BiomedBench
收藏资源简介:
BiomedBench是由洛桑联邦理工学院嵌入式系统实验室创建的一个生物医学应用基准套件,专为低功耗可穿戴设备设计。该数据集包含8个端到端应用,涵盖了从信号采集到处理的各个阶段,旨在评估和比较不同平台的性能。数据集内容丰富,包括心电图(ECG)、表面肌电图(sEMG)等多种信号处理,适用于生物工程系统和TinyML应用设计。创建过程中,研究团队考虑了各种计算负载和活动与空闲时间的比例,确保数据集能有效代表低功耗可穿戴设备领域的挑战。BiomedBench的应用领域广泛,主要用于提高可穿戴设备的计算能力和能效,增强患者隐私保护,减少响应延迟,并支持小型电池和稀疏充电周期的无缝操作。
BiomedBench is a biomedical application benchmark suite developed by the Embedded Systems Laboratory at École Polytechnique Fédérale de Lausanne (EPFL), tailored specifically for low-power wearable devices. This suite includes 8 end-to-end applications spanning all stages from signal acquisition to processing, aiming to evaluate and compare the performance across various platforms. It boasts diverse content covering multiple signal processing tasks such as electrocardiogram (ECG) and surface electromyogram (sEMG), and is suitable for the design of bioengineering systems and TinyML applications. During its development, the research team considered various computational loads as well as the ratio of active to idle time, ensuring that the benchmark suite effectively represents the challenges in the low-power wearable device domain. BiomedBench has a wide range of application scenarios, and is primarily used to improve the computing capability and energy efficiency of wearable devices, enhance patient privacy protection, reduce response latency, and support seamless operation with small batteries and infrequent charging cycles.

- 1BiomedBench: A benchmark suite of TinyML biomedical applications for low-power wearables洛桑联邦理工学院嵌入式系统实验室 · 2024年



