Genetic Algorithm for Optimal Feature Vector Selection in Facial Recognition

Facial Recognition has a large number of real world applications in Security, Marketing, Healthcare, etc. In order for these applications to be feasible with respect to real time applications it's important that the feature vectors from these facial images be represented such that it is computationally efficient without compromising accuracy. Principal component analysis has often been used for dimensionality reduction. However using an evolutionary algorithm like genetic algorithm can ensure that only the best features are selected to increase efficiency and accuracy while ensuring overfitting doesn't take place. This paper demonstrates a way to effectively represent vectors of a face image using genetic algorithm and its comparison against other similar forms of representation.