心率与血压是表征个人健康状态的重要生理参数，长期监测可以有效预防疾病。成像式光电容积描记是通过摄像头捕捉心脏跳动周期造成皮肤颜色变化的方法，利用该方法获取脉搏信号进行非接触式的心率与血压估计研究，不仅克服了传统仪器需要接触式传感器带来的弊端，还具有丰富的应用场景。本文将基于成像式光电容积描记提取脉搏信号完成心率与血压的估计，主要工作如下：

（1）非接触式心率与血压估计的关键在于提取高质量的脉搏信号，本文提出一种自适应皮肤像素选择算法（SA-HSV）加入脉搏信号提取的过程中，减少非皮肤像素的干扰，改善传统皮肤像素选择算法中存在的误判现象。将该算法应用于VIPL-HR公开数据集与自采集数据集并设置多种实验场景对比分析SA-HSV的性能。在自采集数据中得到心率估计的平均绝对误差为3.56bpm，满足中国人民共和国医药行业规定的误差范围小于5bpm。

（2）通过提取脉搏信号中与血压相关的特征值，建立基于脉搏信号特征的血压估计模型。利用自采集数据集分别搭建径向基神经网络与BP神经网络对血压估计。通过仿真BP神经网络的估计误差低于径向基神经网络，其中舒张压的估计误差为1.45 11.10mmHg，收缩压的估计误差为0.71 8.87 mmHg。为了提高估计准确度，将智能优化算法与神经网络结合，分别研究了遗传算法以及粒子群算法优化的神经网络。优化后的网络估计性能均优于BP神经网络，其中粒子群算法的优化效果最好，最终收缩压的估计误差为0.13 7.91mmHg，舒张压的估计误差为-0.12 6.40mmHg。估计误差满足AAMI与BHS血压测量标准，在一定估计误差范围内收缩压的估计准确率为92.39%，舒张压估计准确率为98.36%。

实验结果验证了自适应像素选择算法应用于非接触式心率与血压估计的有效性，本文所提出的方法为非接触式生理参数估计领域提供了实际应用的参考。

Heart rate and blood pressure are important physiological parameters that characterize an individual's health status. Imaging photoplethysmography is a method of capturing skin color changes caused by the heart beating cycle through a camera. The pulse signal obtained by this method can be used for the estimation of heart rate and blood pressure. This paper will complete the estimation of heart rate and blood pressure based on imaging photoplethymography. The main tasks are as follows:

（1）The key to non-contact heart rate and blood pressure estimation is to extract high-quality pulse signals. In this paper, an adaptive skin pixel selection algorithm (SA-HSV) is proposed to process the video frame images of the extracted pulse signal and improve the misjudgment existing in the traditional skin pixel selection algorithm. phenomenon, improve the recognition accuracy. The algorithm is applied to the VIPL-HR public data set and the self-collected data set, and a range of experimental scenarios are set up to evaluate and analyze SA-HSV performance. The average absolute error of heart rate estimation obtained from self-collected data is 3.56bpm.

（2）By extracting the features related to blood pressure in the pulse signal, a blood pressure estimation model based on the features of the pulse signal is established. The self-collected data sets were used to build radial basis neural network and BP neural network respectively to verify the blood pressure estimation. The estimation error of the simulated BP is lower than radial basis neural network. The particle swarm method has a better optimization effect, with a systolic blood pressure estimate error of 0.13 7.91mmHg and a diastolic blood pressure estimation error of -0.12 6.40mmHg. The estimated error matched the blood pressure measurement standards of AAMI and BHS. Within a specific estimation error range, the estimated accuracy of systolic blood pressure was 92.39%, the estimated accuracy of diastolic blood pressure was 98.36%.

The experimental results validate the efficacy of the adaptive pixel selection technique when used to estimate non-contact heart rate and blood pressure. The method proposed in this paper provides a reference for practical applications in the field of non-contact physiological parameter estimation.

[1] Pearson MJ, Smart NA. Exercise therapy and autonomic function in heart failure patients: a systematic review and meta-analysis[J]. Heart Fail Rev. 2018,23(1):91-108.

[2] Wang Y, Yin L, Hu B, et al. Association of heart rate with cardiovascular events and mortality in hypertensive and normotensive population: a nationwide prospective cohort study[J]. Ann Transl Med, 2021,9(11):917.

