Energy-Efficient AI at the edge for Biomedical Applications

Abstract

This paper presents a AI-on-the-edge System-on-Chip (SoC) for biomedical applications. For ambulatory tracking and effective treatment of neurological disorders such as seizure and epilepsy, long-term monitoring wearable SoCs is essential to 'close the loop'. To satisfy the wearable form factor, the design challenges at techniques of feature extraction and classification to improve seizure detection accuracy at the Digital Back-End (DBE) must be addressed at a system perspective. Furthermore, future trends of the epilepsy tracking system are discussed.