Deep Long Short-Term Memory Adaptive Beamforming Networks For Multichannel Robust Speech Recognition

11/21/2017
by   Zhong Meng, et al.
0

Far-field speech recognition in noisy and reverberant conditions remains a challenging problem despite recent deep learning breakthroughs. This problem is commonly addressed by acquiring a speech signal from multiple microphones and performing beamforming over them. In this paper, we propose to use a recurrent neural network with long short-term memory (LSTM) architecture to adaptively estimate real-time beamforming filter coefficients to cope with non-stationary environmental noise and dynamic nature of source and microphones positions which results in a set of timevarying room impulse responses. The LSTM adaptive beamformer is jointly trained with a deep LSTM acoustic model to predict senone labels. Further, we use hidden units in the deep LSTM acoustic model to assist in predicting the beamforming filter coefficients. The proposed system achieves 7.97 evaluation set.

READ FULL TEXT
research
08/17/2017

An Improved Residual LSTM Architecture for Acoustic Modeling

Long Short-Term Memory (LSTM) is the primary recurrent neural networks a...
research
12/05/2022

Indoor room Occupancy Counting based on LSTM and Environmental Sensor

This paper realizes the estimation of classroom occupancy by using the C...
research
01/16/2021

A Novel Approach for Earthquake Early Warning System Design using Deep Learning Techniques

Earthquake signals are non-stationary in nature and thus in real-time, i...
research
09/20/2018

LSTM-based Whisper Detection

This article presents a whisper speech detector in the far-field domain....
research
12/08/2021

Active Sensing for Communications by Learning

This paper proposes a deep learning approach to a class of active sensin...
research
05/31/2021

EchoFilter: End-to-End Neural Network for Acoustic Echo Cancellation

Acoustic Echo Cancellation (AEC) whose aim is to suppress the echo origi...
research
11/08/2021

Learning via Long Short-Term Memory (LSTM) network for predicting strains in Railway Bridge members under train induced vibration

Bridge health monitoring using machine learning tools has become an effi...

Please sign up or login with your details

Forgot password? Click here to reset