Human age estimation based on multi-level local binary pattern and regression method

In this paper, a novel method for human age estimation is proposed. This research is novel in the following four ways. First, the in-plane rotation of face region is compensated based on the detected positions of two eyes by Adaboost method. The region of interest (ROI) for extracting age features in the detected face region is re-defined based on the distance between two eyes. Second, multi-level local binary pattern (MLBP) method is applied in order to extract the features for age estimation. Third, in order to solve the problem of age estimation by active appearance model (AAM), we extract whole texture information by MLBP which takes low processing time. Fourth, the human age is estimated using support vector regression based on the texture features. The experimental results show that the proposed method can estimate the human age with the mean absolute error (MAE) of 6.58 years.