A 65nm Uncertainty-Quantifiable Ventricular Arrhythmia Detection Engine with 1.75µJ per Inference

📋 论文概要

该论文提出了一款基于65nm CMOS工艺的不确定性可量化心室心律失常检测引擎，用于植入式心脏复律除颤器（ICD）中，以降低不适当电击率并提高可靠性。通过集成深度学习与不确定性量化（UQ）方法，该引擎在每次推理仅消耗1.75µJ能量的同时，实现了对室性心动过速和室颤的高精度检测。

💡 主要创新点

• 首次在硬件层面实现了不确定性量化（UQ）功能，增强了深度学习模型在医疗决策中的透明度和可靠性。
• 采用超低功耗设计，每次推理仅需1.75µJ能量，适合植入式设备长期运行。
• 在65nm工艺节点上集成了完整的检测与量化引擎，实现了高能效与高精度的平衡。

🏷 关键词

📄 原文摘要

for preventing Sudden Cardiac Death (SCD) by identifying life-threatening heart rhythms, such as ventricular tachycardia (VT) and ventricular fibrillation (VF) [1], and enabling timely intervention via implantable cardioverter defibrillators (ICD). Although deep-learning (DL) methods have improved VA detection by reducing Inappropriate Shock Rates (ISR) and minimizing manual parameter tuning compared to traditional rule-based systems [2], they suffer from a lack of transparency, particularly in uncertainty quantification (UQ), which limits their reliability in critical medical decisions and impedes widespread adoption in trustworthy smart health applications. Bayesian neural networks (BNNs) address this issue by providing UQ through model sampling, allowing for robust diagnostic interventions, such as from medical experts or

👥 作者与机构

Jianbo Liu, Zephan Enciso, Boyang Cheng, Likai Pei, Steven Davis, Yifan Qin,

Zhenge Jia, Xiaobo Sharon Hu, Yiyu Shi, Ningyuan Cao University of Notre Dame, Notre Dame, IN Detecting Ventricular Arrhythmia (VA) is critical