Facial Feature Point Localization
Many facial analysis approaches rely on robust and accurate automatic facial landmarking to function correctly. We have developed a statistical method for automatic facial landmark localization. Our landmarking relies on a parsimonious mixture model of Gabor wavelet features, computed in coarse-to-fine fashion and complemented with a shape prior. Here you can download the dataset protocols, we have used for our 2D landmark localization tests in the following paper:
H. Dibeklioğlu, A.A. Salah, T. Gevers, "A Statistical Method for 2-D Facial Landmarking", Image Processing, IEEE Transactions on, vol.21, no.2, pp.844-858, February 2012.
Dataset Protocols for Landmark Localization Tests
| Database | Abbr. |
Training
|
Validation
|
Test
|
| AR | AR |
–
|
 |
–
|
| BioID | BioID |
–
|
–
|
|
| Bosphorus (Frontal) | BOS |
|
|
|
| Bosphorus (+10°) | BOS-R10 |
|
|
|
| Bosphorus (+20°) | BOS-R20 |
|
|
|
| BU-4DFE | BU-4DFE |
–
|
–
|
|
| Cohn-Kanade (Neutral) | CK-N |
–
|
–
|
|
| Cohn-Kanade (Expressive) | CK-E |
–
|
–
|
|
| CK (Neutral + Expressive) | CK-All |
|
|
–
|
| Extended M2VTS | XM2VTS |
|
–
|
–
|
| FRGC (Spring 2004) | FRGC |
|
|
|