[3] Seravalle G, Grassi G. Heart rate as cardiovascular risk factor[J]. Postgraduate Medicine, 2020,132(4):358-367.

[4] 胡盛寿. 中国心血管健康与疾病报告2019[R]. 北京:国家心血管病中心, 2019.

[5] Rouast P V, Adam M T P, R Chiong, et al. Remote heart rate measurement using low-cost RGB face video[J]. Frontiers of Computer Science, 2018,12(5):858- 872.

[6] Saini V. Driver drowsiness detection system and techniques: A Review[J]. International Journal of Computer Science & Information Technology, 2014,5(3):4245-4249.

[7] Vilke G M, Jesse M A, Cronin A O, et al. Clinical features of patients with COVID-19: is temperature screening useful? [J]. ScienceDirect. The Journal of Emergency Medicine 2020,59(6):952-956.

[8] Stave G M, Smith S E, Hymel P A, et al. Worksite Temperature Screening for COVID-19[J]. Journal of occupational and environmental medicine, 2021,63(8):638-641.

[9] Natarajan A, Su H W, Heneghan C. Assessment of physiological signs associated with COVID-19 measured using wearable devices[J]. Digital Medicine, 2020,3(1):156.

[10] Pavri B B, Kloo J, Farzad D, et al. Behavior of the PR interval with increasing heart rate in patients with COVID-19[J]. Heart Rhythm, 2020,17(9):1434-1438.

[11] Ting W, Vladimir B, Hans J, et al. Photoplethysmography imaging: a new noninvasive and noncontact method for mapping of the dermal perfusion changes[C]// European Biomedical Optics Week. International Society for Optics and Photonics, 2000:62-70.

[12] Verkruysse W, Svaasand L. Remote plethysmographic imaging using ambient light[J]. Optics express, 2008,16(26):21434-21445.

[13] Mcduff D, Gontarek S, Picard R W. Remote detection of photoplethysmographic systolic and diastolic peaks using a digital camera[J]. Biomedical Engineering IEEE Transactions on, 2014,61(12):2948-2954．

[14] 何璇. 基于面部视频的非接触式血流信号检测系统研究与实现[D].安徽大学, 2017.

[15] Harford M, Catherall J, Gerry S, et al. Availability and performance of image- based, non-contact methods of monitoring heart rate, blood pressure, respiratory rate, and oxygen saturation: a systematic review[J]. Physiological Measurement, 2019,40(6).

[16] Gonzalez V C, Fuentes S, Torrico D, et al. Non-contact heart rate and blood pressure estimations from video analysis and machine learning modelling applied to food sensory responses: A case study for chocolate[J]. Sensors (Basel, Switzerland), 2018,18(6):1802-1820.

[17] Yang Zhao, Yang Xuezhi, Wu Xiu. Motion-tolerant heart rate estimation from face videos using derivative filter[J]. Multimedia Tools and Applications, 2019,78(18): 26747-26757．

[18] Lin Yuchen, Chou Naikuan, Lin Guanyou, et al. A real-time contactless pulse rate and motion status monitoring system based on complexion tracking[J]. Sensors, 2017,17(7):1490-1503.

[19] Song Rencheng, Zhang Senle, Li Chang, et al. Heart rate estimation from facial videos using a spatiotemporal representation with convolutional neural networks[J]. IEEE Transactions on Instrumentation and Measurement, 2020,69(10):7411-7421.

[20] Favilla R, Zuccala V C, Coppini G.Heart rate and heart rate variability from single-channel video and ICA integration of multiple signals[J]. IEEE Journal of Biomedical and Health Informatics, 2018,23(6):2398-2408.

[21] 蔡凯. 非接触式心率测量技术研究[D]. 南京邮电大学, 2019.

[22] Fouad R M, Omer O A, Aly M H. Optimizing Remote Photoplethysmography Using Adaptive Skin Segmentation for Real-Time Heart Rate Monitoring[J]. IEEE Access, 2019,7(99):76513-76528.

[23] Kaito I, Ryota M, Takashi G, et al. Removing the influence of light on the face from display in iPPG[J]. Artificial Life and Robotics, 2020,25(3):377-382.

