计算机科学 ›› 2017, Vol. 44 ›› Issue (7): 299-303.doi: 10.11896/j.issn.1002-137X.2017.07.054
梁文乐,黄元元,胡作进
LIANG Wen-le, HUANG Yuan-yuan and HU Zuo-jin
摘要: 动态手语可以利用其轨迹与关键手型加以描述。大量的统计实验数据表明,大多数的常用手语通过轨迹曲线的匹配即可实现识别,因此,提出一种针对动态手语的分级匹配识别算法。首先利用体感设备获取手势轨迹,并根据轨迹的点密度分布设计了一种关键帧检测算法以提取手势的关键手型,结合轨迹的曲线特征,实现对动态手语的精确描述。然后利用优化的动态时间规整(DTW)算法完成对手语的一级匹配,即轨迹匹配。若此时可以得到识别结果,那么识别过程可以结束,否则进入二级匹配,即针对关键手型再做匹配识别,从而得到最终的识别结果。实验证明,所提算法不仅实时性好,识别的准确率也较高。
[1] STARNER T,PENTLAND A.Real-time american sign lan-guage recognition from video using hidden markov models[M]∥Motion-Based Recognition.Springer Netherlands,1997:227-243. [2] MARAQA M,AL-ZBOUN F,DHYABAT M,et al.Recognition of Arabic Sign Language(ArSL) Using Recurrent Neural Networks[C]∥Applications of Digital Information and Web Techno-logyies,2008(ICADIWT 2008).IEEE,2008:41-52. [3] ARGYROS A,LOURAKIS M I A.Binocular hand tracking and reconstruction based on 2D shape matching[C]∥18th International Conference on Pattern Recognition,2006(ICPR 2006).IEEE,2006:207-210. [4] VOGLER C,METAXAS D.Parallel hidden markov models for american sign language recognition[C]∥IEEE International Conference on Computer Vision,1999.IEEE,1999:116-122. [5] JANG Y.Gesture recognition using depth-based hand tracking for contactless controller application[C]∥2012 IEEE International Conference on Consumer Electronics(ICCE).2012:297-298. [6] CHAI X,LI G,LIN Y,et al.Sign language recognition andtranslation with kinect[C]∥IEEE Conf.on AFGR.2013:1-2. [7] MARIN G,DOMINIO F,ZANUTTIGH P.Hand gesture recognition with jointly calibrated Leap Motion and depth sensor[J].Multimedia Tools and Applications,2016,72(22):14991-15015. [8] SHEN Y L.Analysis of Chinese sign language morpheme [J].Chinese Journal of Special Education,1993(1):1-13.(in Chinese) 沈玉林.中国手语语素分析[J].特殊教育研究,1993(1):1-13 [9] DOLIOTIS P,STEFAN A,MCMURROUGH C,et al.Comparing Gesture Recognition Accuracy Using Color and Depth Information[C]∥Proceedings of the Fourth International Confe-rence on Pervasive Technologies Related to Assistive Environments(PETR) Crete.Greece,2011:1-7. [10] CARMONA J,CLIMENT J.A performance evaluation of hmm and dtw for gesture recognition[C]∥Progress in Pattern Recognition,Image Analysis,Computer Vision,and Applications.2012:236-243. [11] RAHEJA J L,MINHAS M,PRASHANTH D,et al.Robustgesture recognition using Kinect:A comparison between DTW and HMM[J].Optik-International Journal for Light and Electron Optics,2015,126(11):1098-1104. [12] ZHUANG Y T,RUI Y,HUANG T S,et al.Adaptive KeyFrame Extraction Using Unsupervised Clustering[C]∥Proc.of IEEE Int.Conf.on Image Processing.1998:866-870. |
No related articles found! |
|