Age change for predicting future faces

This paper proposes an age change method for predicting the future face. The first step of the method is to extract face-changing components according to ages from the facial images using principal component analysis (PCA) and a 3D facial shape model (FSM). The second synthesizes the future face by adding the extracted age change components (ACC) to the current face. We divide overall lifetime into three parts such as childhood, manhood and old age for extracting approximated linear ACC from the nonlinearly changing face. This piecewise linear treatment provides us convenience for extraction of the ACC and synthesis of the future faces. The synthesized images show high quality and the only age-changed faces without changing other facial features.