Implantable cardioverter defibrillators (ICDs) detect and treat dangerous cardiac arrhythmia. This paper describes a low-power VLSI neural network chip which acts as an intracardiac tachycardia classification system. A robust neural network reduces the impact of noise, drift and offsets that are inherent in analog applications. Our system uses a Kohonen self-organizing map, and has a typical power consumption of a few milliwatts
Published in:
Computer-Based Medical Systems, 1998. Proceedings. 11th IEEE Symposium on
Date of Conference: 12-14 Jun 1998