Speaker-independent machine lip-reading with speaker-dependent viseme classifiers
This work addresses speaker variability in lip-reading for applications like assistive technology, but it is incremental as it builds on existing phoneme-clustering methods without introducing a new paradigm.
The paper tackled the problem of speaker-dependent visual speech in machine lip-reading by creating new phoneme-to-viseme maps for individual and multiple speakers, concluding that speakers share a common set of mouth gestures but differ in how they use them.
In machine lip-reading, which is identification of speech from visual-only information, there is evidence to show that visual speech is highly dependent upon the speaker [1]. Here, we use a phoneme-clustering method to form new phoneme-to-viseme maps for both individual and multiple speakers. We use these maps to examine how similarly speakers talk visually. We conclude that broadly speaking, speakers have the same repertoire of mouth gestures, where they differ is in the use of the gestures.