A Deep Learning Model for Classification of Gender and Age from Facial Images

Abstract

Gender classification and Age identification play an important role in our social lives. Gender is central characteristics of personality, and it is essential in our life. Age is important for our identity. Security, biometric system, and treatment are part of gender classification and age prediction. Age prediction can help to authorize people from buying adult products or other kind of restricted goods. In this project, for classification of gender and prediction of age from pictures deep learning model is used. The objective of this study is to create a model for gender classification and an age estimation using convolutional neural networks and ResNet50. The image's feature extraction and categorization are included by CNN. Feature extraction gives the features corresponding to gender and age from the face pictures whereas the classification classify the image into correct age and gender.ResNet50 is the convolutional network that have 50 layers. Age prediction is the regression problem and prediction of gender is a binary classification problem. The model is evaluated using the UTKFace dataset, a sizable face dataset with a broad age range. Deep learning algorithm is used to obtain higher accuracy and lower MAE, also MAE of the both algorithm is compared to obtain which algorithm more efficient.