FEI Face Database


Disclaimer

The FEI face database is available for research purposes only. Permission to use but not distribute or reproduce this database is granted to all researchers.

Description and Download

The FEI face database is a Brazilian face database that contains a set of face images taken between June 2005 and March 2006 at the Artificial Intelligence Laboratory of FEI in São Bernardo do Campo, São Paulo, Brazil. There are 14 images for each of 200 individuals, a total of 2800 images (a morphed version of these images with lots of synthetically modified samples can be found here). All images are colourful and taken against a white homogenous background in an upright frontal position with profile rotation of up to about 180 degrees. Scale might vary about 10% and the original size of each image is 640x480 pixels. All faces are mainly represented by students and staff at FEI, between 19 and 40 years old with distinct appearance, hairstyle, and adorns.  The number of male and female subjects are exactly the same and equal to 100. Figure 1 shows some examples of image variations from the FEI face database.

 

Figure 1. Some examples of image variations from the FEI face database.

Download:

originalimages_part1 (~ 84MB)

originalimages_part2 (~ 87MB)

originalimages_part3 (~ 86MB)

originalimages_part4 (~ 87MB)

In addition, we provide a subset of the FEI face database composed of only frontal face images previously aligned to a common template so that the pixel-wise features extracted from the images correspond roughly to the same location across all subjects. In this manual alignment, we have randomly chosen the frontal image of a subject as template and the directions of the eyes and nose as a location reference. All these frontal images were then cropped to the size of 360x260 pixels. Since the number of subjects is equal to 200 and each subject has two frontal images (one with a neutral or non-smiling expression and the other with a smiling facial expression), there are 400 full frontal face images manually registered to evaluate experiments on a controlled environment.

Download:

frontalimages_manuallyaligned_part1 (~ 6MB)

frontalimages_manuallyaligned_part2 (~ 6MB)

More recently, with the support of CNPq under the grant 472386/2007-7, we have proposed and implemented a sequence of image processing steps to automatically normalize, equalize and crop frontal face images. A computational tool has been developed to minimize image variations due to the acquisition protocol. Figure 2 shows the average face images of the neutral (left) and smiling (right) facial expressions before and after the normalization only. We believe that these additional sets of frontal face images are particularly useful for evaluating experiments on synthesizing realistic expressions and aging estimation in 2D face data sets.

Download:

frontalimages_spatiallynormalized_cropped_equalized_part1 (~ 4MB)

frontalimages_spatiallynormalized_cropped_equalized_part2 (~ 4MB)

frontalimages_spatiallynormalized_cropped_equalized_averagefaceimages (~ 0.01MB)

frontalimages_spatiallynormalized_part1 (~ 6MB)

frontalimages_spatiallynormalized_part2 (~ 6MB)

frontalimages_spatiallynormalized_averagefaceimages (~ 0.03MB)

 

Figure 2. Average face images of the neutral (left) and smiling (right) facial expressions before and after the normalization only.

Using a set of public tools written by Professor Timothy F. Cootes, we have also annotated manually the same corresponding 46 points on each spatially normalized (not cropped) frontal face image, as exemplified in Figure 3. To download all the 400 frontal annotated shapes, please click here.

 

Figure 3. A landmark example of the 46 points consistently and manually annotated on each frontal face image of the database.

For more details on the FEI face database, its acquisition protocol, annotation, and the image processing steps to automatically normalize frontal face images, please read and reference the following article (in english) or papers (in portuguese):

C. E. Thomaz and G. A. Giraldi. A new ranking method for Principal Components Analysis and its application to face image analysis, Image and Vision Computing, vol. 28, no. 6, pp. 902-913, June 2010.

E. Z. Tenorio and C. E. Thomaz. Analise multilinear discriminante de formas frontais de imagens 2D de face. In proceedings of the X Simposio Brasileiro de Automacao Inteligente SBAI 2011, pp. 266-271, Universidade Federal de Sao Joao del Rei, Sao Joao del Rei, Minas Gerais, Brazil, 18th - 21st September 2011.

V. Amaral, C. Figaro-Garcia, G. J. F. Gattas, C. E. Thomaz. "Normalizacao espacial de imagens frontais de face em ambientes controlados e nao-controlados" (in portuguese), Periodico Cientifico Eletronico da FATEC Sao Caetano do Sul (FaSCi-Tech), vol. 1, no. 1, October 2009.

V. Amaral and C. E. Thomaz. "Normalizacao Espacial de Imagens Frontais de Face". Technical Report 01/2008 (in portuguese), Department of Electrical Engineering, FEI, São Bernardo do Campo, São Paulo, Brazil, October 2008.

L. L. de Oliveira Junior and C. E. Thomaz. "Captura e Alinhamento de Imagens: Um Banco de Faces Brasileiro". Undergraduate Technical Report (in portuguese), Department of Electrical Engineering, FEI, São Bernardo do Campo, São Paulo, Brazil, June 2006.

Inquiries for further information may be made to:

Dr. Carlos Eduardo Thomaz

Image Processing Laboratory

Department of Electrical Engineering

Centro Universitario da FEI, São Bernardo do Campo, São Paulo, Brazil

Phone: +55 (0) 11 4353-2910

e-mail: cet@fei.edu.br

[back]


Last updated: novembro 30, 2023.