CardiacWave: A mmWave-based Scheme of Non-Contact and High-Definition Heart Activity Computing
summary
A mmWave system claims non-contact, ECG-like heart monitoring by sensing the cardiac electromagnetic field via the Cardiac-mmWave scattering effect (CaSE) and extracting frequency shifts with learnable mask filters on a denoised response matrix. A neural profiler embeds heart domain knowledge using LSTM with attention and trains on a fidelity loss defined by Discrete Wavelet Transform coefficient distance. Evaluation with 40 subjects and three-channel ECG ground truth shows P/T/QRS timing errors of 0.67%, 0.71%, 0.49% and mean event delay 5.3 ms.
AI-written summary · about
related talks
06mHeartQuake: Accurate Low-Cost Non-Invasive ECG Monitoring Using Bed-Mounted Geophones
01h 13mHealth Monitoring with Wireless Sensors and Machine Learning
46mAnalyzing the ElectroCardioGram (ECG)
28mHow Smart is the ECG? Lessons from Advanced ECG Analytics
06mBodyCompass: Monitoring Sleep Posture with Wireless Signals
26mTed Willke, Senior Principal Engineer, Intel Labs
11mConnected brains
48mRecoverMonitor: Edge-enabled Open-source Wearable Devices for Monitoring Healthcare Remote Patient Recovery with Precision and Efficiency