21	12	27	15
14		2	13	5	13	12	12	22	12	28	14
15		2	13	7	11	11	10	21	10	26	11

16	$D$	2	13	8	10	12	9	20	7	23	9
17		3	13	6	11	14	9	23	9	24	9
18		3	14	8	12	14	10	21	9	24	9
19		1	13	6	12	15	7	21	8	24	10
20		2	9	5	9	15	10	22	12	27	13

21	$E$	3	9	14	6	19	7	21	10	27	12
22		2	10	8	7	14	8	21	11	27	13
23		2	9	8	7	18	8	21	10	27	13
24		3	11	6	11	14	8	20	9	27	13
25		3	10	5	10	15	7	21	10	28	14

下载: 导出CSV

表 3 左侧足底压力数据聚类分析辨识结果

Table 3 The result of left plantar pressure analysis

序号	训练组别	训练数据个数	测试组别	测试数据个数	辨识准确率 (%)
1	$A, B, C, D$	40	$E$	10	90
2	$A, B, C, E$	40	$D$	10	50
3	$A, B, D, E$	40	$C$	10	60
4	$ A, C, D, E$	40	$B$	10	100
5	$B, C, D, E$	40	$A$	10	80

下载: 导出CSV

表 4 正常测试者右侧足底压力数据特征序列表

Table 4 COP feature vectors of right feet in the normal group

序号	组别	$C_1x$	$C_1y$	$C_2x$	$C_2y$	$C_3x$	$C_3y$	$C_4x$	$C_4y$	$C_5x$	$C_5y$
1	$A$	3	8	5	7	14	8	7	7	27	6
2		3	10	4	8	10	8	22	5	27	5
3		2	8	4	8	12	8	21	9	27	7
4		2	5	4	4	11	7	22	9	28	9
5		2	4	4	4	13	8	22	10	27	9

6	$B$	3	5	5	4	5	4	22	10	28	10
7		3	4	4	4	16	9	23	9	28	9
8		3	5	4	5	13	8	22	9	28	8
9		2	4	4	4	13	8	22	10	27	9
10		3	5	5	4	5	4	22	10	28	10

11	$C$	3	5	4	4	12	7	22	9	28	10
12		4	7	5	7	12	8	21	8	23	9
13		2	7	6	6	16	7	22	6	26	3
14		3	9	4	9	14	8	21	6	27	4
15		2	5	5	5	13	6	21	7	27	4

16	$D$	3	6	4	6	12	7	21	8	27	5
17		3	7	8	8	16	8	21	9	27	6
18		2	8	7	10	12	10	22	10	28	7
19		2	7	4	8	15	9	9	9	27	6
20		3	7	9	8	15	8	22	7	28	4

21	$E$	3	9	9	12	16	12	22	9	27	5
22		3	4	5	4	14	6	5	4	27	9
23		3	7	5	7	13	10	23	13	27	14
24		2	5	5	5	10	6	6	5	27	8
25		3	6	4	6	7	6	4	6	27	8

下载: 导出CSV

表 5 右膝ACL断裂测试者右侧足底压力数据特征序列表

Table 5 COP feature vectors of right feet in the group of right side ACLD

序号	组别	$C_1x$	$C_1y$	$C_2x$	$C_2y$	$C_3x$	$C_3y$	$C_4x$	$C_4y$	$C_5x$	$C_5y$
1	$A$	2	4	8	8	17	10	21	10	27	8
2		3	7	4	7	15	9	20	9	27	7
3		2	6	5	6	17	10	21	9	27	8
4		3	6	7	8	18	9	22	9	27	8
5		2	6	7	8	17	10	21	10	27	8

6	$B$	2	6	7	7	13	7	23	8	27	8
7		3	4	5	4	14	7	22	9	27	11
8		2	4	7	6	12	6	23	8	27	9
9		3	5	8	7	14	8	23	8	27	9
10		2	6	5	6	15	7	22	8	27	8

11	$C$	2	4	5	5	12	6	21	9	27	8
12		3	5	5	6	12	8	22	10	27	10
13		2	7	6	7	16	9	14	10	27	6
14		2	8	11	11	16	10	21	9	27	6
15		3	7	5	8	16	9	21	9	27	5

16	$D$	2	6	6	7	13	9	21	9	27	4
17		2	6	7	8	16	9	21	8	27	4
18		3	7	6	8	17	10	22	9	28	6
19		2	6	4	6	13	8	22	9	28	8
20		2	7	4	7	10	7	21	9	27	7

21	$E$	3	9	6	9	12	10	21	10	27	6
22		3	7	6	7	15	9	22	8	27	7
23		3	6	6	6	17	8	23	7	28	6
24		3	6	5	6	16	9	21	10	27	5
25		3	7	8	8	15	10	22	11	28	8

下载: 导出CSV

表 6 右侧足底压力数据聚类分析辨识结果

Table 6 The result of right plantar pressure analysis

序号	训练组别	训练数据个数	测试组别	测试数据个数	辨识准确率 (%)
1	$ A, B, C, D$	40	$E$	10	80
2	$A, B, C, E$	40	$D$	10	70
3	$ A, B, D, E$	40	$C$	10	80
4	$ A, C, D, E$	40	$B$	10	70
5	$ B, C, D, E$	40	$A$	10	80

