An Active Age Estimation of Facial image using Anthropometric Model and Fast ICA

In this paper, an efficient feature extraction method based on the Kande-Lucas-Tomasi (KLT) using fast independent component analysis (Fast ICA) & Anthropometric Model as the distance measure is proposed. Each face is extracted facial organs are marked for Anthropometric Model (AM) distance measure. The KLT facial coefficients of low & high frequency in different scales & various angles are obtained. The coefficients are utilized as a feature vector for further processing. Considering the extracted face image & adopt the Fast-ICA algorithm based on entropy to extract the face feature information. Finally, according to the Anthropometric distance to classify face feature & Artificial Neural network (ANN) used to estimate age for all kinds of facial databases. Experiments are done by using the YALE & FERET databases. An experimental outcome shows that the recognition rate Mean Absolute Error (MAE) of the proposed algorithm is acceptable & very promising, & confirm the success of the proposed face feature extraction approach.