Enabling a Secure Match over Private Image Collections

Abstract

Image matching techniques play an essential role in many real world applications such as content based image retrieval (CBIR), computer vision, and near duplicate images. The state of the art methods are generally assumed that the content of images is not private. This reduces the utilization of these methods to work within only environments where images are publicly access. Essentially, this assumption limits more practical applications, e.g., image matching between two security agencies, where images are confidential. This paper addresses the problem of privacy-preserving image matching between two parties where images should not be revealed to each other. The descriptor set of the queried party needs to be generated and encrypted properly using a secret key at the queried party side before being transferred to the other party. We have developed a secure scheme to measure the cosine similarity between two descriptor sets without decryption. Several experiments are conducted to investigate the performance of the proposed scheme.