This set of pixels varies as the lip boundary changes during speech. The vertical profile is mapped to a feature vector. Both the shape-based and ...
確定! 回上一頁