Signal Enhancement Applied to Pulse Transit Time Measurement

High arterial blood pressure, or hypertension, is a major risk factor for cardiovascular diseases. Conventional noninvasive methods for estimating the arterial blood pressure rely on a brachial cuff that is placed around the upper arm. The cuff is inflated to a pressure that exceeds the systolic blood pressure. The heart pulse is monitored while the cuff is slowly deflated. The pressure at which the heart pulse sound can be detected corresponds to the systolic pressure and the pressure at which the heart pulse sound can no longer be detected corresponds to the diastolic pressure. This approach cannot be used on a continuous basis and has many disadvantages, including the fact that the system is cumbersome and causes discomfort to the patient, which may affect their blood pressure value. This thesis proposes two approaches to accurately measure the pulse transit time (PTT). The PTT can be used to estimate arterial blood pressure variations noninvasively, on a continuous basis without the use of a cuff. The correlation between the arterial blood pressure and PTT are verified using bio-signals from the MIMIC II database. The first approach for measuring the PTT relies on the electrocardiogram (ECG) and photo-plethysmograph (PPG) signals. Algorithms based on the empirical mode decomposition (EMD) method and the adaptive filtering techniques are proposed to enhance the corrupted ECG signal. Gaussian-based curve fitting algorithms are proposed to model and extract the embedded characteristic parameters from the original PPG signal. The second approach relies on a novel framework: the Eulerian video magnification (EVM). The color component of the video signal is amplified to reveal subtle change in the skin color. These changes correspond to the blood flow into the arteries. Simulations and experiments are conducted to validate the proposed framework. The results demonstrate that the enhanced ECG and PPG signals can improve the accuracy of the PTT in estimating the arterial blood pressure. The EVM method can capture the characteristics of the wrist pulse signal and detect the heart pulse on different areas of the human skin. This paves the way towards the usage of cameras for continuous arterial blood pressure monitoring.