The Applications Of Garbled Circuits In Watermarking And Classification Protocols

Aiming at the problem of consumers’ rights and interests in watermarking, this thesis designs a simple and secure watermarking protocol based on garbled circuit theory. Considering the privacy preserving problem of the two-party minimum Euclidean distance computation and matching protocol, this thesis puts forward a privacy-preserving two-party minimum Euclidean distance computation and matching protocol based on homomorphic cryptosystems and garbled circuit. Garbled circuit is one of the basic methods of secure two-party computation, whose routine is Boolean circuit. It allows two-party to carry out secure computation for plaintext which doesn’t use encryption.Watermarking protocol is an effective means of digital copyright protection, which plays the role of protecting buyer-seller’ rights and interests, traitor tracing and so on. Privacy-preserving two-party minimum Euclidean distance computation and matching protocol includes the related problems of distance computation, minimum selection and minimum distance, which are crucial building blocks of many cryptographic schemes such as privacy preserving biometric authentication, privacy preserving classifier learning, etc.In this thesis, based on the theory and application of garbled circuit, combined with homomorphic encryption, the research work is as follows:1) Based on garbled circuit as theory and implementation, a secure and practical buyer- seller watermarking protocol is proposed by introducing watermark server as a trading platform, which can reduce the burden of buyers and sellers, while improving the efficiency of transactions.2) According to the buyer-seller watermarking protocol which is designed in this thesis, combined with the watermark algorithm, watermarking embedding circuit and watermarking extracting circuit is designed through the basic circuits and previous work, while completing the design and implementation of the corresponding circuit embedded file and circuit extraction file.3) A privacy-preserving two-party minimum Euclidean distance computation and matching protocol is proposed. Both the server and the client’s data are encrypted. It designs and optimizes the secure two-party computation of the Euclidean distance by using homomorphic cryptosystems to be popularized and applied easily.4) A matching circuit is designed through the introduction of the previous design of minimum circuit and threshold comparison circuit, and the server and the client can complete the secure two party computations, including minus the blinding factor, threshold comparison and minimum selection.