多目标优化的静态手语识别算法研究
|
摘要
手势是人类与计算机交互的直观方式,随着人工智能技术的发展,以机器为核心的计算模式正朝着以人为中心的计算模式转变,自然且符合人类习惯的人机交互(HCI)方式逐渐成为目前研究的热点。一个高效的人机交互系统应该以良好的识别精度和识别速度为目标。文中提出了一种基于深度信息的静态手势识别方法,包含了手势大小、光照和旋转变化等因素的影响。从识别精度和速度两个方面,综合比较了如Hu矩、Zernike矩、伪Zernike矩、傅里叶描述符和Gabor特征等几种常见的图像特征描述符。手势识别采用多层感知器,它具有结构灵活、识别速度快等特点。为了提高识别精度、减少计算量,特征向量和神经网络均通过基于NSGA-II的多目标进化算法调整。在对手语识别的进一步探究中,对所提出的方法的有效性进行验证。
Gesture is an intuitive way for human beings to interact with computers.With the development of artificial intelligence technology,the machine-centric computing mode is shifting toward the human-centered computing mode,and the human-computer interaction(HCI) methods that are natural and in line with human habits are gradually become the focus of current research.An efficient humancomputer interaction system should aim for better recognition accuracy and recognition speed.For this,we present a static gesture recognition method based on depth information,including the influence of gesture size,lighting,and rotation changes.Several common image feature descriptors such as Hu moment,Zernike moment,pseudo Zernike moment,Fourier descriptor and Gabor feature are comprehensively compared from their respective recognition accuracy and speed.Gesture recognition uses a multi-layer sensor which has a flexible structure and fast recognition speed.In order to improve the recognition accuracy and reduce the computational complexity,the eigenvectors and the neural network are adjusted by the multi-objective evolutionary algorithm based on NSGA-II.In the further exploration of the recognition of the opponent language,the validity of the proposed method is verified.
Gesture is an intuitive way for human beings to interact with computers.With the development of artificial intelligence technology,the machine-centric computing mode is shifting toward the human-centered computing mode,and the human-computer interaction(HCI) methods that are natural and in line with human habits are gradually become the focus of current research.An efficient humancomputer interaction system should aim for better recognition accuracy and recognition speed.For this,we present a static gesture recognition method based on depth information,including the influence of gesture size,lighting,and rotation changes.Several common image feature descriptors such as Hu moment,Zernike moment,pseudo Zernike moment,Fourier descriptor and Gabor feature are comprehensively compared from their respective recognition accuracy and speed.Gesture recognition uses a multi-layer sensor which has a flexible structure and fast recognition speed.In order to improve the recognition accuracy and reduce the computational complexity,the eigenvectors and the neural network are adjusted by the multi-objective evolutionary algorithm based on NSGA-II.In the further exploration of the recognition of the opponent language,the validity of the proposed method is verified.
引文
[1]侯亭亭,肖秦琨,杨永侠.基于动态贝叶斯网络的手势识别[J].国外电子测量技术,2015,34(1):36-39.
[2]赵一丹,肖秦琨,高嵩.基于模糊神经网络和图模型推理的动作识别[J].计算机技术与发展,2018,28(8):80-85.
[3]GUO Yun,LIU Chunping,GONG Shengrong.Improved algorithm for Zernike moments[C]//International conference on control,automation and information sciences.Changshu,China:IEEE,2015:307-312.
[4]吴良圆,魏书宁,周棒棒,等.基于PCA降维的分层超限学习机手势识别方法[J].电子测量技术,2017,40(3):82-88.
[5]蒲兴成,王涛,张毅.基于改进Hu矩算法的Kinect手势识别[J].计算机工程,2016,42(7):165-172.
[6]CHAUDHARY A,RAHEJA J,DAS K,et al.Intelligent approaches to interact with machines using hand gesture recognition in natural way:a survey[J].International Journal of Computer Science&Engineering Survey,2011,2(1):122-133.
[7]HASAN H,KAREEM S.Static hand gesture recognition using neural networks[J].Artificial Intelligence Review,2014,41(2):147-181.
[8]AOWAL M A,ZAMAN A S,RAHMAN S M M,et al.Static hand gesture recognition using discriminative 2D Zernike moments[C]//IEEE region 10 conference.Bangkok,Thailand:IEEE,2015:1-5.
[9]LEUNG F H F,LAM H K,LING S H,et al.Tuning of the structure and parameters of a neural network using an improved genetic algorithm[J].IEEE Transactions on Neural Networks,2003,14(1):79-88.
[10]何阳清,葛元,王林泉.应用几何矩和边缘检测的手势识别算法[J].计算机工程,2005,31(15):165-166.