[24] 李晓媛, 武鹏, 刘允, 等. 基于人脸视频的心率参数提取[J].光学精密工程, 2020, 28(03):548-557.

[25] Grabovskis A, Marcinkevics Z, Rubins U, et al. Effect of probe contact pressure on the photoplethysmographic assessment of conduit artery stiffness[J]. Journal of Biomedical Optics, 2013,18(2):27004.

[26] Hsiu H, Hsu C L, Chen C T, et al. Correlation of harmonic components between the blood pressure and photoplethysmography waveforms following local-heating stimulation[J]. 2012,2(4):248-253.

[27] Wang L, Wei Z, Ying X, et al. A novel neural network model for blood pressure estimation using photoplethesmography without electrocardiogram[J]. Journal of Healthcare Engineering, 2018.

[28] Yongbo Liang, Guiyong Liu, Mohamed Elgendi. A new, short-recorded photo- plethysmogramdataset for blood pressure monitoring in China[J]. Scientific data, 2018.

[29] Liu D, Görges M, Jenkins S A. University of Queensland vital signs dataset: Development of an accessible repository of anesthesia patient monitoring data for research[J]. Anesthesia and analgesia, 2012,114(3):584-589.

[30] Khalid S G, Zhang J, Chen F, et al. Blood pressure estimation using photoplethysmography only: comparison between different machine learning approaches[J]. Journal of Healthcare Engineering, 2018.

[31] Satu R, Harri L, Tapio T. Comparison of photoplethysmogram measured from wrist and finger and the effect of measurement location on pulse arrival time[J]. Physiological Measurement, 2018,39(7):075010.

[32] Simjanoska M, Gjoreski M, Gams M, Madevska B A. Non-invasive blood pressure estimation from ECG using machine learning techniques[J]. Sensors (Basel, Switzerland), 2018,18(4): 1160.

[33] Xiaoman Xing, Zhimin Ma, Mingyou Zhang, et al. An unobtrusive and calibration-free blood pressure estimation method using photoplethysmography and biometrics[J]. Scientific Reports, 2019,9(1):8611.

[34] Nath R K, Thapliyal H, Caban-Holt A. Towards photoplethysmogram based non-invasive blood pessure classification[C]// 2018 IEEE International Symposium on Smart Electronic Systems (iSES) (Formerly iNiS). IEEE, 2018.

[35] Ahmed S A, Kemal P, Abdullah A, et al. Gaussian process regression (GPR) based non-invasive continuous blood pressure prediction method from cuff oscillometric signals[J]. Applied Acoustics, 2020,164(7):107256.

[36] Mousavi S S, Firouzmand M, Charmi M, et al. Blood pressure estimation from appropriate and inappropriate PPG signals using A whole-based method[J]. Biomedical signal processing and control, 2019,47(1):196-206.

[37] Zadi A S, Alex R, Rong Z, et al. Arterial blood pressure feature estimation using photoplethysmography[J]. Computers in Biology and Medicine, 2018,102(1):104-111.

[38] Mohebbian M R, Dinh A, Wahid K, et al. Blind, cuff-less, calibration-free and continuous blood pressure estimation using optimized inductive group method of data handling[J]. Biomedical Signal Processing and Control, 2020,57(3):101682.

[39] Monika J, Sujay D, AV Subramanyam. Face video based touchless blood pressure and heart rate estimation[C]//2016 IEEE 18th International Workshop on Multimedia Signal Processing (MMSP). 2016.

[40] Makoto Y, Norihiro S, Akira T, et al. Non-contact blood pressure estimation using video pulse waves for ubiquitous health monitoring [C]//2017 IEEE 6th Global Conference on Consumer Electronics (GCCE). IEEE, 2017.

[41] Patil O R, Wang W, Gao Y, et al. A camera-based pulse transit time estimation approach towards non-intrusive blood pressure monitoring[C]// 2019 IEEE International Conference on Healthcare Informatics (ICHI). IEEE, 2019.

[42] Gonzalez V C, Fuentes S, Torrico D D, Dunshea F R. Non-contact heart rate and blood pressure estimations from video analysis and machine learning modelling applied to food sensory responses: a case study for chocolate[J]. Sensors (Basel, Switzerland), 2018,18(6):1802.

