Combining articulatory and acoustic information for speech recognition in noisy and reverberant environments | Synapse