Derin öğrenme tekniklerine dayalı yüz yaş grubu sınıflandırıcısı

Recently, researchers in the computer vision field have become increasingly interested in extracting important information from the human face, including estimating demographic characteristics such as age, gender, and race, due to their increasing applications in the real world. Models based on convolutional neural network (CNN) have proven that they are one of the most important techniques capable of extracting facial features automatically from images of human faces. In this thesis, a novel CNN-based model to extract facial discriminative features from face images and classify those images into the corresponding age group is proposed. The overall performance of the proposed model is evaluated on both UTKFace and Facial-age datasets. The results showed that the model has a better performance compared with recent studies in terms of classification accuracy. The model achieved age group classification accuracy of 87.82% on the mentioned facial aging datasets.