Comparison of facial image datasets
Appearance
In computer vision, facial images have been used extensively to develop facial recognition systems, face detection, and many other projects that use facial images. This article compares various facial image datasets.
Database | Subjects | Samples per subject | Total images | Expressions | Poses | 3D | Color | Institution | License |
---|---|---|---|---|---|---|---|---|---|
Bosphorus[1] | 105 | 31-54 | 4652 | 34: action units and 6 expressions, labeled; 24 facial landmarks labeled | 13 pitch, yaw, and cross rotations | Yes; structured light acquisition | Yes | Bogazici University, Turkey | case-by-case, non-commercial, privacy protections[2] |
IMPA-FACE3D[3] | 38 | 14 | 532 | neutral face, 6 expressions (joy, sadness, surprise, anger, disgust, fear), and another 5 facial expressions | n/a | Yes; structured light acquisition | Yes | IMPA & IME/USP, Brazil | ? |
York-3DFace[4] | 350 | 15 | 5250 | neutral face, 5 expressions: anger, happiness, sadness, eyes closed, eye-brows raised | Uncontrol led up and down | Yes | ? | University of York, United Kingdom | ? |
ND2006[5] | 888 | 1-63 | 13450 | 5: happiness, sadness, surprise, disgust, other | n/a | Yes | ? | University of Notre Dame, United States | ? |
CASIA[6] | 123 | 37-38 | 4624 | 5: Anger, smile, laugh, surprise, closed eyes | n/a | Yes | ? | Institute of Automation chinese Academy of Sciences, China | ? |
BU-3DFE[7] | 100 | 25 | 2500 | neutral face, and 6 expressions: anger, happiness, sadness, surprise, disgust, fear (4 levels) | n/a | Yes | ? | Binghamton University, United States | ? |
FRAV3D[8] | 106 | 16 | 1696 | netral face, smaile and mouth and eyes open | Left, right, up, down ? | Yes; structured light acquisition | Yes | Rey Juan Carlos University, Spain | ?
|
FRGC v.2 | 466 | 1-22 | 4007 | 6: anger, happiness, sadness, surprise, disgust, puffy | n/a | Yes | ? | ?
| |
GavabDB | 61 | 9 | 549 | Smile, frontal accentuated laugh, frontal random gesture | Left, right, up, down | Yes | ? | publications must reference; available for immediate download | |
3DRMA | 100 | 20+ | 9971 | Mostly neutral, some unconstrained; neutral are labeled, unconstrained not further labeled | head turning from left to right, head nodding, raised/lowered head turning left to right | Yes | Yes | ? | |
PUT | 120 | 6 | 720 | n/a? | Slight left/right and up/down | No | ? | case-by-case, non-commercial, publications must reference; available for download after email correspondence[9] |
The above table was collected from the paper entitled, "Bosphorus Database for 3D Face Analysis." "
References
- ^ Savran et al.(2008) A. Savran, N. Alyüz, H. Dibeklioğlu, O. Celiktutan, B. Gökberk, B. Sankur, and L. Akarun. Bosphorus database for 3D face analysis. Biometrics and Identity Management, pp 47–56.
- ^ Bosphorus - How to obtain
- ^ Mena-Chalco et al.(2008) J.P. Mena-Chalco, R.M. Cesar-Jr, and L. Velho. Banco de dados de faces 3D: IMPA-FACE3D. Technical report, National Institute for Pure and Applied Mathematics - IMPA - VISGRAF Laboratory, Rio de Janeiro, RJ, Brazil. TR01.
- ^ Heseltine et al.(2008) T. Heseltine, N. Pears, and J. Austin. Three-dimensional face recognition using combinations of surface feature map subspace components. Image and Vision Computing. v. 26, n. 3, pp 382–396.
- ^ Faltemier et al.(2007) T. Faltemier, K. Bowyer, and P. Flynn. Using a multi-instance enrollment representation to improve 3D face recognition. pp 1–6.
- ^ Zhong et al.(2007) C. Zhong, Z. Sun, and T. Tan. Robust 3D face recognition using learned visual codebook. In IEEE Conference on Computer Vision and Pattern Recognition, 2007. CVPR, pp 1–6.
- ^ Yin et al.(2006) L.J. Yin, X.Z. Wei, Y. Sun, J. Wang, and M.J. Rosato. A 3D facial expression database for facial behavior research. In 7th International Conference on Automatic Face and Gesture Recognition (FGR06), pp. 211–216.
- ^ Conde(2006) C. Conde. Biometría: Reconocimiento facial mediante fusión 2D y 3D. Dykinson SL, Madrid.
- ^ PUT license page