Children Detection Algorithm Based on Statistical Models and LDA in Human Face Images

Advances in different software and systems, and the everyday increasing demand on internet networks highlights the need for a system being able to give service to its clients based on their age. Children and adolescents are the most vulnerable group in the society. Therefore we have to look for an algorithm that can categorize immature from adult. In this paper, a practical algorithm in children classification from adults by their facial image is proposed. In this algorithm, statistical modelling of the face is used to extract the age dependent face features and then by applying Linear Discriminant Analysis (LDA) on the face parameters, useful specifications are extracted. By transferring into a one dimension feature space Euclidean distance is used as a dissimilarity function. The proposed algorithm obtains accuracy rate of 85% on a standard FG-NET aging face database.