Patents
ECG & Biosignal Processing Patents
Seven patents awarded for work at Hewlett-Packard / Agilent Technologies, Andover MA (1980–1995). These algorithms represented the first system in the industry to successfully monitor patients with pacemakers in real-time clinical environments, and remain deployed in intensive care units worldwide.
This body of work is the direct precursor to modern medical ML — built using signal decomposition, feature engineering, and real-time classification under noise before deep learning frameworks existed. The same principles now inform modern biosignal ML approaches.
- US 4,664,116 — Pace Pulse Identification Apparatus
- US 4,832,041 — Pace Pulse Eliminator
- US 4,539,999 — Method and Device for Subtracting a Pacer Signal from an ECG Signal
- US 4,838,278 — Paced QRS Complex Classifier
- US 5,447,164 — Interactive Medical Information Display System and Method for Displaying User-Definable Patient Events
- A61B5/7264 — Classification of Physiological Signals Using Neural Networks, Statistical Classifiers, and Expert Systems
Relevance to Modern Medical ML
The core problems these patents solved — detecting and classifying low-SNR physiological events in real time, handling pacemaker artifact, defining clinically meaningful event taxonomies — remain open research problems in modern form. Work in PCG (heart sound) classification, atrial fibrillation detection, and wearable biosignal processing all draw on the same signal processing foundations.