Secure Image Processing

OF THESIS SECURE IMAGE PROCESSING In todays heterogeneous network environment, there is a growing demand for distrusted parties to jointly execute distributed algorithms on private data whose secrecy needed to be safeguarded. Platforms that support such computation on image processing purposes are called secure image processing protocols. In this thesis, we propose a new security model, called quasi information theoretic (QIT) security. Under the proposed model efficient protocols on two basic image processing algorithms – linear filtering and thresholding – are developed. For both problems we consider two situations: 1) only two parties are involved where one holds the data and the other possesses the processing algorithm; 2) an additional non-colluding third party exists. Experiments show that our proposed protocols improved the computational time significantly compared with the classical cryptographical couterparts as well as providing reasonable amount of security as proved in the thesis.