Table 3 Total error values for original CLNF model and the periodical spline interpolation method (expressed in pixels and cm)
From: Inner lips feature extraction based on CLNF with hybrid dynamic template for Cued Speech
RMSE | Speaker 1 | Speaker 2 | Speaker 3 | Total |
|---|---|---|---|---|
CLNF | 39.64 (2.24 cm) | 39.94 (2.26 cm) | 32.55 (1.84 cm) | 37.40 (2.12 cm) |
The proposed method | 8.03 (0.45 cm) | 5.84 (0.33 cm) | 5.07 (0.28 cm) | 6.11 (0.35 cm) |