摘要
传统动态手势识别方法中存在硬件成本高,推广难度大等局限性.文章提出一种基于信道状态信息的复杂动态手势识别方法CSI-Num,该方法可用来实现对空中数字手势的高效识别.CSI-Num识别过程主要分为两个阶段:数据提取处理阶段与手势匹配识别阶段.提取处理阶段,是将采集到的数据,选取能够反映手势动作的子载波特征值作为被选信号,通过小波阈值函数和五点三次平滑方法对信号进行降噪平滑;匹配识别阶段,提取有效手势数据,使用k均值聚类算法和动态时间规整算法特性相融合的K-DTW匹配算法识别出不同数字的手势动作.实验结果表明,针对不同环境的室内场景,相应地调整参数设置,CSI-Num可以高效地识别出不同数字的手势动作,且具有较高鲁棒性.
The traditional dynamic gesture recognition method has some limitations,such as high hardware cost and difficulty in popularizing.The paper proposes a complex dynamic gesture recognition method CSI-Num,based on channel state information,which can realize the efficient recognition of digital gestures in the air.The process of CSI-Num recognition consists of two stages:data extraction processing and gesture matching recognition.In the stage of extracting and processing,the acquired data is selected as the selected signal,and by wavelet threshold function and five-spot triple smoothing method make the signal de-noised and smooth.In the phase of matching recognition,the valid gesture data is extracted,and using the K-DTW matching algorithm(combined k-means clustering algorithm and the characteristics of Dynamic Time Warping)can recognize the gesture actions with different numbers.The experimental results showthat CSI-Num can efficiently recognize different digital gestures and is robust to different indoor scenes and adjust the parameters accordingly.
作者
党小超
刘洋
郝占军
曹渊
DANG Xiao-chao;LIU Yang;HAO Zhan-jun;CAO Yuan(College of Computer Science and Engineering,Northwest Normal University,Lanzhou 730070,China;Gansu Province Internet of Things Engineering Research Center,Lanzhou 730070,China)
出处
《小型微型计算机系统》
CSCD
北大核心
2020年第1期200-205,共6页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(61762079,61662070)资助
甘肃省科技重点研发项目(17YF1GA015)资助
