A granular resampling method based energy-efficient architecture for heartbeat classification in ECG

Abstract

SVM-based granular resampling method is put forward to obtain a robust classification model for energy-efficient ECG systems. The classification model consists of a low-complexity model to filter most easy-to-learn heartbeats and a high-accuracy classifier to identify the remained heartbeats. Energy-efficient hardware architecture for multi-class heartbeat classification is implemented based on the classification model. The architecture optimizations include memory segmentation to reduce energy consumption and time domain reuse to save resources. We adopt 40-nm CMOS process to implement the proposed design. It provides an average prediction speedup by 57.21% and a significant energy dissipation reduction by 52.22% per classification compared with the design without low-complexity models.