Contactless Cardiovascular Monitoring Using Neck-Based Video Analysis

Background

Continuous cardiovascular monitoring is essential for the early detection and management of cardiac conditions, driving a growing demand for accessible diagnostic tools in both clinical and remote telemedicine settings. Traditional monitoring methods, such as electrocardiography and fingertip photoplethysmography, require physical sensors attached to the body, which can cause patient discomfort, restrict mobility, and limit long-term compliance. Consequently, the field of non-invasive diagnostics has increasingly turned toward remote photoplethysmography (rPPG), a contactless technology that utilizes standard optical cameras to detect subtle, blood-volume-induced color changes in the skin to estimate vital metrics like heart rate and heart rate variability.

However, current rPPG approaches predominantly rely on analyzing video recordings of the patient’s face, which introduces significant biometric privacy risks by exposing identifiable facial features. Furthermore, face-based methods often suffer from spatial signal variability and decreased reliability across diverse skin tones. Shifting the focus to alternative anatomical regions, such as the neck, avoids facial exposure but presents severe technical hurdles due to high spatial heterogeneity. In these non-facial areas, desired arterial pulsations frequently overlap with competing venous signals, while routine physiological movements like swallowing or sudden shifts introduce substantial motion artifacts that degrade signal quality and diagnostic accuracy.

Technology

Developed at Lehigh University, this software-based diagnostic technology enables contactless cardiovascular monitoring by analyzing video recordings of the anterior neck to extract remote photoplethysmogram (rPPG) signals. Utilizing standard hardware like smartphone cameras, the system defines a standardized Region of Interest using anatomical landmarks and tracks it across video frames. This region is partitioned into a 200-cell spatial grid to extract mean green channel intensities. The signals then undergo a rigorous preprocessing pipeline involving bandpass filtering, artifact rejection for swallowing or motion, two-stage peak detection, and outlier rejection. Finally, the framework utilizes reference-guided or autonomous region selection strategies to identify vascular “hotspots” and generate a high signal-to-noise waveform, accurately calculating Heart Rate and Heart Rate Variability metrics.

This technology is uniquely differentiated by its privacy-preserving, neck-focused approach, which eliminates the biometric privacy risks associated with traditional face-based rPPG systems. Furthermore, its granular grid-based architecture and adaptive selection algorithms effectively overcome the complex spatial heterogeneity of the neck, successfully isolating arterial signals from overlapping venous signals and motion artifacts. Capable of operating completely autonomously without external reference sensors, the solution demonstrates exceptional accuracy and remains robust across diverse skin phototypes, making it ideal for accessible telemedicine.

Benefits

• Preserves patient biometric privacy by analyzing the neck rather than the face
• Enables contactless and non-invasive monitoring of heart rate and heart rate variability
• Operates on standard consumer hardware like smartphone cameras for accessible telemedicine and home use
• Automatically compensates for subject movement and swallowing artifacts to ensure high signal accuracy
• Provides reliable cardiovascular measurements across diverse skin tones
• Functions autonomously without requiring external reference sensors or wearable devices

Applications

• Telemedicine cardiac diagnostics
• Privacy-preserving remote patient monitoring
• Smartphone-based consumer heart monitoring
• Contactless clinical vital sign screening
• In-hospital continuous patient monitoring

Additional info

Dr. Amirtahà Taebi is an Assistant Professor with the Department of Bioengineering at Lehigh University. His work focuses on developing AI-powered sensing technologies, wearable and contactless monitoring systems, and digital twin modeling to advance noninvasive cardiovascular diagnostics. His research has been supported by local, state, and federal agencies, including the NSF, NIH, and Mississippi Institutions of Higher Learning.