Comparing Deep Learning and Genetic Algorithms Techniques for Age and Gender Classification

Age and gender demographic attributes represent key components which are the basis of a wide range of practical applications including: biometric security and authentication, medical applications or facial recognition. Therefore, in this paper we want to compare the efficiency of both classical Deep Learning (DL) and Genetic Algorithms (GA) approaches to train a model for age and gender demographic attribute using the UTK-Face dataset. In addition, a preprocessing algorithm that approaches unbalanced labels is presented. We obtained competitive results in terms of overall accuracy with 89% for the DL approach as well as 82% for the GA approach.