Distinguishing children from adults via facial
image analysis has lots of potential real-world applications such
as security access control and human computer interaction.
However, it is still a challenging problem for the computer
vision systems to automatically and effectively distinguish
children from adults. In this paper, we introduce a novel
children recognition method, which improves both the accuracy
and reliability of the latest work on this subject simultaneously.
Results on the FG-NET aging database show that using the
minimum distance classifier on the one dimensional feature
space created by using Active Appearance Model (AAM)
followed by Linear Discriminant Analysis (LDA), we can
recognize the children and adults with accuracies up to 89%
and 90%, respectively.
Age estimation, active appearance model,
LDA, minimum distance classifier.
A. K. Choobeh is with the Department of Electrical engineering,
Buinzahra Branch, Islamic Azad University, Buinzahra, Iran (e-mail:
Alireza Keshavarz Choobeh, "
An Image-Based Method of Distinguishing Children from
Adults," International Journal of Information and Electronics Engineering vol. 3, no. 5, pp.