基于菲涅尔区模型的年龄段识别方法研究

基于步态特征的行为识别技术是一种新兴的人机交互技术,在安保监控、图像处理等许多领域正发挥着巨大作用.目前已经提出的各种行为识别的技术手段中,基于商用的WIFI设备的相关技术因其易于部署、设备成本不高、对应用环境要求低等优势而得以广泛应用.现有的步态识别研究只可以识别出人的步态,但无法根据步态识别年龄段.在本文中,探讨了菲涅尔区域理论,提出了一种基于WiFi物理层信息的年龄段识别方法,即利用步态识别年龄段.使用菲涅尔区模型来学习步态特征,首先将动作数据集进行去噪处理来收集信道状态信息(CSI).然后基于菲涅尔区域模型提取信道相位信息的特征值来描述步态动作,最后利用分类算法实现年龄段识别.实验结果表明,该方法具有良好的鲁棒性,在不同的实验环境中的平均识别准确率达到91.2％.因此,本文提出的系统可以应用于复杂的室内场景,具有广泛的应用前景.

Research on age recognition method based on Fresnel zone model

Gait based behavior recognition technology is a new human-computer interaction technology,which plays an important role in many fields such as security monitor ing,image pr ocessing and soon.At present,among all kinds of behavior recognition technologie s,the tech nology based on commercial WiFi equipment is widely used because of its advantag es of easy deployment,low cost of equipment and low requirements for application environment.The existing gait recognition research can only recognize the human gait,but can not recognize the age according to the gait.In this paper,we discuss the F resnel region theory,and propose an age recognition method based on WiFi physical layer infor mation. That is,using gait to recognize age.This work use Fresnel zone model to learn gait fe atures.Firstly, we denoise the motion data set to collect channel state information (CSI).Then, based on the Fresnel region model,the eigenvalues of the channel phase information are extra cted to describe the gait action.Finally,the classification algorithm is used to reali ze the age recognition.Experimental results show that the method has good robustness,and the average recognition accuracy in different experimental environments reaches 91.2%.There fore,the system proposed in this paper can be applied to complex indoor scenes,a nd has a wide range of application prospects.