[1]吴常铖,曹青青,费飞,等.基于数据手套和神经网络的数字手势识别方法[J].东南大学学报(自然科学版),2020,50(3):563-569.[doi:10.3969/j.issn.1001-0505.2020.03.020]
 Wu Changcheng,Cao Qingqing,Fei Fei,et al.Digital gesture recognition method based on data glove and neural networks[J].Journal of Southeast University (Natural Science Edition),2020,50(3):563-569.[doi:10.3969/j.issn.1001-0505.2020.03.020]
点击复制

基于数据手套和神经网络的数字手势识别方法()
分享到:

《东南大学学报(自然科学版)》[ISSN:1001-0505/CN:32-1178/N]

卷:
50
期数:
2020年第3期
页码:
563-569
栏目:
出版日期:
2020-05-20

文章信息/Info

Title:
Digital gesture recognition method based on data glove and neural networks
作者:
吴常铖12曹青青3费飞1杨德华1陆熊1徐宝国2曾洪2宋爱国2
1南京航空航天大学自动化学院, 南京 211106; 2东南大学仪器科学与工程学院, 南京 210096; 3南京工业职业技术学院航空工程学院, 南京 210023
Author(s):
Wu Changcheng12 Cao Qingqing3 Fei Fei1 Yang Dehua1 Lu Xiong1 Xu Baoguo2Zeng Hong2 Song Aiguo2
1College of Automation Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 211106, China
2School of Instrument Science and Engineering, Southeast University, Nanjing 210096, China
3School of Aviation Engineering, Nanjing Institute of Industry Technology, Nanjing 210023, China
关键词:
数据手套 自学习 神经网络 数字手势识别
Keywords:
data glove self-learning neural networks digital gesture recognition
分类号:
TP911.7;TP241
DOI:
10.3969/j.issn.1001-0505.2020.03.020
摘要:
针对使用数据手套进行数字手势识别时存在个体差异的问题,使用弯曲电阻片设计了数据手套并提出了基于神经网络的数字手势识别方法.首先,在分析测量电路原理的基础上结合弯曲电阻片的特性优选了电路参数,使手指弯曲角度测量的灵敏度最大化.其次,针对用户在手指长度、手势习惯上存在个体差异的情况,提出了一种基于弯曲信号自学习和广义回归神经网络(GRNN)的数字手势识别方法.数据手套信号测试及数字手势试验结果表明,采用优选的电路参数时测量电路的输出振幅最大;在全体评估试验和个体交叉评估试验中,经过自学习预处理后的数字手势识别平均准确率分别为99.2%和96.1%,与未进行自学习处理的识别结果相比分别提高了2.8%和10.7%.在全体评估试验和个体交叉评估试验中,GRNN的识别结果均优于决策树的识别结果.
Abstract:
Aiming at the problem of individual differences in digital gesture recognition based on data gloves, a data glove was designed by using bending resistors and a digital gesture recognition method based on neural networks was proposed. First, a circuit parameter was optimized based on the characteristics of the bending resistor and the principle of the measuring circuit, so as to maximize the sensitivity of the finger bending angle measurement. Secondly, according to the individual differences on the finger length and gesture habits of different users, a gesture recognition method based on signal self-learning and generalized regression neural network(GRNN)was proposed. The results of the data glove testing and the digital gesture recognition experiments show that the output amplitude of the measuring circuit is the largest when the optimized circuit parameter is adopted. In the whole evaluation and the individual cross evaluation experiments, the average accuracy of the digital gesture recognition based on self-learning and GRNN is 99.2% and 96.1%, respectively, which is 2.8% and 10.7% higher than the recognition results without self-learning. The recognition results of GRNN are better than those of the decision tree in the whole evaluation and the individual cross evaluation experiments.

参考文献/References:

[1] 秦超龙,宋爱国,吴常铖,等.基于Unity3D与Kinect的康复训练机器人情景交互系统[J].仪器仪表学报,2017,38(3):530-536.DOI:10.19650/j.cnki.cjsi.2017.03.003.
Qin C L,Song A G,Wu C C,et al.Scenario interaction system of rehabilitation training robot based on Unity3D and Kinect[J].Chinese Journal of Scientific Instrument,2017,38(3):530-536.DOI:10.19650/j.cnki.cjsi.2017.03.003. (in Chinese)
[2] Li I,Liu Y X,Zang X Z.Human-machine interaction control system of a remote maintenance manipulator for tokamak[C]//2018 Chinese Automation Congress(CAC).Xi’an,China,2018:336-340.DOI:10.1109/cac.2018.8623475.
[3] Song A G,Wu C C,Ni D J,et al.One-therapist to three-patient telerehabilitation robot system for the upper limb after stroke[J].International Journal of Social Robotics,2016,8(2):319-329.DOI:10.1007/s12369-016-0343-1.
[4] Shin A,Oh J,Lee J.Apprentice of oz:Human in the loop system for conversational robot wizard of oz[C]//2019 14th ACM/IEEE International Conference on Human-Robot Interaction(HRI).Daegu,South Korea, 2019:516-517.DOI:10.1109/hri.2019.8673205.
[5] Baribina N,Oks A,Baltina I,et al.Development of pressure sensitive glove prototype[J].Key Engineering Materials,2019,800:326-330.DOI:10.4028/www.scientific.net/kem.800.326.
[6] Jha C K,Chakraborty A L.A fiber Bragg grating strain sensor-based glove to accurately measure the bend angle of the finger flexed at the proximal interphalangeal joints[C]//2018 IEEE Sensors.New Delhi,India,2018:1-4.DOI:10.1109/icsens.2018.8589898.
[7] Fang B,Sun F C,Liu H P,et al.A novel data glove for fingers motion capture using inertial and magnetic measurement units[C]//2016 IEEE International Conference on Robotics and Biomimetics(ROBIO).Qingdao,China,2016:2099-2104.DOI:10.1109/robio.2016.7866639.
[8] Connolly J,Condell J,O′Flynn B,et al.IMU sensor-based electronic goniometric glove(iSEG-Glove)for clinical finger movement analysis[J].IEEE Sensors Journal,2018,18(3):1273-1281.DOI:10.1109/jsen.2017.2776262.
[9] 赵裕沛,费飞,杨德华,等.面向手部康复应用的数据手套技术研究[J].仪器仪表学报,2017,38(10):2415-2421.DOI:10.3969/j.issn.0254-3087.2017.10.008.
Zhao Y P,Fei F,Yang D H,et al.Research on data glove for hand rehabilitation[J].Chinese Journal of Scientific Instrument,2017,38(10):2415-2421.DOI:10.3969/j.issn.0254-3087.2017.10.008. (in Chinese)
[10] 刘续丹,侯红霞,杨钰淇,等.基于数据手套的交互系统设计[J].通讯世界,2019,26(4):313-314.DOI:10.3969/j.issn.1006-4222.2019.04.202.
Liu X D,Hou H X,Yang Y Q,et al.Interactive system design based on data glove[J].Telecom World,2019,26(4):313-314.DOI:10.3969/j.issn.1006-4222.2019.04.202. (in Chinese)
[11] 林焕辉,陈德为.基于力反馈数据手套控制的仿生机械手设计研究[J].机电技术,2018,41(1):7-9.DOI:10.19508/j.cnki.1672-4801.2018.01.003.
Lin H H,Chen D W.Design and research of bionic manipulator based on force feedback data glove control[J].Mechanical & Electrical Technology,2018,41(1):7-9.DOI:10.19508/j.cnki.1672-4801.2018.01.003. (in Chinese)
[12] 孙中圣,包钢,李军,等.力反馈数据手套的角度测量[J].哈尔滨工业大学学报,2008,40(1):69-72.DOI:10.3321/j.issn:0367-6234.2008.01.019.
Sun Z S,Bao G,Li J,et al.Angle measurement of force feedback dataglove[J].Journal of Harbin Institute of Technology,2008,40(1):69-72.DOI:10.3321/j.issn:0367-6234.2008.01.019. (in Chinese)
[13] Jiang L J,Xia H L,Guo C L.A model-based system for real-time articulated hand tracking using a simple data glove and a depth camera[J].Sensors,2019,19(21):4680.DOI:10.3390/s19214680.
[14] Huang X A,Wang Q,Zang S Y,et al.Tracing the motion of finger joints for gesture recognition via sewing RGO-coated fibers onto a textile glove[J].IEEE Sensors Journal,2019,19(20):9504-9511.DOI:10.1109/jsen.2019.2924797.
[15] Spectra Symbol.Resistive flex sensors[EB/OL].(2016-11-30)[2019-07-25].https://www.spectrasymbol.com/product/flex-sensors/.
[16] 冯志鹏,宋希庚,薛冬新,等.基于广义回归神经网络的时间序列预测研究[J].振动、测试与诊断,2003,23(2):105-109.
  Feng Z P,Song X G,Xue D X,et al.General regression neural network based prediction of time series[J].Journal of Vibration,Measurement & Diagnosis,2003,23(2):105-109.(in Chinese)

相似文献/References:

[1]颜廷虎,毛玉良,钟秉林.基于联想记忆模型的模糊关系矩阵自学习机制[J].东南大学学报(自然科学版),1997,27(5):1.[doi:10.3969/j.issn.1001-0505.1997.05.001]
 Yan Tinghu,Mao Yuliang,Zhong Binglin.A Self Learning Mechanism for Fuzzy Relation Matrix Based on Association Memory Model[J].Journal of Southeast University (Natural Science Edition),1997,27(3):1.[doi:10.3969/j.issn.1001-0505.1997.05.001]
[2]吕剑虹,张世军,陈来九.自学习多变量汽温控制系统[J].东南大学学报(自然科学版),1995,25(5):33.[doi:10.3969/j.issn.1001-0505.1995.05.006]
 Lu Jianhong,Zhang,Shijun,et al.A Study on Self-Learning Multivariable Temperature Control System[J].Journal of Southeast University (Natural Science Edition),1995,25(3):33.[doi:10.3969/j.issn.1001-0505.1995.05.006]

备注/Memo

备注/Memo:
收稿日期: 2019-12-30.
作者简介: 吴常铖(1987—),男,博士,讲师,changchengwu@nuaa.edu.cn.
基金项目: 国家自然科学基金资助项目(61773205, 61803201)、中央高校基本科研业务费资助项目(NS2018023)、中国博士后科学基金资助项目(2019M661686)、江苏省自然科学基金资助项目(BK20170803).
引用本文: 吴常铖,曹青青,费飞,等.基于数据手套和神经网络的数字手势识别方法[J].东南大学学报(自然科学版),2020,50(3):563-569. DOI:10.3969/j.issn.1001-0505.2020.03.020.
更新日期/Last Update: 2020-05-20