下载: 导出CSV

参考文献(18)

[1]	Taylor A J, Menz H B, Keenan A M. The influence of walking speed on plantar pressure measurements using the two-step gait initiation protocol. The Foot, 2004, 14 (1): 49-55
[2]	Stolwijk N M, Duysens J, Louwerens J W K, Keijsers N L. Plantar pressure changes after long-distance walking. Medicine and Science in Sports and Exercise, 2010, 42 (12): 2264-2272
[3]	Lai Y C, Lin H S, Pan H F, Chang W N, Hsu C J, Renn J H. Impact of foot progression angle on the distribution of plantar pressure in normal children. Clinical Biomechanics, 2014, 29 (2): 196-200 doi: 10.1016/j.clinbiomech.2013.11.012
[4]	Keijsers N L W, Stolwijk N M, Nienhuis B, Duysens J. A new method to normalize plantar pressure measurements for foot size and foot progression angle. Journal of Biomechanics, 2009, 42 (1): 87-90 http://www.academia.edu/14125878/A_new_method_to_normalize_plantar_pressure_measurements_for_foot_size_and_foot_progression_angle
[5]	Kohonen T. The self-organizing map. Proceedings of the IEEE, 1990, 78 (9): 1464-1480
[6]	Kohonen T. Self-Organizing Maps. New York: Springer-Verlag, 2001.
[7]	Carpenter G A, Grossberg S, Rosen D B. Fuzzy ART: fast stable learning and categorization of analog patterns by an adaptive resonance system. Neural Networks 1991, 4 (06): 759-771 doi: 10.1016/0893-6080(91)90056-B
[8]	Huang G B, Wang D H, Lan Y. Extreme learning machines: a survey. International Journal of Machine Learning and Cybernetics, 2011, 2 (2): 107-122 doi: 10.1007/s13042-011-0019-y
[9]	Stolwijk N M, Duysens J, Louwerens J W, van de Ven Y H, Keijsers N L. Flat feet, happy feet? Comparison of the dynamic plantar pressure distribution and static medial foot geometry between Malawian and Dutch adults. PLoS One, 2013, 8 (2): e57209 doi: 10.1371/journal.pone.0057209
[10]	Cavanagh P R, Rodgers M M. The arch index: a useful measure from footprints. Journal of Biomechanics, 1987, 20 (5): 547-551 doi: 10.1016/0021-9290(87)90255-7
[11]	Andriacchi T P, Dyrby C O. Interactions between kinematics and loading during walking for the normal and ACL deficient knee. Journal of Biomechanics, 2005, 38 (2): 293-298 doi: 10.1016/j.jbiomech.2004.02.010
[12]	黄红拾, 于媛媛, 郭秦炜, 徐雁, 敖英芳.前交叉韧带断裂患者步行时足底压力时间特征.中国运动医学杂志, 2015, 34 (3): 271-274 http://www.cnki.com.cn/Article/CJFDTOTAL-YDYX201503009.htm Huang Hong-Shi, Yu Yuan-Yuan, Guo Qin-Wei, Xu Yan, Ao Ying-Fang. Temporal characteristics of plantar pressure in patients with anterior cruciate ligament rupture during walking. Chinese Journal of Sports Medicine, 2015, 34 (3): 271-274 http://www.cnki.com.cn/Article/CJFDTOTAL-YDYX201503009.htm
[13]	Huang G B, Chen Y Q, Babri H A. Classification ability of single hidden layer feedforward neural networks. IEEE Transactions on Neural Networks, 2000, 11 (3): 799-801 doi: 10.1109/72.846750
[14]	Feng G R, Huang G B, Lin Q P, Gay R. Error minimized extreme learning machine with growth of hidden nodes and incremental learning. IEEE Transactions on Neural Networks, 2009, 20 (8): 1352-1357 https://www.researchgate.net/publication/26665344_Error_Minimized_Extreme_Learning_Machine_With_Growth_of_Hidden_Nodes_and_Incremental_Learning
[15]	Huang G B, Zhou H M, Ding X J, Zhang R. Extreme learning machine for regression and multiclass classification. IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), 2012, 42 (2): 513-529 http://www.doc88.com/p-9197171944428.html
[16]	DeFrate L E, Papannagari R, Gill T J, Moses J M, Pathare N P, Li G A. The 6 degrees of freedom kinematics of the knee after anterior cruciate ligament deficiency: an in vivo imaging analysis. American Journal of Sports Medicine, 2006, 34 (8): 1240-1246 https://www.researchgate.net/publication/7145174_The_6_Degrees_of_Freedom_Kinematics_of_the_Knee_After_Anterior_Cruciate_Ligament_Deficiency_An_In_Vivo_Imaging_Analysis
[17]	Ikeda H, Kurosawa H, Kim S G. Quadriceps torque curve pattern in patients with anterior cruciate ligament injury. International Orthopaedics, 2002, 26 (6): 374-376 doi: 10.1007/s00264-002-0402-0
[18]	Chiu M C, Wu H C, Chang L Y, Wu M H. Center of pressure progression characteristics under the plantar region for elderly adults. Gait and Posture, 2013, 37 (3): 408-412 doi: 10.1016/j.gaitpost.2012.08.010