Classification of Demographic Attributes from Facial Image by using CNN

Human facial appearance is emphatically influenced by demo-graphical features like definite age, ethnicity and sexual orientation with each category advance apportioned into categories White, Dark, East Asian, Male, Female, Child (1-18), Young (19-36), Center Age (37-54) and old (55-above). Most subjects share a more comparative appearance with the possess demographic class than with other demographic class. We assess here the precision of automatic facial verification for subjects having a place to changing age, ethnicity, and gender categories. For this reason, we utilize transfer learning technique in which pre-trained convolutional neural network is used for feature mining and to present that our strategy yields a satisfactory execution on person demographics for development of a viable facial recognition system. We have used IMDB data set for training of our system. We have concluded that results on ethnicity group white are relatively lower than other ethnicity group's result. We talk about the outcomes and make recommendations for improving facial image classification over changing demographics, in expansion to the advancement of a framework.