Introduction - If you have any usage issues, please Google them yourself
To achieve a combination of color space, transform and deformable template automatic lip localization and lip contour extraction, tracking the establishment of the first color model in the space of face detection, location, geometric features of the face by the rough location and then combined with lip lip transformation so that the skin color model, lip color difference visible, presented the results according to the luminance information to transform the image after pretreatment use of segmentation, and further validated by the lip model to achieve precise positioning and then use the deformation of the lip template for lip contour extraction , to enhance the robustness of location within the outline, by the brightness of the pretreatment and the lip model validation by oral Quyu edge map to fit curve to achieve the final positioning within the contour, the lip-reading image sequence in the last frame of the lips After positioning results extend the forecast area as the current frame