Paper Info
Title
Improved voice activity detection combining noise reduction and subband divergence measures
Abstract
Currently, new trends in wireless communications are demanding reliable human-machine interaction in real-life environments. How- ever, there are obstacles inhibiting automatic speech recognition systems (ASR) working in noisy environments. The main difficulty is the degradation suffered by ASR systems due to a mismatch be- tween training and test conditions. This paper shows an improved voice activity detector (VAD) combining noise reduction and sub- band divergence estimation for improving the reliability of speech recognizers operating in noisy environments. The algorithm for- mulates the decision rule by measuring the divergence between the subband spectral magnitude of speech and noise using the Kullback- Leibler (KL) distance on the denoised signal. Experiments demon- strate a sustained advantage over different VAD methods including standard VADs such as G.729 and AMR, which are used as a refer- ence, recently reported algorithms, and the VADs of the advanced frontend (AFE) for distributed speech recognition (DSR).
Year
Venue
Keywords
2004
INTERSPEECH
voice activity detection,noise reduction
Field
DocType
Citations 
Noise reduction,Decision rule,Divergence,Wireless,Pattern recognition,Computer science,Voice activity detection,Speech recognition,Artificial intelligence,Detector
Conference
0
PageRank 
References 
Authors
0.34
10
5
Name
Order
Citations
PageRank
Javier Ramírez165668.23
José C. Segura248138.14
M. Carmen Benítez330325.05
Ángel de la Torre448234.91
Antonio J. Rubio546636.14