Automated Facial Features Points Localization for Age Estimation Based on Ideal Frontal Symmetry and Proportion of the Face

Abstract

Age Estimation Components, i.e. the extraction and localization of facial features landmark points, which represent ageing pattern are very important steps in the age estimation process. A good age estimation technique will contain enough variation of landmark points and consider all important points to express the full complexity of the problem. Major progress has been achieved recently in the area of facial landmark localization and extraction method for age estimation. Moreover, measuring facial features landmark points for age classification algorithm has become an interesting subject when dealing with automatic localization process. However, the difficulties to measure the points automatically divert to wrong result if the method unable to locate exact facial features landmark points properly at critical area with complex appearance. Such as measuring points at upper region of face when dealing with individuals with no hair or hair that covered part of the forehead. Therefore, we address this issue by proposing a new method to automatically localize optimal facial landmark points from an input of face image based on Ideal Frontal Symmetry and Proportion of the face. The performance of the proposed algorithm is evaluated with baseline localization using qualitative evaluations. The proposed method achieved a satisfying outcome, which is an average of 80% detection rate for every detected landmark points. The advantage of this method is to accurately identify the points with automatic processing. Each of the point’s position localization process was learned independently so that it is suitable to be implemented in real-time face tracking application.