Microvibrometric Blood Pressure Estimation via Korotkoff Sound–Coupled Pulses Using a Semiconductor Piezoresistive Sensor

Accurate blood pressure (BP) monitoring traditionally relies on acoustic auscultation, but the mechanical microvibrations underlying Korotkoff sound (K-sound) generation remain insufficiently characterized. This study proposes a microvibrometric sensing strategy using a high-sensitivity semiconductor piezoresistive sensor (K-sensor) to capture pulse-resolved mechanical signatures during cuff deflation, circumventing the limitations of conventional air-conducted acoustic detection. Through expert-consensus annotation with simultaneous acoustic references, a coupling relationship between microvibration morphology and K-sound occurrence was established. A convolutional neural network (CNN) was implemented to automate the identification of K-sound–coupled (cK) pulses from microvibration signals. Across seven independent train–test splits in 49 healthy participants, the model achieved a recall of 93.1% ± 4.3% and a balanced accuracy of 95.0% ± 2.0%, with mean biases of 0.30 ± 2.25 mmHg for systolic BP (SBP) and −0.14 ± 2.14 mmHg for diastolic BP (DBP). Beyond classification performance, two distinct morphological archetypes—characterized as notch-type and shoulder-type waveforms—were consistently observed among cK pulses, reflecting differentiated patterns of arterial wall dynamics associated with K-sounds. These findings support microvibration sensing as a physiologically grounded and sensor-centric framework for automated noninvasive cardiovascular assessment beyond conventional acoustic paradigms.