Outsourcing image search services to public clouds is an ever-increasing trend. However, directly outsourcing image datasets to untrusted clouds introduces privacy concerns. Several secure image retrieval schemes have been proposed recently. However, most of them require participation from image owners when building secure indexes, which wastes many computational resources of the image owners. Several schemes are proposed to solve this problem, but they suffer from low search accuracy on large datasets. In this paper, we propose the first secure image retrieval scheme that simultaneously solves these two problems. To obtain higher search accuracy, we extract image features via fine-tuned convolutional neural networks. Then, the image features are encrypted by using the secure k-Nearest Neighbor algorithm. To improve search speed and reduce the cost of image owners, we let cloud servers locally build a secure hierarchical index graph by using the encrypted image features. Besides, the secure index can be built and updated in parallel. We provide security analysis for the proposed scheme. Performance evaluations on the CIFAR-10 dataset show that the proposed scheme is practical. Moreover, compared with a recent scheme, our scheme can save more index construction time and cost of image owners when building secure indexes.

Introduction

With the rapid development of multimedia devices, a great many images are created every day. Image retrieval is a promising technology that helps us quickly find the images we are interested in. Content-based image retrieval (CBIR) plays a vital role in image retrieval. CBIR aims to use the visual content of a query image to search similar images from an image database, which is useful in many fields, such as remote diagnosis [1], face recognition [2], and online shopping [3]. Concerning the huge storage and the complicated maintenance cost, more and more image owners prefer to outsource huge image datasets and image search services to public clouds. Then, authorized users can retrieve similar images from the clouds. Unfortunately, although the public clouds reduce the cost of the image owners, they also introduce new security threats [4]. Since the outsourced image datasets may contain sensitive information, directly outsourcing the image datasets to untrusted cloud servers may cause the sensitive images to be stolen by commercial opponents, hackers, and cloud providers. Using traditional cryptosystems to protect the images can avoid the disclosure of the sensitive information, but the public clouds can no longer provide image search services.