Research Article
Collaborative Research on Mouth Shape and Lyrics in Singing Practice Based on Image Processing
Table 2
Comparison of three extraction methods.
| Method | Typical algorithm | Advantages | Disadvantages |
| Shape-based method | Geometric feature method, snake model method, and active contour model method | Low feature dimension; it is not easy to change | The lip movement information reflected is not comprehensive; the image has clear edges | Pixel-based method | Method based on image transformation | All pixels represent visual information together, and the loss of information is relatively small | High feature dimension; image geometric transformation sensitivity | Hybrid methods | Method based on motion analysis | Combine shape-based and pixel-based features | The algorithm is complex; it is difficult to extract the ideal contour in processing |
|
|