Face Validation with Facial Model using Genetic Algorithms

In this paper, the Genetic Algorithms (GA) is used to learn the facial model consisting of eyes, nose and mouth for face validation. This facial model improves the detection rate for the Maximum-Likelihood (ML) head detector [9], which produces ellipse-like objects as face candidates, by applying a second validation stage to verify these candidates. In formulating the genetic algorithm, a two-dimensional binary genome mapped from the facial images is used to encode the chromosomes. We also propose to employ properties, such as the eye position and the eye-point preservation, to evaluate the fitness in the genetic algorithm. We demonstrate the experimental results of using two initialization models, namely the blank model and the average edge model, to learn the facial model for face validation.