基于深度时空特征模型的非接触式心率估计方法  

A Non-Contact Heart Rate Estimation Method Based on a Deep Spatiotemporal Feature Model

作　　者：乔圣洋 徐慧英[1] 朱信忠[1] 周宇豪 魏远旺 张峻嘉 Qiao Shengyang;Xu Huiying;Zhu Xinzhong;Zhou Yuhao;Wei Yuanwang;Zhang Junjia(School of Computer Science and Technology,Zhejiang Normal University,Jinhua,Zhejiang 321004;College of Information Science and Engineering,Jiaxing University,Jiaxing,Zhejiang 314001;Research Institute of Information Network&Artificial Intelligence,Jiaxing University,Jiaxing,Zhejiang 314001;China MCC20 Group Corp.,Ltd,Baoshan,Shanghai 201900)

机构地区：[1]浙江师范大学计算机科学与技术学院,浙江金华321004 [2]嘉兴大学信息科学与工程学院,浙江嘉兴314001 [3]嘉兴大学信息网络与智能研究院,浙江嘉兴314001 [4]中国二十冶集团有限公司,上海宝山201900

出　　处：《嘉兴大学学报》2024年第6期36-45,共10页Journal of Jiaxing University

基　　金：国家自然科学基金项目(62376252);浙江省自然科学基金重点项目(LZ22F030003);嘉兴市级公益性研究计划项目(2023AY11030);嘉兴学院大学生科研训练(SRT)计划项目(8517231499)。

摘　　要：为了提升心率计算在不同光照和运动条件下的准确率,提出了一种基于深度时空特征模型的创新方法——非接触式心率计算方法:对输入的人脸视频分别利用卷积神经网络(CNN)提取空间特征和利用长短期记忆网络(LSTM)提取时序特征,将这两种特征融合之后,模型对微弱的血液容积脉冲(BVP)信号展现出更强的表征能力.在自采的ZJXU-Phys数据集上进行验证,新方法在各种条件下均表现出色,其有效性和实用性得到了验证.In order to improve the accuracy of heart rate calculation under different lighting and exercise conditions,we propose an innovative non-contact heart rate calculation method based on a deep spatiotemporal feature model.We use convolutional neural networks(CNN)to extract spatial features from the input face video,and long short-term memory networks(LSTM)to extract temporal features.By fusing these two features,our model exhibits a stronger representation ability for weak blood volume pulse(BVP)signals.To verify the effectiveness of this method,we conduct experiments on the self-collected ZJXU-Phys dataset,and the experimental results show that the proposed method performs well under various conditions,verifying its effectiveness and practicality.

关 键 词：深度时空特征模型 心率估计 远程光电容积脉搏波描记法 ZJXU-Phys数据集 

分 类 号：TP302.7[自动化与计算机技术—计算机系统结构]