Distinctive feature fusion for recognition of australian English consonants
Trent W. Lewis, David M W Powers · 2008
Audio-Visual Automatic Speech Recognition offers to make speech recognition possible in noisy environments. Early and late fusion approaches dominate the field but may ignore lin-guistically relevant features. Distinctive features offer an alter-native unit for fusion and research has shown that this is feasible on subsets of phonemes [1]. This paper outlines two extended models, multi-class and binary, and results suggest that it is pos-sible to achieve a 20dB gain over audio-only recognition in low SNR environments. Index Terms: audio-visual fusion, distinctive features