近些年來，由於運動風氣的普及，越來越多人用穿戴式裝置在運動時同時監測心率，然而配戴穿戴式裝置需與肌膚進行接觸，長時間的使用可能造成不適。考量到非接觸式量測技術的優點，我們希望將該技術應用到運動器材場域，使得使用者可以在更舒適、便利且安全的情況下達到運動監測的目的。在研究目前基於影像非接觸式心率量測技術應用於運動器材場域的相關論文後，我們相信仍存在部分問題有待改善，其主要問題點整理如下：
(1) 在運動狀態啟動演算法，往往會因為沒有過往的心率參考值而造成估計誤差。
(2) 在靜態和動態運動轉換時，容易造成心率估計誤差。
基於上述問題，本論文基於電腦和攝影機開發了一應用於實際運動場域的高準確性非接觸式即時心率量測系統並提出全新的心率估計演算法來改善上述問題。在腳踏車、踏步機、跑步機、初始於動態狀態之運動和靜動態快速轉換運動五項實驗項目中10個人的測試集平均絕對誤差(Mean Absolute Error，MAE)/均方根誤差(Root-Mean-Square Error，RMSE)分別為1.62/2.58、1.67/2.78、2.09/3.67、2.68/5.05、2.42/3.08 (BPM)，而SR_5/SR_10分別為0.96/0.99、0.95/0.98、0.89/0.96、0.85/0.94、0.86/0.95。此實驗結果顯示本論文所提出之方法除能有效改善上述問題並進一步提升心率量測準確度。

Regular exercise and fitness training are more emphasized in modern days, and thus monitoring our heart rate during exercise with wearable devices had become more and more common. However, these products may cause discomfort or even skin allergies while sweating during exercising due to the contact with the users. Considering the increasing development of non-contact physiological signal measurement technology based on remote photoplethysmography (remote-PPG, rPPG), we wish to apply such technology on exercise and fitness training to help users exercise in a more comfortable, convenient, and safe condition while remain monitoring their physiological signal. After studying current camera-based rPPG methods which were applied to the sports field, we believe that there are still some problems that need to be solved, and the main issues are listed as follows: 1) If the subject exercises from dynamic state, most algorithms might fail to find the correct peaks and cause estimation error due to the lack of previous estimations. 2) Most research have relatively larger estimation error during exercise state conversion such as from steady state to exercise state.
The contribution of this paper is to develop a non-contact real-time heart rate measurement system with high accuracy and low preparation time based on the computer and camera matched with sports equipment while also proposing a new pulse rate calculation algorithm to solve the problems above. The mean absolute error and root mean square error (MAE/RMSE) of the 10 subjects test dataset in the five experimental items (bicycle, stepper, treadmill, exercise from dynamic state, and static and dynamic fast conversion) are 1.62/2.58, 1.67/2.78, 2.09/3.67, 2.68/5.05 and 2.42/3.08 (BPM); and SR_5/SR_10 are 0.96/0.99, 0.95/0.98, 0.89/0.96, 0.85/0.94 and 0.86/0.95. The experimental results show that our method can solve the problems above and improve the accuracy of pulse rate measurement as well.

[1] 台灣趨勢股份有限公司, "TTR 台灣趨勢研究報告, 產業分析:運動服務業發展趨勢," 2018. [Online]. Available: http://www.twtrend.com/upload/shares/a_15299856880.pdf.
[2] J. H. Wilmore, D. L. Costill, and W. L. Kenney, Physiology of Sport and Exercise. Human Kinetics, 1994.
[3] F. P. Wieringa, F. Mastik, and A. F. van der Steen, "Contactless Multiple Wavelength Photoplethysmographic Imaging: A First Step Toward “SpO2 Camera” Technology," Annual Review of Biomedical Engineering, vol. 33, no. 8, pp. 1034-41, 2005.
[4] K. Humphreys, T. Ward, and C. Markham, "Noncontact Simultaneous Dual Wavelength Photoplethysmography: A Further Step Toward Noncontact Pulse Oximetry," Review of Scientific Instruments, vol. 78, no. 4, p. 044304, 2007.
[5] W. Verkruysse, L. O. Svaasand, and J. S. Nelson, "Remote Plethysmographic Imaging Using Ambient Light," Opt. Express, vol. 16, no. 26, pp. 21434-45, 2008.
[6] L. A. Aarts et al., "Non-Contact Heart Rate Monitoring Utilizing Camera Photoplethysmography in the Neonatal Intensive Care Unit - A Pilot Study," Early Human Development, vol. 89, no. 12, pp. 943-8, 2013.
[7] G. A. Ramirez, O. Fuentes, S. L. Crites, M. Jimenez, and J. Ordonez, "Color Analysis of Facial Skin: Detection of Emotional State," in Proceedings of IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), pp. 474-479, 2014.
