Back

RadiHeart · Measurement technology and clinical evidence

FROM SENSING TO ANALYSIS

From optical sensing to AI. The science of fingertip measurement.

Optical sensing captures fingertip photoplethysmography (PPG) signals. Signal quality control and AI waveform analysis then estimate the arterial blood pressure waveform, showing pressure changes throughout the cardiac cycle.

  1. Optical sensing

    Dual-wavelength optical sensing

    Red and infrared light capture fingertip pulse waves. A stable optical sensor and ambient-light rejection reduce interference from external light sources, preserving waveform detail and variation.

    • Japanese-made optical module
    • Ambient-light suppression
  2. Signal processing

    On-device filtering and quality assessment

    On-device filtering and signal-quality assessment reduce noise and select pulse signals that meet quality thresholds, providing reliable input for subsequent AI analysis.

    • On-device signal processing
    • Signal quality control
  3. AI analysis

    AI blood pressure waveform analysis

    Personalized deep learning models estimate arterial blood pressure waveforms from optical pulse signals and derive 32 cardiovascular indicators, revealing physiological information beyond systolic and diastolic pressure.

    • Personalized deep learning model
    • 32 cardiovascular metrics

CLINICAL VALIDATION

Innovation starts with technology. Trust is built on evidence.

Measurements from 111 participants at Hualien Tzu Chi Hospital were compared with a mercury sphygmomanometer to assess blood pressure measurement performance against ISO 81060-2:2018.

Participants
111 participants
Measurement pairs
333 pairs
Mean error · Systolic / Diastolic
−2.10 / 2.44 mmHg
ISO 81060-2:2018

RadiOxi compared with a mercury sphygmomanometer

Company-provided summary of RadiOxi blood pressure measurement validation. Values are in mmHg.
Assessment criterionThresholdSystolicDiastolicResult
Mean errormmHg±5−2.102.44Met
Standard deviation of errormmHg≤86.205.82Met
Standard deviation of paired differencesmmHg≤6.62 / 6.516.606.35Met

Study registration NCT06104501 ↗