Classifier performance is often enhanced through combining multiple streams of information. In the context of multi-stream HMM/ANN systems in ASR, a confidence measure widely used in classifier combination is the entropy of the posteriors distribution output from each ANN, which generally increases as classification becomes less reliable. The rule most commonly used is to select the ANN with the minimum entropy. However, this is not necessarily the best way to use entropy in classifier combination. In this article, we test three new entropy based combination rules in a full-combination multi-stream HMM/ANN system for noise robust speech recognition. Best results were obtained by combining all the classifiers having entropy below average using a weighting proportional to their inverse entropy.
[1]
Claude E. Shannon,et al.
The mathematical theory of communication
,
1950
.
[2]
H Hermansky,et al.
Perceptual linear predictive (PLP) analysis of speech.
,
1990,
The Journal of the Acoustical Society of America.
[3]
Hynek Hermansky,et al.
RASTA processing of speech
,
1994,
IEEE Trans. Speech Audio Process..
[4]
Ronald A. Cole,et al.
New telephone speech corpora at CSLU
,
1995,
EUROSPEECH.
[5]
Hervé Glotin,et al.
Multi-stream adaptive evidence combination for noise robust ASR
,
2001,
Speech Commun..
[6]
Martin Heckmann,et al.
Noise Adaptive Stream Weighting in Audio-Visual Speech Recognition
,
2002,
EURASIP J. Adv. Signal Process..