A newem estimationof dynamic stream weights for coupled-HMM-based audio-visual ASR.

Ahmed Hussen Abdelaziz Steffen Zeiler Dorothea Kolossa

Published in: ICASSP (2014)

Keyphrases

audio visual
multi stream
multi modal
hidden markov models
audio visual speech recognition
visual information
temporal context
visual data
data streams
automatic speech recognition
emotion recognition
multimedia
person authentication
image classification
data sets
low level
three dimensional
databases