Security Analysis on Privacy-Secure Image Trading Framework Using Blind Watermarking

We propose and evaluate a privacy-secure image trading framework using blind fingerprinting technique which embeds a user ID using watermark technique. The framework provides a privacy-secure trading even though there is no confidential relation between a content provider and a user. In a conventional fingerprinting scheme, a content provider offers a user ID embedded image to a user in order to trace a pirated piece and an illegal user when an image is pirated. In spite the fact that privacy protection becomes an important issue along with augmentation of digital content trading, it has not been well considered in a conventional fingerprinting scheme. Some blind fingerprinting techniques satisfy privacy protection by concealing user information using cryptography. Semi-blind fingerprinting is another type of blind technique that is proposed with a concept of blinding information within feasible processing cost. Feasibility is satisfied by avoiding use of cryptography which causes heavy overhead in computation. In our proposed image trading framework, we use semi-blind technique which is well optimized for our framework. In order to establish a robust framework, we carry out security analysis by specifying various kinds of incidents and solutions.