Research Article
A Robust Shape Reconstruction Method for Facial Feature Point Detection
Table 1
Comparison of different parameter sets on LFPW (68 points) dataset. Here
is fixed to 2.
| | | Size of | Size of | | (tr ts) | Mean errors |
| 2 | 2 | 256 | 256 | 5 | 1 1 | 0.079381 | 256 | 512 | 5 | 1 1 | 0.08572 | 512 | 256 | 5 | 1 1 | 0.085932 | 512 | 512 | 5 | 1 1 | 0.086731 |
| 2 | 4 | 256 | 256 | 5 | 1 1 | 0.081958 | 256 | 512 | 5 | 1 1 | 0.083715 | 512 | 256 | 5 | 1 1 | 0.086187 | 512 | 512 | 5 | 1 1 | 0.087157 |
| 2 | 6 | 256 | 256 | 5 | 1 1 | 0.07937 | 256 | 512 | 5 | 1 1 | 0.07986 | 512 | 256 | 5 | 1 1 | 0.075987 | 512 | 512 | 5 | 1 1 | 0.084429 |
| 2 | 8 | 256 | 256 | 5 | 1 1 | 0.075863 | 256 | 512 | 5 | 1 1 | 0.082588 | 512 | 256 | 5 | 1 1 | 0.077048 | 512 | 512 | 5 | 1 1 | 0.082644 |
| 2 | 10 | 256 | 256 | 5 | 1 1 | 0.076178 | 256 | 512 | 5 | 1 1 | 0.076865 | 512 | 256 | 5 | 1 1 | 0.080907 | 512 | 512 | 5 | 1 1 | 0.088414 |
|
|