Noise reduction based on mouth area movement recognition

摘要: A computing device can capture video data of at least a portion mouth area (e.g., mouth, lips, tongue, chin, jaw) user the device. The also sound including voice as well noise (e.g. background noise). be processed to detect movement area. analyzed and compared with models characteristic oral communication speech, song). If corresponds one model communication, then indicates that is likely engaging in communication. Noise reduction applied and/or increased on captured reduce turn enhance user's voice.