Other Face Databases From Other Research Groups

 

 

  1. WIDER FACE

o       Source: this database is built by Multimedia Laboratory at the Chinese University of Hong Kong.

o       Purpose: this database is primarily used for face detection and contains "in-the-wild" images collected from publicly available WIDER dataset.

o       Properties:

Properties

Descriptions

# of subjects

N/A

# of images/videos

Training set: 12k+ images and labeled 157k+ faces

Validation set:3k+ images and labeled 39k+ faces.

Test set:16k+ images and labeled 196k+ faces.

Static/Videos

Static

Single/Multiple faces

Multiple

Gray/Color

24-bit RGB

Resolution

N/A

Face pose

Annotated with typical/atypical pose variation

Facial expression

Annotated with typical/exaggerate facial expression

Illumination

Annotated with normal/extreme illumination

Occlusion

Annotated with no/partial/heavy occlusion

Blur

Annotated with clear/normal/heavy blur

Ground truth

Cropped face region

o       Reference: please refer to paper: Yang, Shuo and Luo, Ping and Loy, Chen Change and Tang, Xiaoou, WIDER FACE: A Face Detection Benchmark, IEEE Conference on Computer Vision and Pattern Recognition, 2016.

 

  1. MIT CBCL Face Database #1

o       Source: this database is constructed by Center for Biological and Computational Learning at MIT.

o       Purpose: this database is primarily used for face detection.

o       Properties:

Properties

Descriptions

# of subjects

N/A

# of images/videos

Training set: 2,429 faces and 4,548 non-faces

Test set: 472 faces and 23,573 non-faces.

Static/Videos

Static

Single/Multiple faces

Single

Gray/Color

Eight-bit gray

Resolution

19*19

Face pose

Moderate pose variation

Facial expression

Moderate facial expression changes

Illumination

Moderate illumination variation

Accessories

Glasses

3D data

N/A

Ground truth

Cropped face region

 

  1. Spacetime Face Database

o       Source: this database is constructed by the University of Washington Graphics and Imaging Laboratory.

o       Purpose: this database is primarily used for face modeling and animation.

o       Properties: 384 face meshes, each with about 23K vertices.

o       Reference: Please refer to the paper: Li Zhang, Noah Snavely, Brian Curless, and Steve Seitz, Spacetime Faces: High-resolution capture for modeling and animation, Proc. of ACM SIGGRAPH2004”.

 

  1. BioID Database

o       Source: this database is constructed by HumanScan company.

o       Purpose: this database can be used for face detection, face recognition and eye detection.

o       Properties:

Properties

Descriptions

# of subjects

23

# of images/videos

1521

Static/Videos

Static

Single/Multiple faces

Single

Gray/Color

Gray

Resolution

382*288 (eye distance ~ 50pixels)

Face pose

Frontal

Facial expression

Various

Illumination

Various lighting conditions

Accessories

Various

3D data

N/A

Ground truth

Coordinates of eyes.

Coordinates of 20 feature points (Eyebrow corners, eye, mouth and tip of chin)

·        Jesorsky, K. Kirchberg, R. Frischholz, Robust Face Detection Using the Hausdorff Distance, Audio and Video based Person Authentication - AVBPA 2001, pages 90-95. Springer, 2001..


  1. CVL Face Database

o       Source: this database is constructed by Peter Peer, Computer Vision Laboratory of the University of Ljubljana.

o       Purpose: this database can be used for face detection, feature detection, face recognition and 3D face modeling.

o       Properties:

Properties

Descriptions

# of subjects

114

# of images/videos

798

Static/Videos

Static

Single/Multiple faces

Single

Gray/Color

Color

Resolution

640*480

Face pose

Horizontal: -90, -45, 0, 45, 90

Facial expression

3 facial expressions: serious, smiling( showing teeth and showing no teeth)

Illumination

N/A

Accessories

N/A

3D data

N/A

Ground truth

N/A

·        Reference: Please contact Peter Peer (Peter.peer@fri.uni-lj.si) and Computer Vision Laboratory of the University of Ljubljana for the database.



  1. NIST Mugshot Identification Database (MID)

o       Source: this database is constructed by the National  Institue of Standards and Technology.

o    Purpose: this database is primarily for use in development and testing of automated mugshot identification systems.

o       Properties:

Properties

Descriptions

# of subjects

1573 (1495 male and  78 female)

# of images/videos

3248

Static/Videos

Static

Single/Multiple faces

Single?

Gray/Color

Gray

Resolution

Varying in size from 1" - 21/2" in height

Face pose

front and profile

Facial expression

N/A

Illumination

N/A

Accessories

N/A

3D data

N/A

Ground truth

Subject identity


  1. The University of Oulu Physics-Based Face Database

o       Source: this database is collected at the Machine Vision and Media Processing Unit, University of Oulu.

o       Purpose: this database can be used for research in face recognition and color.

o       Properties:

Properties

Descriptions

# of subjects

125

# of images/videos

N/A (each person has 16 frontal views and an additional 16 if the person has glasses)

Static/Video

Static

Single/Multiple faces

Single

Gray/Color

Color

Resolution

428*569

Face pose

Minor various poses

Facial expression

N/A

Illumination

Four illuminants: Horizon, Incandescent, Fluorescent and Daylight.

Accessories

Glasses

3D data

N/A

Ground truth

N/A

·        Reference: Please refer to the papsers "A physics-based face database for color research, Journal of Electronic Imaging Vol. 9 No. 1 pp.  32-38." and "Color correction of face images under different illuminants by RGB eigenfaces,  Proc. 2nd Audio- and Video-Based Biometric Person Authentication Conference (AVBPA99), March 22-23, Washington DC USA pp. 148-153."


  1. UCD Color Face Image (UCFI) Database

o       Source: this database is constructed by the Department of Electronic & Electrical Engineering, University College Dublin. The images are required form a wide variety of sources such as digital cameras, pictures scanned using photo-scanner, other face databases and the world wide web.

o       Purpose: this database is primarily to test new face detection algorithms using color information.

o       Properties:

Properties

Descriptions

# of subjects

N/A

# of images/videos

299

Static/Videos

Static

Single/Multiple faces

Single?

Gray/Color

Color

Resolution

Vaious

Face pose

Frontal, profile, intermediate, upright and rotated

Facial expression

Various

Illumination

Various

Accessories

Glasses, sunglasses, beards, moustaches

3D data

N/A

Ground truth

Hand segmantation