[11]ZERNIKE V F.Beugungstheorie des schneidenver-fahrens und seiner verbesserten form,der phasenkontrastmethode[J].Physica,1934,1(7):689-704.
[12]CHONG C W,RAVEENDRAN P,MUKUNDAN R.An efficient algorithm for fast computation of Pseudo-Zernike moments[J].International Journal of Pattern Recognition&Artificial Intelligence,2003,17(6):1011-1023.
[13]ZHANG Dengsheng,LU Guojun.A comparative study of curvature scale space and Fourier descriptors for shape-based image retrieval[J].Journal of Visual Communication&Image Representation,2003,14(1):39-57.
[14]付余伟.基于非高斯Gabor滤波器的掌纹识别算法研究[D].昆明昆明理工大学,2017.
[15]HUANG D Y,HU W C,CHANG S H.Vision-based hand gesture recognition using PCA+gabor filters and SVM[C]//5th international conference on intelligent information hiding and multimedia signal processing.Kyoto,Japan:IEEE,2009:1-4.
[16]路艳雪.NSGA-Ⅱ多目标优化算法的改进及应用研究[D].太原:太原理工大学,2017.
[17]SEADA H,DEB K.U-NSGA-III:a unified evolutionary optimization procedure for single,multiple,and many objectives:proof-of-principle results[C]//International conference on evolutionary multi-criterion optimization.[s.l.]:[s.n.],2015:34-49.
[18]KOHAVI R,JOHN G H.Wrappers for feature subset selection[J].Artificial Intelligence,1997,97(1-2):273-324.
[2]赵一丹,肖秦琨,高嵩.基于模糊神经网络和图模型推理的动作识别[J].计算机技术与发展,2018,28(8):80-85.
[3]GUO Yun,LIU Chunping,GONG Shengrong.Improved algorithm for Zernike moments[C]//International conference on control,automation and information sciences.Changshu,China:IEEE,2015:307-312.
[4]吴良圆,魏书宁,周棒棒,等.基于PCA降维的分层超限学习机手势识别方法[J].电子测量技术,2017,40(3):82-88.
[5]蒲兴成,王涛,张毅.基于改进Hu矩算法的Kinect手势识别[J].计算机工程,2016,42(7):165-172.
[6]CHAUDHARY A,RAHEJA J,DAS K,et al.Intelligent approaches to interact with machines using hand gesture recognition in natural way:a survey[J].International Journal of Computer Science&Engineering Survey,2011,2(1):122-133.
[7]HASAN H,KAREEM S.Static hand gesture recognition using neural networks[J].Artificial Intelligence Review,2014,41(2):147-181.
[8]AOWAL M A,ZAMAN A S,RAHMAN S M M,et al.Static hand gesture recognition using discriminative 2D Zernike moments[C]//IEEE region 10 conference.Bangkok,Thailand:IEEE,2015:1-5.
[9]LEUNG F H F,LAM H K,LING S H,et al.Tuning of the structure and parameters of a neural network using an improved genetic algorithm[J].IEEE Transactions on Neural Networks,2003,14(1):79-88.
[10]何阳清,葛元,王林泉.应用几何矩和边缘检测的手势识别算法[J].计算机工程,2005,31(15):165-166.
[11]ZERNIKE V F.Beugungstheorie des schneidenver-fahrens und seiner verbesserten form,der phasenkontrastmethode[J].Physica,1934,1(7):689-704.
[12]CHONG C W,RAVEENDRAN P,MUKUNDAN R.An efficient algorithm for fast computation of Pseudo-Zernike moments[J].International Journal of Pattern Recognition&Artificial Intelligence,2003,17(6):1011-1023.
[13]ZHANG Dengsheng,LU Guojun.A comparative study of curvature scale space and Fourier descriptors for shape-based image retrieval[J].Journal of Visual Communication&Image Representation,2003,14(1):39-57.
[14]付余伟.基于非高斯Gabor滤波器的掌纹识别算法研究[D].昆明昆明理工大学,2017.
[15]HUANG D Y,HU W C,CHANG S H.Vision-based hand gesture recognition using PCA+gabor filters and SVM[C]//5th international conference on intelligent information hiding and multimedia signal processing.Kyoto,Japan:IEEE,2009:1-4.
[16]路艳雪.NSGA-Ⅱ多目标优化算法的改进及应用研究[D].太原:太原理工大学,2017.
[17]SEADA H,DEB K.U-NSGA-III:a unified evolutionary optimization procedure for single,multiple,and many objectives:proof-of-principle results[C]//International conference on evolutionary multi-criterion optimization.[s.l.]:[s.n.],2015:34-49.
[18]KOHAVI R,JOHN G H.Wrappers for feature subset selection[J].Artificial Intelligence,1997,97(1-2):273-324.