Humans are often able to compensate for noise degradation and uncertainty in speech information by augmenting the received audio with visual information. Such bimodal perception generates a rich combination of information that can be used in the recognition of speech. However, due to wide variability in the lip movement involved in articulation, not all speech can be substantially improved by audio-visual integration. This paper describes a feature-fusion audio-visual speech recognition (AVSR) system that extracts lip geometry from the mouth region using a combination of skin color filter, border following and convex hull, and classification using a Hidden Markov Model. The comparison of the new approach with conventional audio-only system is made when operating under simulated ambient noise conditions that affect the spoken phrases. The experimental results demonstrate that, in the presence of audio noise, the audio-visual approach significantly improves speech recognition accuracy compared with audio-only approach.
展开▼
