高级检索

  • ISSN 1006-3080
  • CN 31-1691/TQ

基于卷积神经网络的手势识别算法

朱雯文 叶西宁

朱雯文, 叶西宁. 基于卷积神经网络的手势识别算法[J]. 华东理工大学学报(自然科学版), 2018, (2): 260-269. doi: 10.14135/j.cnki.1006-3080.20170327001
引用本文: 朱雯文, 叶西宁. 基于卷积神经网络的手势识别算法[J]. 华东理工大学学报(自然科学版), 2018, (2): 260-269. doi: 10.14135/j.cnki.1006-3080.20170327001
ZHU Wen-wen, YE Xi-ning. Convolution Neural Networks for Gesture Recognition[J]. Journal of East China University of Science and Technology, 2018, (2): 260-269. doi: 10.14135/j.cnki.1006-3080.20170327001
Citation: ZHU Wen-wen, YE Xi-ning. Convolution Neural Networks for Gesture Recognition[J]. Journal of East China University of Science and Technology, 2018, (2): 260-269. doi: 10.14135/j.cnki.1006-3080.20170327001

基于卷积神经网络的手势识别算法

doi: 10.14135/j.cnki.1006-3080.20170327001
基金项目: 

国家自然科学基金(60974066)

Convolution Neural Networks for Gesture Recognition

  • 摘要: 手势识别是人机交互、智能假肢、医疗康复等领域的研究热点。为了满足手势识别实时性和准确性的需求,本文以成本较小的加速度信号作为数据,在对LeNet-5卷积神经网络进行分析的基础上,提出了一种适合加速度信号的LeNet-A网络。该网络针对基于加速度的手势分类特有的复杂性,增加Dropout层,改变卷积核大小、卷积核数量、激活函数以及分类器。在Ninapro数据集上的实验结果表明,该网络在正常受试者和截肢者的识别率上均表现出很大的优势,平均精度分别为90.37%和79.99%,比目前最佳分类器提升了12%和31%左右。该网络还具有较好的实时性和抗噪性。

     

  • [1] VANBANG, 朱煜, 赵江坤, 等. 基于深度图像HOG特征的实时手势识别方法[J]. 华东理工大学学报(自然科学版), 2015, 41(5):698-702.
    [2] ATZORI M, GIJSBERTS A, MVLLER H, et al. Classification of hand movements in amputated subjects by sEMG and accelerometers[C]//201436th Annual International Conference of the IEEE Engineering in Medicine and Biology Society. UAS:IEEE, 2014:3545-3549.
    [3] GIJSBERTS A, CAPUTO B. Exploiting accelerometers to improve movement classification for prosthetics[C]//2013 IEEE International Conference on Rehabilitation Robotics (ICORR). USA:IEEE, 2013:1-5.
    [4] PENG L, HOU Z, CHEN Y, et al. Combined use of semg and accelerometer in hand motion classification considering forearm rotation[C]//201335th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC). USA:IEEE, 2013:4227-4230.
    [5] GIJSBERTS A, ATZORI M, CASTELLINI C, et al. Movement error rate for evaluation of machine learning methods for sEMG-based hand movement classification[J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2014, 22(4):735-744.
    [6] NIEZEN G, HANCKE G P. Evaluating and optimising accelerometer-based gesture recognition techniques for mobile devices[C]//Proceedings of IEEE AFRICON-09. Kenya:IEEE, 2009:1-6.
    [7] 荆雷, 马文君, 常丹华. 基于动态时间规整的手势加速度信号识别[J]. 传感技术学报, 2012, 25(1):72-76.
    [8] 谢仁强, 曹俊诚. 基于加速度传感器的可扩展手势识别[J]. 传感技术学报, 2016, 29(5):659-664.
    [9] LIN H I, HSU M H, CHEN W K. Human hand gesture recognition using a convolution neural network[C]//2014 IEEE International Conference on Automation Science and Engineering (CASE).USA:IEEE, 2014:1038-1043.
    [10] 蔡娟, 蔡坚勇, 廖晓东, 等. 基于卷积神经网络的手势识别初探[J]. 计算机系统应用, 2015, 24(4):113-117.
    [11] 操小文, 薄华. 基于卷积神经网络的手势识别研究[J]. 微型机与应用, 2016, 35(9):55-57.
    [12] ATZORI M, COGNOLATO M, MVLLER H. Deep learning with convolutional neural networks applied to electromyography data:A resource for the classification of movements for prosthetic hands[J]. Frontiers in Neurorobotics, 2016, 10(9):1-10.
    [13] LECUN Y, BOTTOU L, BENGIO Y, et al. Gradient-based learning applied to document recognition[J]. Proceedings of the IEEE, 1998, 86(11):2278-2324.
    [14] ATZORI M, GIJSBERTS A, CASTELLINI C, et al. Electromyography data for non-invasive naturally-controlled robotic hand prostheses[J]. Nature, 2014, 1:605-610.
    [15] 乐毅.深度学习——Caffe之经典模型详解与实战[M].北京:电子工业出版社, 2016:147-149.
    [16] 许凡, 程华, 房一泉. 基于CLSTM的步态分类方法[J]. 华东理工大学学报(自然科学版), 2017, 43(4):553-558.
    [17] NAIR V, HINTON G E. Rectified linear units improve restricted boltzmann machines[C]//Proceedings of the 27th International Conference on Machine Learning (ICML-10). Haifa, Israel:ACM, 2010:807-814.
    [18] 陈耀丹, 王连明. 基于卷积神经网络的人脸识别方法[J]. 东北师大学报(自然科学版), 2016, 48(2):70-76.
    [19] BOUREAU Y L, BACH F, LECUN Y, et al. Learning mid-level features for recognition[C]//2010 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). USA:IEEE, 2010:2559-2566.
    [20] ENGLEHART K, HUDGINS B. A robust, real-time control scheme for multifunction myoelectric control[J]. IEEE Transactions on Biomedical Engineering, 2003, 50(7):848-854.
    [21] SRIVASTAVA N, HINTON G E, KRIZHEVSKY A, et al. Dropout:A simple way to prevent neural networks from overfitting[J]. Journal of Machine Learning Research, 2014, 15(1):1929-1958.
    [22] LECUN Y, JACKEL L D, BOTTOU L, et al. Comparison of learning algorithms for handwritten digit recognition[C]//International Conference on Artificial Neural Networks.USA:IEEE, 1995:53-60.
  • 加载中
图(1)
计量
  • 文章访问数:  1455
  • HTML全文浏览量:  253
  • PDF下载量:  550
  • 被引次数: 0
出版历程
  • 收稿日期:  2017-03-31
  • 刊出日期:  2018-04-22

目录

    /

    返回文章
    返回