[43] Senturk U, Yucedag I, Polat K. Repetitive neural network (RNN) based blood pressure estimation using PPG and ECG signals[C]// 2018 2nd International Symposium on Multidisciplinary Studies and Innovative Technologies (ISMSIT), 2018.

[44] Takahashi R, Ogawa-Ochiai K, Tsumura N. Non-contact method of blood pressure estimation using only facial video[J]. Artificial Life and Robotics, 2020, 25(8):343-350.

[45] Xijian Fan, Qiaolin Ye, Xubing Yang, et al. Robust blood pressure estimation using an RGB camera[J]. Journal of Ambient Intelligence and Humanized Computing, 2020,11(11):4329-4336.

[46] Patil O R, Yang G, Li B, et al. CamBP: a camera-based, non-contact blood pressure monitor[C]// the 2017 ACM International Joint Conference on Pervasive and Ubiquitous Computing and the 2017 ACM International Symposium on Wearable Computers. ACM, 2017:524-529.

[47] Djeldjli D, Bousefsaf F, Maaoui C, et al. Imaging photoplethysmography: signal waveform analysis[C]// 2019 10th IEEE International Conference on Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications (IDAACS). IEEE, 2019:830-834.

[48] Luo Hong, Yang D, Barszczyk A, et al. Smartphone-based blood pressure measurement using transdermal optical imaging technology[J]. Circulation. Cardiovascular imaging, 2019,12(8): 8857.

[49] Yuseok B, Sang-Ki K, Sooyeon K, et al. Face detection based on skin color likelihood[J]. Pattern Recognition: The Journal of the Pattern Recognition Society, 2014,47(4):1573-1585.

[50] Bousefsaf F, Maaoui C, Pruski A. Peripheral vasomotor activity assessment using a continuous wavelet analysis on webcam photoplethysmographic signals[J]. Bio Medical Materials & Engineering, 2016,27(5):527-538.

[51] Sanyal S, Nundy K K. Algorithms for monitoring heart rate and respiratory rate from the video of a user's face[J]. IEEE Journal of Translational Engineering in Health & Medicine, 2018,1(6):1-11.

[52] Ramadhan J M, Khaled M E. A video steganography algorithm based on Kanade-Lucas-Tomasi tracking algorithm and error correcting codes[J]. Multimedia Tools and Applications, 2016,75(17):10311-10333.

[53] Unakafov A M. Pulse rate estimation using imaging photoplethysmography: generic framework and comparison of methods on a publicly available dataset[J]. Biomedical Physics & Engineering Express, 2017,4(4):045001.

[54] Haan D , Gerard, Jeanne, et al. Robust Pulse Rate From Chrominance-Based rPPG.[J]. IEEE Transactions on Biomedical Engineering, 2013,60(10):2878-2886.

[55] Wang W, den Brinker A C, Stuijk S, De Haan G. Algorithmic principles of remote PPG [J]. IEEE Transactions on Biomedical Engineering, 2016,64(7):1479-1491.

[56] Duan W, Zhiyang C, Qiang L, et al. The study on continuous blood pressure information based on the pulse characteristic parameters[J]. Applied Mechanics& Materials, 2013,346(8):103-108.

[57] Takahashi R, Ogawa-Ochiai K, Tsumura N. Non-contact method of blood pressure estimation using only facial video[J]. Artificial Life and Robotics, 2020, 25(3):343-350.

[58] Rong M, Li K. A blood pressure prediction method based on imaging photoplethy-smography in combination with machine learning[J]. Biomedical Signal Processing and Control, 2021,64(2):102328.

[59] Yogarajah P, Condell J, Curran K, et al. A dynamic threshold approach for skin segmentation in color images[J]. International Journal of Biometrics, 2012,4(1):38-55.

[60] Xuesong Niu, Shiguang Shan, Hu Han, Xilin Chen. RhythmNet: End-to-end Heart Rate Estimation from Face via Spatial-temporal Representation[J]. IEEE Transactions on Image Processing, 2020,29(10):2409-2423.

[61] Xuesong Niu, Hu Han, Shiguang Shan, Xilin Chen. VIPL-HR: A Multi-modal Database for Pulse Estimation from Less-constrained Face Video[C] //Asian Conference on Computer Vision, ACCV. 2018:562-576.