[8] Y. C. Tsai, P. W. Lai, P. W. Huang, T. M. Lin, and B. F. Wu, "Vision-Based Instant Measurement System for Driver Fatigue Monitoring," IEEE Access, vol. 8, pp. 67342-67353, 2020.
[9] G. de Haan and V. Jeanne, "Robust Pulse Rate from Chrominance-Based rPPG," IEEE Transactions on Biomedical Engineering, vol. 60, no. 10, pp. 2878-2886, 2013.
[10] G. Balakrishnan, F. Durand, and J. Guttag, "Detecting Pulse from Head Motions in Video," in Proceedings of IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 3430-3437, 2013.
[11] G. de Haan and A. J. van Leest, "Improved Motion Robustness of Remote-PPG by Using the Blood Volume Pulse Signature," Physiological measurement, vol. 35, no. 9, pp. 1913-1926, 2014.
[12] R. Y. Huang and L. R. Dung, "A Motion-Robust Contactless Photoplethysmography Using Chrominance and Adaptive Filtering," in Proceedings of IEEE Biomedical Circuits and Systems Conference (BioCAS), pp. 1-4, 2015.
[13] Z. Wu and N. E. Huang, "Ensemble Empirical Mode Decomposition: A Noise-Assisted Data Analysis Method," Advances in Adaptive Data Analysis, vol. 1, no. 01, pp. 1-41, 2009.
[14] K. Lin, D. Chen, and W. Tsai, "Face-Based Heart Rate Signal Decomposition and Evaluation Using Multiple Linear Regression," IEEE Sensors Journal, vol. 16, no. 5, pp. 1351-1360, 2016.
[15] W. Wang, A. C. d. Brinker, S. Stuijk, and G. d. Haan, "Algorithmic Principles of Remote PPG," IEEE Transactions on Biomedical Engineering, vol. 64, no. 7, pp. 1479-1491, 2017.
[16] B. F. Wu et al., "Motion Resistant Image-Photoplethysmography Based on Spectral Peak Tracking Algorithm," IEEE Access, vol. 6, pp. 21621-21634, 2018.
[17] Q. Zhu, C. W. Wong, C. H. Fu, and M. Wu, "Fitness Heart Rate Measurement Using Face Videos," in Proceedings of IEEE International Conference on Image Processing (ICIP), 2017.
[18] Y. C. Lin and Y. H. Lin, "Step Count and Pulse Rate Detection Based on the Contactless Image Measurement Method," IEEE Transactions on Multimedia, vol. 20, no. 8, pp. 2223-2231, 2018.
[19] K. Xie, C. H. Fu, H. Liang, H. Hong, and X. Zhu, "Non-Contact Heart Rate Monitoring for Intensive Exercise Based on Singular Spectrum Analysis," in Proceedings of IEEE Conference on Multimedia Informaton Processing and Retrieval (MIPR), pp. 228-233, 2019.
[20] R. Spetlik, V. Franc, J. Cech, and J. Matas, "Visual Heart Rate Estimation with Convolutional Neural Network," in Proceedings of British Machine Vision Conference, pp. 1-12, 2018.
[21] W. Sun, H. Wei, and X. Li, "No-Contact Heart Rate Monitoring Based on Channel Attention Convolution Model," in Proceedings of SPIE International Conference on Graphics and Image Processing (ICGIP), vol. 11373, pp. 113732T-1-7, 2019.
[22] O. Perepelkina, M. Artemyev, M. Churikova, and M. Grinenko, "HeartTrack: Convolutional Neural Network for Remote Video-Based Heart Rate Monitoring," in Proceedings of IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 288-289, 2020.
[23] M. Sabokrou, M. Pourreza, X. Li, M. Fathy, and G. Zhao, "Deep-HR: Fast Heart Rate Estimation from Face Video Under Realistic Conditions," in Expert Systems with Applications, vol. 186, pp. 115596, 2021.
[24] Y. C. Chou, B. Y. Ye, H. R. Chen and Y. H. Lin, "A Real-Time and Non-contact Pulse Rate Measurement System on Fitness Equipment," in IEEE Transactions on Instrumentation and Measurement, vol. 71, pp. 1-11, 2022.
[25] 葉柏毅, "進階的非接觸式心率量測系統在運動器材的應用," 國立臺灣科技大學碩士論文, 2020.
[26] Dlib, "Dlib C++ library." [Online]. Available: https://github.com/davisking/dlib.
[27] Z. Zhang, "Photoplethysmography-Based Heart Rate Monitoring in Physical Activities via Joint Sparse Spectrum Reconstruction," in IEEE Transactions on Biomedical Engineering, vol. 62, no. 8, pp. 1902-1910, Aug. 2015.
[28] "Polar H7 heart rate sensor." [Online]. Available: https://www.polar.com/en/products/accessories/H7_heart_rate_sensor.
[29] "Vernier Go Direct EKG sensor." [Online]. Available: https://www.vernier.com/product/go-direct-ekg-sensor/.