We list some face databases widely used for face related
studies, and summarize the specifications of these databases as below.
o
Source: the FERET
database is sponsored by the Defense Advanced Research Products Agency (DARPA).
o
Purpose: the FERET database
is widely used as the standard face database to evaluate the face recognition
systems. It may also be used for face pose estimation and eye detection.
o
Properties:
Properties |
Descriptions |
# of subjects |
1199 |
# of images/videos |
14051 |
Static/Videos |
Static |
Single/Multiple
faces |
Single |
Gray/Color |
eight-bit gray |
Resolution |
256*384 |
Face pose |
7 categories: Frontal,
quarter-left, quarter-right, half-left, half-right, full-left, full-right |
Facial expression |
Slight facial
expression changes |
Illumination |
Controlled
illumination |
3D data |
N/A |
Ground truth |
Positions of eyes,
nose, and mouth Identifications of
subjects |
o
Reference: refer to
the paper: P. J. Phillips, Hyeonjoon Moon, S. A. Rizvi, and P. J. Rauss, The FERET evaluation methodology for face
recognition algorithm, IEEE Trans. on PAMI, vol. 22, no. 10, pp. 1090-1104,
October
o
Source: CAS-PEAL
database is obtained from
o
Purpose: CAS-PEAL
database is used to evaluate the face recognition systems. It may also be used
for eye detection, face pose estimation, and facial expression recognition.
o
Properties:
Properties |
Descriptions |
# of subjects |
1040 (595 males and
445 females) of Asians |
# of images/videos |
30,900 |
Static/Videos |
Static |
Single/Multiple
faces |
Single |
Gray/Color |
eight-bit gray |
Resolution |
360*480 |
Face pose |
21 pose angles vertical: up,
middle, and down horizontal: left to
right (67, 45, 22, 0, -22, -45, -67) |
Facial expression |
6 facial expressions: neutral, eye
closing, frown, smile, surprise, and
mouth open |
Illumination |
15 lighting
conditions |
Accessories |
3 kinds of glasses
and 3 kinds of caps |
3D data |
N/A |
Ground truth |
Positions of eyes Identifications of
subjects Face pose angles Facial expression
labels Illumination
positions |
o
Reference: Please
refer to the technical report JDL-TR-04-FR-001 The CAS-PEAL Large-Scale
Chinese Face Database and Baseline Evaluations.
o
Source: The PIE
database is obtained from the Robotics Institute of Carnegie Mellon University.
o
Purpose: PIE
database is used to evaluate the face recognition systems. It may also be used
for facial feature detection, face pose estimation, and facial expression
recognition.
o
Properties:
Properties |
Descriptions |
# of subjects |
68 |
# of images/videos |
41,368 |
Static/Videos |
Static |
Single/Multiple
faces |
Single |
Gray/Color |
Color |
Resolution |
640*486 |
Face pose |
13 pose angles in
vertical and horizontal |
Facial expression |
4 facial
expressions: neutral, eye
closing, smiling, and talking |
Illumination |
N/A |
Accessories |
Glasses |
3D data |
N/A |
Ground truth |
Some feature point
data Identifications of
subjects Measured locations
of camera Head pose Facial expression
labels Illumination
positions |
Additional
materials |
Background images |
o
Reference: Please
refer to the CMU Technical Report CMU-RI-TR-01-02 The CMU Pose, Illumination,
and Expression (PIE) Database of Human Faces.
o
Source: this
database is obtained from the Robotics Institute of Carnegie Mellon University.
It combines images collected at CMU and MIT.
o
Purpose: this
database is primarily used for face detection task. It may also be used for eye
detection and facial feature detection.
o
Properties:
Properties |
Descriptions |
# of subjects |
N/A |
# of images/videos |
169 frontal-view
face images 202 profile face
images |
Static/Videos |
Static |
Single/Multiple
faces |
Multiple |
Gray/Color |
eight-bit gray |
Resolution |
N/A |
Face pose |
Frontal and profile |
Facial expression |
Various facial
expressions |
Illumination |
Various lighting
conditions |
Accessories |
Various |
3D data |
N/A |
Ground truth |
Positions of eyes, nose
tip, mouth corners, and mouth center for each face (frontal-view face); Positions of eye
corner, eye, nose, nose tip, mouth corner, mouth center, chin, earlobe, and
ear tip for each face (profile face) |
o
Source: this
database is constructed by
o
Purpose: this
database can be used for face recognition and facial expression recognition.
o
Properties:
Properties |
Descriptions |
# of subjects |
15 |
# of images/videos |
165 |
Static/Videos |
Static |
Single/Multiple
faces |
Single |
Gray/Color |
eight-bit gray |
Resolution |
320*243 |
Face pose |
Frontal view |
Facial expression |
6 facial expressions: neutral, happiness,
sadness, sleepiness, surprise, and wink |
Illumination |
3 lighting
conditions: center-light, left-light, and right-light |
Accessories |
Glasses |
3D data |
N/A |
Ground truth |
Identifications of
subjects Facial expression
labels Illumination
positions |
o
Reference: Please
refer to the paper P. N. Belhumeur, J. P. Hespanha, and D. J. Kriegman, Eigenfaces vs. Fisherfaces:
Recognition using class specific linear projection, IEEE Trans. on PAMI, vol.
19, no. 7, pp. 711-720, July
o Source: this database is constructed by
o
Purpose: this database
can be used for face recognition, face pose estimation, and eye detection.
o
Properties:
Properties |
Descriptions |
# of subjects |
10 |
# of images/videos |
5760 |
Static/Videos |
Static |
Single/Multiple
faces |
Single |
Gray/Color |
Gray |
Resolution |
640*480 (eye
distance ~ 90pixels) |
Face pose |
9 poses |
Facial expression |
Neutral |
Illumination |
64 lighting
conditions and 1 ambient illumination |
Accessories |
N/A |
3D data |
N/A |
Ground truth |
Identifications of subjects Face pose Illumination
positions Coordinates of eyes
and mouth (frontal view) Coordinates of face center (other views) |
o
Reference: Please
refer to the paper A. S. Georghiades and P. N. Belhumeur, From Few to Many: Illumination Cone Models
for Face Recognition under Variable Lighting and Pose, IEEE Trans. on.
PAMI, vol.23, no.6, pp.643-660, June
o
Source: this database
is constructed by Aleix Martinez and Robert Benavente in the Computer Vision Center (CVC) at the UAB.
o
Purpose: this
database is primarily used for face recognition. It may also be used for facial
expression recognition.
o
Properties:
Properties |
Descriptions |
# of subjects |
126 (70 male and 56
female) |
# of images/videos |
4000 |
Static/Videos |
Static images and
image sequences |
Single/Multiple
faces |
Single |
Gray/Color |
color |
Resolution |
768*576 |
Face pose |
Nearly frontal-view
or quarter-profile images |
Facial expression |
4 facial
expressions: neutral, smile,
anger, and scream |
Illumination |
3 illumination
conditions: left, right, and
all side lights on |
Accessories |
Sun glasses, scarf |
3D data |
N/A |
Ground truth |
Identifications of
subjects Facial expression
labels |
o Reference: Please refer to the technical report A. M.
Martinez and R. Benavente, The AR Face Database,
CVC Technical Report #24, June