Face-Age Modeling: A Pattern Recognition Analysis for Age Estimation

This One way to explain the concept of face recognition is to see it as the computational ability engaged in verifying the identity of a person on the basis of the face image regardless of background, face-attachment, face-makeup and illumination. Face recognition has been researched and applied in real life problems such as security access controls, banking identity verification, immigrant verification and licensing. This study examines popular techniques and approaches involved in face recognition and its application in age estimation. The identification characteristics possessed by the face with other demographic information such as age, sex and ethnicity can be predicted using facial features extracted from facial images. The application of face recognition in these areas especially age estimation has not been richly researched. Consequently, this study presents an explorative write-up of face-age models with reference to the application of the commonly used face recognition technique to computationally achieve age prediction. In this study, it's germane to consider face recognition techniques since an important part of age estimation pipeline is the feature extraction which is also important in face recognition. This study shows that most of the researches in age estimation were tested with image taken under controlled or semi-controlled environment which are not sufficient to capture challenges in real life conditions. It was observed that most existing researches uses FG-NET and MORPH 2 face database in testing age estimation systems with scarcity of black face feature across several existing systems. However, more work is desirable, owing to failure of some system to accurately recognize or identify specific faces or using long turnaround processing time for identification vis-a-vise age estimation. This calls for the development of robust black-face database for local testing and implementation among African population. It is also evident that more work is required to show the interaction of epigenetics and its analysis in face-age modeling.