Paper Info
Title
An 8-Element Frequency-Selective Acoustic Beamformer and Bitstream Feature Extractor with 60 Mel-Frequency Energy Features Enabling 95% Speech Recognition Accuracy
Abstract
A synergistic approach to beamforming and feature extraction, reduces processing complexity and die area, and delivers the high SNR required for reliable speech recognition. The 1.1mm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> IC combines frequency-selective bitstream beamforming, bitstream Mel frequency-band feature extraction, and an array of continuous-time sigma-delta modulators (SDMs) without area/power-intensive decimation. When coupled with a DNN, the prototype achieves 95.3% accuracy in recognizing spoken words from the Tensorflow dataset.
Year
DOI
Venue
2020
10.1109/VLSICircuits18222.2020.9162783
2020 IEEE Symposium on VLSI Circuits
Keywords
DocType
ISSN
8-element frequency-selective acoustic beamformer,bitstream feature extractor,speech recognition accuracy,synergistic approach,reliable speech recognition,frequency-selective bitstream beamforming,Mel frequency-band feature extraction,mel-frequency energy features
Conference
2158-5601
ISBN
Citations 
PageRank 
978-1-7281-9943-6
0
0.34
References 
Authors
0
6
Name
Order
Citations
PageRank
Seungjong Lee100.68
Taewook Kang201.69
Bell, J.T.3192.90
Mohammad R. Haghighat400.34
Alberto J. Martinez500.34
Michael P. Flynn642167.71