Abstract: In today’s world, sharing of personal information like images more easily due to the enhancement in content sites. Number of images user can share with more number of people. Increasing use of social media violates the privacy of contents due to large amount of sharing of the data. With the purpose to help the users to compose privacy settings for the images shared through social sites and improve content search method. This paper introduces a system called Adaptive Privacy Policy Prediction (A3P) system through which users can create their privacy settings for images and to satisfy other needs. The different contents like social context, image content, and metadata are administrated for users’ privacy preferences. According to the history available on the site previously paper implement two level frameworks to provide a best privacy for the images which will be uploaded by the users. The images uploaded are classified according to the image categories and policy prediction algorithms on which our implementation depends and generates a particular strategy for user uploaded images and characteristic.

Keywords: Content sharing sites; Metadata; Privacy; Clustering; Cluster Labeling.