AIRoV ? The First Austrian Symposium on AI, Robotics, and Vision
Sprache des Tagungstitel:
Englisch
Original Kurzfassung:
The problem of QRS detection in ECG signals addresses the identification of heart beats in electrocardiogram signals. This pattern detection problem for bio-signals is taken as a use case to investigate the feasibility and effectiveness of spike-based encoding of time-varying signals by means of threshold-based sampling. We compare the state-of-the-art detection algorithm based on equidistant sampling with threshold-based approaches, comprising signal processing techniques and spiking neural networks (SNN). In contrast to equidistant sampling, this approach allows a reduction of the input data stream by a factor of 10 on average while maintaining comparable detection performance.
Forschungs-