Frontiers in Neuroscience · 2021 · 11 citations · 44 references
EngineeringBiomedical EngineeringElectrical EngineeringEdge ComputingElectrophysiologyNeural EngineeringComputational ElectrophysiologyDeep LearningEdge DevicesMemristive MaterialsNeural Signal ProcessingSpiking Neural NetworksAccuracyArrhythmia DetectionArrhythmia RecognitionEmbedded Hardware Co-designEnergy-efficient ComputationMemristive MemoryNeuromorphic HardwarePower Consumption
In this work, a memristive spike-based computing in memory (CIM) system with adaptive neuron (MSPAN) is proposed to realize energy-efficient remote arrhythmia detection with high accuracy in edge devices by software and hardware co-design. A multi-layer deep integrative spiking neural network (DiSNN) is first designed with an accuracy of 93.6% in 4-class ECG classification tasks. Then a memristor-based CIM architecture and the corresponding mapping method are proposed to deploy the DiSNN. By evaluation, the overall system achieves an accuracy of over 92.25% on the MIT-BIH dataset while the area is 3.438 mm 2 and the power consumption is 0.178 μJ per heartbeat at a clock frequency of 500 MHz. These results reveal that the proposed MSPAN system is promising for arrhythmia detection in edge devices.
44
Dmitri B. Strukov, Gregory S. Snider, Duncan R. Stewart et al. · Nature · 2008 · 11.5K citations
Engineering, Electrical Engineering, Electronic Materials +5
Memristors with diffusive dynamics as synaptic emulators for neuromorphic computing
Zhongrui Wang, Saumil Joshi, Sergey Savel’ev et al. · Nature Materials · 2016 · 2.3K citations · Full text
Zachi I. Attia, Peter A. Noseworthy, Francisco López-Jiménez et al. · The Lancet · 2019 · 1.5K citations