Age prediction using a supervised facial model

Facial rejuvenation has driven a lot of research in the field of dermatology and plastic surgery, leading to many medical procedures. This paper proposes an age prediction method that could be used to better understand the ageing process and to evaluate the benefits of a rejuvenating treatment, for example. A supervised Facial Model (SFM) is built using Partial Least Squares regression (PLSR) to capture and summarize age related changes from a database of front face images. The model describes the changes related to the shape and proportions of facial features, color and texture of the face. Experimental results from a database of 173 Caucasian women pictures demonstrate that the model matches human perception.