簡易檢索 / 詳目顯示

研究生: 劉明烜
LIU, MING-HSUAN
論文名稱: 具備個人化校正機制之智慧型手機rPPG生理資訊量測
Smartphone rPPG Vital Sign Measurement with Personalized Calibration Mechanism
指導教授: 侯廷偉
Hou, Ting-Wei
學位類別: 碩士
Master
系所名稱: 工學院 - 工程科學系
Department of Engineering Science
論文出版年: 2026
畢業學年度: 114
語文別: 中文
論文頁數: 73
中文關鍵詞: 遠端光體積描記智慧型手機非接觸生理量測血壓推算個人化校正血氧飽和度呼吸率
外文關鍵詞: remote photoplethysmography, smartphone, contactless vital signs, cuff-less blood pressure, personalized calibration, oxygen saturation, respiratory rate
相關次數: 點閱:10下載:0
分享至:
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報
  • 本論文以智慧型手機為平台,探討一套著重於系統穩定性與工程可行性之rPPG (Remote Photoplethysmography)生理資訊量測架構。研究重點不在於單一演算法之創新,而是從系統層級出發,整合品質導向之訊號處理概念與個人化校正架構,分析多階段品質控管與校正策略對量測穩定性之影響,並評估其於實務量測情境中的可行性。
    實驗結果顯示,導入品質導向設計概念與個人化校正機制後,整體量測結果在穩定性與一致性方面呈現改善趨勢,特別是在訊號品質變動較大的量測情境中更為明顯。相關結果顯示,系統層級之設計取捨與整合策略,對於rPPG 技術由研究方法走向實際裝置實作具有關鍵影響。
    本研究聚焦於工程層級之方法探討與可行性分析,未以臨床診斷或醫療決策為研究目標。所提出之架構與觀察結果,可作為未來發展具實務可行性之rPPG 生理量測系統之參考。

    This thesis investigates a smartphone-based remote photoplethysmography (rPPG) framework with an emphasis on system robustness and engineering feasibility. By integrating quality-oriented signal processing concepts and a personalized calibration framework, the proposed approach aims to improve measurement stability under unconstrained conditions. Experimental results demonstrate improved consistency and robustness of rPPG-based measurements. This work focuses on system-level analysis rather than clinical validation, and provides insights for practical deployment of rPPG on consumer devices.

    摘要 I Extended Abstract II 目錄 IX 表目錄 XII 圖目錄 XIII 第一章 緒論 1 1.1 研究背景與動機 1 1.2 研究目的 2 1.3 研究貢獻 2 1.4 論文架構 3 第二章 文獻探討 4 2.1 Photoplethysmography (PPG) 原理與波形特徵 4 2.1.1 研究挑戰與背景 4 2.1.2 PPG 量測原理與光學配置 4 2.1.3 波形形態特徵及其生理意涵 5 2.2 Remote PPG(rPPG)基礎:臉部流程與訊號生成 7 2.2.1 臉部 ROI 取樣與追蹤(multi-ROI / patches) 9 2.2.2 rPPG 訊號生成方法(POS / 投影類) 9 2.2.3 手機端干擾來源與品質控管需求 10 2.3 rPPG 心率(HR)估算方法 12 2.4 rPPG 呼吸率(RR)估算方法 13 2.5 相機式 SpO₂ 估算:ratio-of-ratios 與手機端限制 13 2.6 無袖帶(cuffless)BP估算流程:形態特徵與個體差異 15 2.7 個人化校正與驗證(AHA / Mukkamala 核心精神) 16 2.8 小結 19 第三章 研究方法 21 3.1 系統總覽與資料流程 21 3.2 影像擷取、倒數啟動與 ROI 取樣 23 3.2.1 影像擷取與時間標記建立 23 3.2.2 量測前準備階段與相機穩定化設計 23 3.2.3 Armed 狀態與量測起點定義 24 3.2.4 多 ROI 取樣與 RGB 時序訊號建立 24 3.3 rPPG 訊號生成流程與多路架構設計 25 3.3.1 共同前處理(Pre-processing) 25 3.3.2 Signal formation(POS/投影類方法) 25 3.3.3 多路解耦架構(Multi-track decoupling) 26 3.4 品質指標、Quality Gating 與融合策略 26 3.4.1 品質指標(Quality metrics) 26 3.4.2 Top-K ROI 選擇與 track 融合(概述) 27 3.4.3基於品質權重之保守融合 27 3.5 心率、呼吸率與血氧飽和度估算(HR / RR / SpO₂ Vital Signs Inference)估算流程 29 3.5.1 心率估算(Heart Rate, HR) 29 3.5.2 呼吸率估算(Respiratory Rate, RR) 30 3.5.3 血氧飽和度估算(SpO₂) 31 3.6 BP 形態特徵擷取與 raw BP估算(BP Morphology & Raw Estimation) 34 3.6.1 量測前段處理與避免形態失真之訊號處理設計 34 3.6.2 morphology 特徵點與特徵定義 34 3.6.3 raw BP estimate 的定位 35 3.7 Reference BP protocol 與 raw/ref pairing 35 3.7.1 兩階段量測設計(Pre-cal / Post-cal) 35 3.7.2 Δt 對齊門檻(固定規格) 36 3.7.3 單次 reference 的合理性與風險控管 36 3.8 條件式個人化校正(Conditional Calibration)與安全機制 37 3.8.1 校正點與模式切換 37 3.8.2 固定安全規格(slope clamping / physiological bounds) 38 3.9 Logging / Export 與可追溯性(Traceability-ready Output) 39 第四章 實驗設計與結果分析 43 4.1 研究設計概述 43 4.2 實驗設備與受試者 43 4.2.1 實驗設備 43 4.2.2 受試者與量測條件 44 4.3 量測流程(Pre-cal / Reference / Post-cal) 45 4.3.1 Smartphone rPPG 單次量測流程 45 4.3.2 Reference BP 量測 45 4.3.3 raw/ref pairing 與個人化校正流程 45 4.4 時間對齊、納入規則與品質控管 46 4.4.1 Δt對齊門檻(fixed spec) 46 4.5 評估指標 47 4.5.1 MAE 與 RMSE 47 4.5.2 Bland–Altman 47 4.6 結果 47 4.6.1 校正前後 BP 誤差改善(主要結果) 48 4.6.2 以受試者為單位之 raw/ref pairing 與誤差分布 49 4.7 討論 51 4.8 本章小結 52 第五章 結論與未來展望 53 5.1 結論 53 5.2 未來工作與展望 55 參考文獻 56

