Lip-Motion-Activated Speech Detection

[[abstract]]自動語音識別運用在許多交通工具的人機互動和人車互動應用上。但為了有效的使用，我們必需先找到一種方法來判斷我們是否正準備使用此功能。因此，語音活動檢測是語音識別的關鍵步驟。我們利用脣形動作來驅動語音檢測的功能。 基於脣形啟動之語音偵測系統是使用臉部檢測和臉部識別來找到脣部的位置，並且運用影像處理來設定四個嘴脣的特徵點。使系統可以利用這四個特徵點的距離來決定使用者是否要開啟語音辨識的功能。[[abstract]]Automatic speech recognition is used in many applications for human-computer interaction and human vehicle interaction on mobile and vehicular platforms. To apply this function, we need to find a way to detect whether we are using it or not. So voice activity detection is the essential step in speech recognition. We use lip-motion-activated speech detection to achieve the function. This lip-motion-activated speech detection system uses face detection and facial recognition to find the location of lip, and defines four feature points of lip by image processing. System can decide that the user want to use the record system based on the distance of those four feature points.[[note]]碩