A neural-network based gender detection algorithm on full-face photograph

Gender detection is gaining much more interest for applications such as smart advertisement and authentication. High processing-power available today, makes possible the high process-consuming tasks. In this paper, gender detection of over 18 years old persons is investigated based on their frontal facial images. First the feature-extraction is studied and the best features in a facial image are introduced. A Neural Network (NN) based method is developed for gender detection and a database of 200 females and 200 males is gathered for training and testing the neural network. To eliminate restrictions on race and ethnicity, the database is produced based on some standard Iranian pictures and standard databases exist on this subject. The result of implementation of the proposed method shows completely a noticeable improvement in accuracy, compared to previously performed researches.