Privacy-Preserving Content-Based Image Retrieval in Cloud Computing With EMD

Abstract

Content based image retrieval (CBIR) system plays a vital role in real world image retrieval applications. CBIR is developed to retrieve desired featured image. CBIR is limited because it has been used widely for its computational requirements and security purposes. In this work, we develop a privacy-preserving content based image retrieval design in which the data owner is allowed to obtain the content based image which can be stored in the cloud. Earth Mover’s Distance is utilized to measure the similarity of the viewed images. EMD can be calculated by solving transportation problem and linear programming problem. The proposed design supports local-feature based CBIR with EMD such a way that the cloud can find a solution without experiencing the personal information. This proposed system provides enhanced security and efficient searching.