CONTENT ACCESS USING FACE BIOMETRICS

Abstract

In recent years, privacy and security have become a major challenge to all of us. Each type of data is easily available and accessible to all. This raises a concern that is everyone should have access to every kind of data? Unlike for adults, the access to the data contents for children should be limited and access to adult content like pornography etc. should be controlled and kept out of reach of the children until they become certain years of age. Therefore, we aim to design an approach that allows the access of data contents stored in a system based on the face biometrics of the person trying to access the data. We have thus created a system with data mainly videos that are segregated in two categories mainly adult videos and child videos. Our approach works in real time and allows access to the data to a particular age group only. It works for the prediction of the age group to allow access to the data contents. So, first, the age group of a person is predicted by using the Voila-Jones algorithm to detect the faces or face limitations. After detecting the face, features are extracted using HOG approach to get the feature vector, and then to train the system we have used KNN classifier. A publically available dataset is used for training purposes. We have thus implemented an approach that works in real time and gives efficient results.