The paper considers the problem of audio-visual speech recognition in a simultaneous (target/masker) speaker environment. The paper follows a conventional multistream approach and...
The Multi-Stream automatic speech recognition approach was investigated in this work as a framework for Audio-Visual data fusion and speech recognition. This method presents many ...
Multi-stream hidden Markov models (HMMs) have recently been very successful in audio-visual speech recognition, where the audio and visual streams are fused at the final decision...