Face Database

The Sheffield (previously UMIST) Face Database consists of 564 images of 20 individuals (mixed race/gender/appearance). Each individual is shown in a range of poses from profile to frontal views – each in a separate directory labelled 1a, 1b, … 1t and images are numbered consecutively as they were taken. The files are all in PGM format, approximately 220 x 220 pixels with 256-bit grey-scale.

The authors grant the right to use the face database with the following restrictions:

  • Only images of individuals 1a and 1e may be published (and then only with permission). This is not out of vanity, but for legal reasons.
  • Acknowledgement of use of this database should be provided in any publication. We would, of course, be very interested to hear about any publications.

The database has been used in several of our publications, for example:

Characterizing Virtual Eigensignatures for General Purpose Face Recognition

Daniel B Graham and Nigel M Allinson.
(in) Face Recognition: From Theory to Applications ,
NATO ASI Series F, Computer and Systems Sciences, Vol. 163.
H. Wechsler, P. J. Phillips, V. Bruce, F. Fogelman-Soulie and T. S. Huang (eds), pp 446-456, 1998.


Cropped image example


Uncropped image example

Face Database

Building Image Dataset

This dataset consists of over 3000 low-resolution images of forth different buildings – typically between 70 and 120 images per building. The images are taken from different viewpoints under widely different lighting conditions. They are intended to represent the quality and variety of images obtained by hand-held mobile devices for a range of weather and time-of-day conditions.

The building image dataset has been presenting in the following publication.

Jing Li and Nigel M. Allinson Dimensionality Reduction-Based Building Recognition, in Proceedings of the Ninth IASTED International Conference of Visualisation, Imaging and Image Processing Cambridge (UK), July 13-15, 2009.


Building Image Dataset

REVIEW Dataset

REVIEW is a new retinal vessel reference dataset consisting of four image sets. Four image sets were chosen to assess the accuracy and precision of the vessel width measurement algorithms in the presence of pathology and a central light reflex and to compare the performance of the proposed algorithms with manual measurements.

This dataset includes 16 images with 193 vessel segments, demonstrating a variety of pathologies and vessel types. To see this dataset and download it please follow the link below to one of our Academics website.


Retinal Image Computing and Understanding