第13卷第4期 智能系统学报 Vol.13 No.4 2018年8月 CAAI Transactions on Intelligent Systems Aug.2018 D0:10.11992/tis.201708003 网络出版地址:http:/kns.cnki.net/kcms/detail/23.1538.TP.20180402.1805.012.html 对偶树复小波与空域信息的手势识别分类研究 贾鹤鸣,朱传旭',张森,杨泽文2,何东旭2 (1.东北林业大学机电工程学院,黑龙江哈尔滨150040;2.哈尔滨工程大学自动化学院,黑龙江哈尔滨 150001) 摘要:为提高手势识别中特征获取的有效性,本文提出空域特征与对偶树复小波变换特征相结合的融合特 征,主要包括水平位置、竖直位置、长宽比、矩形度、u矩7个分量,及11雏空域特征与对偶树复小波变换的 16维特征进行融合后得到的27维特征。针对分类器优化算法,提出进行训练样本优选的最优距离-支持向量 机BD-SVM)分类方法。最后的实验结果表明,对“1~9”手势进行测试,当采用径向基核函数时,平均识别精度 最高,为90.33%,平均识别时间为0.026s,说明所提出的方法能够较好地进行静态手势识别,具有较高的训练 速度和辨识精度。 关键词:手势识别;空域特征;对偶树复小波:特征融合;分类器优化;BD-SVM;径向基核函数:静态测试 中图分类号:TP273文献标志码:A文章编号:1673-4785(2018)04-0619-06 中文引用格式:贾鹤鸣,朱传旭,张森,等.对偶树复小波与空域信息的手势识别分类研究.智能系统学报,2018,13(4): 619-624. 英文引用格式:JIA Heming,.ZHU Chuanxu,.ZHANG Sen,etal.Research on gesture recognition and classification of dual-tree complex wavelet and spatial information.CAAI transactions on intelligent systems,2018,13(4):619-624. Research on gesture recognition and classification of dual-tree complex wavelet and spatial information JIA Heming',ZHU Chuanxu',ZHANG Sen',YANG Zewen',HE Dongxu' (1.College of Mechanical and Electrical Engineering,Northeast Forestry University,Harbin 150040,China;2.College of Automa- tion,Harbin Engineering University,Harbin 150001,China) Abstract:To improve the validity of features obtained in gesture recognition,in this paper,we propose a fusion feature that combines spatial and dual-tree complex wavelet transform features.These features mainly include seven compon- ents(horizontal position,vertical position,aspect ratio,rectangular degree,Hu moments,etc.)and 27 dimensional fea- tures,comprising 11 dimensional spatial features and 16 dimensional dual-tree complex wavelet transform features.We employ the optimal distance support vector machine(BD-SVM)classification method to optimize training samples for the classifier optimization algorithm.The experimental results show that,in a test of gestures"1~9"using the RBF ker- nel function,the highest average recognition accuracy is 90.33%and the average recognition time is 0.026 s.These res- ults reveal that the proposed method demonstrates excellent static gesture recognition,a high training speed,and accur- acy in identification. Keywords:gesture recognition:spatial feature:dual-tree complex wavelet:feature fusion:classifier optimization:BD- SVM:radial basis kernel function:static test 收稿日期:2017-08-03.网络出版日期:2018-04-03 手势语言作为一种常用的交流语言,通过不 基金项目:中央高校基本科研业务费专项资金项目(2572014BB03: 国家自然科学基金项目(31470714.51609048):黑龙 同手势的组合、不同手形的变化,能够表达多种 江省研究生教育创新工程项目GXM_HLJ2016014), 通信作者:贾鹤鸣.E-mail:jiaheminglucky99@126.com. 复杂的含义,在非声音传递信息的方式中,如特DOI: 10.11992/tis.201708003 网络出版地址: http://kns.cnki.net/kcms/detail/23.1538.TP.20180402.1805.012.html 对偶树复小波与空域信息的手势识别分类研究 贾鹤鸣1 ,朱传旭1 ,张森1 ,杨泽文2 ,何东旭2 (1. 东北林业大学 机电工程学院,黑龙江 哈尔滨 150040; 2. 哈尔滨工程大学 自动化学院,黑龙江 哈尔滨 150001) 摘 要:为提高手势识别中特征获取的有效性,本文提出空域特征与对偶树复小波变换特征相结合的融合特 征,主要包括水平位置、竖直位置、长宽比、矩形度、Hu 矩 7 个分量,及 11 维空域特征与对偶树复小波变换的 16 维特征进行融合后得到的 27 维特征。针对分类器优化算法,提出进行训练样本优选的最优距离–支持向量 机 (BD-SVM) 分类方法。最后的实验结果表明,对“1~9”手势进行测试,当采用径向基核函数时,平均识别精度 最高,为 90.33%,平均识别时间为 0.026 s,说明所提出的方法能够较好地进行静态手势识别,具有较高的训练 速度和辨识精度。 关键词:手势识别;空域特征;对偶树复小波;特征融合;分类器优化;BD-SVM;径向基核函数;静态测试 中图分类号:TP273 文献标志码:A 文章编号:1673−4785(2018)04−0619−06 中文引用格式:贾鹤鸣, 朱传旭, 张森, 等. 对偶树复小波与空域信息的手势识别分类研究[J]. 智能系统学报, 2018, 13(4): 619–624. 英文引用格式:JIA Heming, ZHU Chuanxu, ZHANG Sen, et al. Research on gesture recognition and classification of dual-tree complex wavelet and spatial information[J]. CAAI transactions on intelligent systems, 2018, 13(4): 619–624. Research on gesture recognition and classification of dual-tree complex wavelet and spatial information JIA Heming1 ,ZHU Chuanxu1 ,ZHANG Sen1 ,YANG Zewen2 ,HE Dongxu2 (1. College of Mechanical and Electrical Engineering, Northeast Forestry University, Harbin 150040, China; 2. College of Automation, Harbin Engineering University, Harbin 150001, China) Abstract: To improve the validity of features obtained in gesture recognition, in this paper, we propose a fusion feature that combines spatial and dual-tree complex wavelet transform features. These features mainly include seven components (horizontal position, vertical position, aspect ratio, rectangular degree, Hu moments, etc.) and 27 dimensional features, comprising 11 dimensional spatial features and 16 dimensional dual-tree complex wavelet transform features. We employ the optimal distance support vector machine (BD-SVM) classification method to optimize training samples for the classifier optimization algorithm. The experimental results show that, in a test of gestures “1~9” using the RBF kernel function, the highest average recognition accuracy is 90.33% and the average recognition time is 0.026 s. These results reveal that the proposed method demonstrates excellent static gesture recognition, a high training speed, and accuracy in identification. Keywords: gesture recognition; spatial feature; dual-tree complex wavelet; feature fusion; classifier optimization; BDSVM; radial basis kernel function; static test 手势语言作为一种常用的交流语言,通过不 同手势的组合、不同手形的变化,能够表达多种 复杂的含义,在非声音传递信息的方式中,如特 收稿日期:2017−08−03. 网络出版日期:2018−04−03. 基金项目:中央高校基本科研业务费专项资金项目 (2572014BB03); 国家自然科学基金项目 (31470714,51609048);黑龙 江省研究生教育创新工程项目 (JGXM_HLJ_2016014). 通信作者:贾鹤鸣. E-mail:jiaheminglucky99@126.com. 第 13 卷第 4 期 智 能 系 统 学 报 Vol.13 No.4 2018 年 8 月 CAAI Transactions on Intelligent Systems Aug. 2018