Abstract

A new age estimation method that takes classifiable ability of each component in age manifold into account is considered. First, we analysis the age classification rate of each component in reduced dimension age manifold. Second, we apply this property to kernel function in popular method such as SVM. This is implemented by weighted kernel function. Finally, we evaluate this method in “wild” face image database. Experimental results demonstrate the effectiveness and robustness of our proposed framework.

Authors

Pak Duho, Ri Kumhyok, Hyon Cungyong
Kim Il Sung University, D.P.R. of Korea

Keywords

Age Estimation, Support Vector Machine, Support Vector Regression