Method and apparatus for audio-visual speech detection and recognition

摘要

Techniques for providing speech recognition comprise the steps of processing a video signal associated with an arbitrary content video source, processing an audio signal associated with the video signal, and recognizing at least a portion of the processed audio signal, using at least a portion of the processed video signal, to generate an output signal representative of the audio signal.