    [1] J. B. Cohen, R. L. Byfield, S. T. Hardy, et al., “Cuffless Devices for the Measurement of Blood Pressure: A Scientific Statement from the American Heart Association,” Hypertension, vol. 82, pp. e00–e00, 2025. doi:10.1161/HYP.0000000000000254.
    [2] R. Mukkamala, M. Yavarimanesh, K. Natarajan, et al., “Evaluation of the Accuracy of Cuffless Blood Pressure Measurement Devices: Challenges and Proposals,” Hypertension, vol. 78, no. 5, pp. 1161–1167, 2021. doi:10.1161/HYPERTENSIONAHA.121.17747.
    [3] R. J. Lee, S. Sivakumar, and K. H. Lim, “Review on Remote Heart Rate Measurements Using Photoplethysmography, ”Multimedia Tools and Applications, vol. 83, pp. 44699–44728, 2024. doi:10.1007/s11042-023 16794-9.
    [4] W. Chen, Z. Yi, L. J. R. Lim, et al., “Deep Learning and Remote Photoplethysmography Powered Advancements in Contactless Physiological Measurement,” Frontiers in Bioengineering and Biotechnology, vol. 12, art. 1420100, 2024. doi:10.3389/fbioe.2024.1420100.
    [5] C. G. Scully, J. Lee, J. Meyer, et al., “Physiological Parameter Monitoring from Optical Recordings With a Mobile Phone, ” IEEE Transactions on Biomedical Engineering, vol. 59, no. 2, pp. 303–306, Feb. 2012.doi:10.1109/TBME.2011.2163157.
    [6] M. Elgendi,“On the Analysis of Fingertip Photoplethysmogram Signals,”Current Cardiology Reviews, vol. 8, no. 1, pp. 14–25, 2012.
    [7] A. Lam and Y. Kuno, "Robust Heart Rate Measurement from Video Using Select Random Patches," in Proceedings of the IEEE International Conference on Computer Vision (ICCV), 2015, pp. 3640–3648.
    [8] S. Tulyakov, X. Alameda-Pineda, E. Ricci, et al. Self-Adaptive Matrix Completion for Heart Rate Estimation from Face Videos under Realistic Conditions, Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 2396–2404, 2016.
    [9] W. Karlen, J. M. Ansermino, and G. A. Dumont, “Multiparameter Respiratory Rate Estimation from the Photoplethysmogram, IEEE Transactions on Biomedical Engineering, vol. 60, no. 7, pp. 1946–1953, 2013.
    [10] J. S. Lee, G. Hwang, M. Ryu, and S. J. Lee, “LSTC-rPPG: Long Short-Term Convolutional Network for Remote Photoplethysmography,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 2023.
    [11] E. Allado, M. Poussel, J. Renno, et al., “Remote Photoplethysmography Is an Accurate Method to Remotely Measure Respiratory Rate: A Hospital-Based Trial, ”Journal of Clinical Medicine, vol. 11, art. 3647, 2022.doi:10.3390/jcm11133647.
    [12] M. van Gastel and W. Verkruysse,“Contactless SpO₂ with an RGB camera: Experimental Proof of Calibrated SpO₂,”Biomedical Optics Express, vol. 13, no. 12, pp. 6791–6802, 2022.doi:10.1364/BOE.471332.
    [13] A. M. Cabanas, N. Sáez, P. O. Collao-Caiconte, et al., “Evaluating AI Methods for Pulse Oximetry: Performance, Clinical Accuracy, and Comprehensive Bias Analysis, ”Bioengineering, vol. 11, art. 1061, 2024. doi:10.3390/bioengineering11111061.
    [14] S. Yang, Y. Zhang, S.-Y. Cho, R. Correia, and S. P. Morgan, “Non-invasive Cuff-less Blood Pressure Estimation Using a Hybrid Deep Learning Model,” Optical and Quantum Electronics, vol. 53, art. 93, 2021. doi:10.1007/s11082-020-02667-0.
    [15] 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.
    [16] International Medical Device Regulators Forum (IMDRF) SaMD Working Group, “Software as a Medical Device (SaMD): Application of Quality Management System,” IMDRF/SaMD WG/N23FINAL, 2015.

    QR CODE