Abstract: Online Social Networks (ONS) is an important part of our everyday life...At present people are very interested to share the photo on online social networks, unluckily which is used for the purpose we not at all imagine. To check possible confidentiality outflow of a group photograph, we prepared a mechanism in which each person can take part in the decision making on the photograph posting, so everyone in the photograph must be recognized. It is important to have a professional facial recognition (FR) system that cans distinguish each person in the photograph. Haar cascade classifier and CBIR (content based image retrieval) algorithm are used in the proposed system for face detection and recognition. Once the faces are recognized from the group photograph then acceptance or reject notifications are sending automatically to each identified people within the close friend circle. The photograph will be posted if all the individuals within the friend circle are accepting the notification and it will not be posted if any of them rejects the notification. We expect that this proposed system would be very useful in protecting users’ confidentiality in photo/image sharing over online social networks.

Keywords: Online Social Network, CBIR, K-Mean Clustering, Face Recognition.