A 1μw Voice Activity Detector Using Analog Feature Extraction and Digital Deep Neural Network.
2018 IEEE INTERNATIONAL SOLID-STATE CIRCUITS CONFERENCE - (ISSCC)（2018）
Key words
voice UIs,energy-harvesting acoustic sensor nodes,battery-operating devices,voice activity detection,background noise,power gating,higher-level speech tasks,speaker identification,deep neural networks,classification accuracy,AFE variation,chip-wise training,mixed-signal decision tree classifier,ultra-low-power input devices,mobile devices,wearable devices,voice user interfaces,digital deep neural network,analog feature extraction,1μW voice activity detector,10dB SNR speech,1-σ standard deviation,digital filter bank,event-encoding A/D interface,digital BNN classifier,1μW VAD system utilizing AFE,ULP implementations,binarized neural networks,embedded systems,conventional floating-point DNNs,low input SNR,inferior classification accuracy,7-node DT,chip-to-chip basis,DT thresholds,machine-learning based calibration,analog signal processing,power 1.0 muW,power 6.0 muW,power 0.63 muW,noise figure 10.0 dB
AI Read Science
Must-Reading Tree
Example
Generate MRT to find the research sequence of this paper
Chat Paper
Summary is being generated by the instructions you defined