Non-contact heart vibration measurement using computer vision-based seismocardiography
Executive Summary
This study introduces a novel method for non-contact seismocardiography (SCG) using smartphone cameras and computer vision techniques, specifically the Lucas–Kanade template tracking algorithm. The method was validated against gold-standard accelerometer-based SCG signals in 14 healthy subjects, showing high dynamic time warping (DTW) similarity indices (0.94–0.95) and good agreement in heart rate estimation (bias = 0.649 bpm). This approach demonstrates potential for low-cost, accessible remote cardiac monitoring.
Answer Machine Insights
Q: How accurate is the vision-based SCG compared to the gold-standard accelerometer-based SCG?
The DTW-based similarity index was 0.94–0.95, indicating high accuracy.
The average DTW-based similarity index between the signals was 0.94 and 0.95 in the right-to-left and head-to-foot directions, respectively.
Q: What is the agreement between heart rate estimated from SCGv and ECG?
The bias in heart rate estimation was 0.649 bpm.
The findings indicated a good agreement between the estimated heart rate values and the gold-standard measurements (bias = 0.649 beats/min).
Key Results
The DTW-based similarity index between vision-based SCG (SCGv) and gold-standard SCG (SCGg) was 0.94–0.95.
Heart rate estimation from SCGv showed a bias of 0.649 bpm compared to ECG-derived heart rate.
Visual Evidence
Clinical Snapshot
Evidence Rating
Relevance
high Priority