Abstract—
In order to protect data privacy, image with sensitive or
private information needs to be encrypted before being outsourced to
the cloud. However, this causes difficulties in image retrieval and data
management. A secure image retrieval method based on orthogonal
decomposition is proposed in the paper. The image is divided into two
different components, for which encryption and feature extraction are
executed separately. As a result, cloud server can extract features from
an encrypted image directly and compare them with the features of the
queried images, so that the user can thus obtain the image. Different
from other methods, the proposed method has no special requirements
to encryption algorithms. Experimental results prove that the proposed
method can achieve better security and better retrieval precision.
Keywords—
Secure image retrieval, secure search, orthogonal
decomposition, secure cloud computing.
I. I
NTRODUCTION
ITH the rapid development of cloud service, storing
images in cloud servers is becoming more and more
popular. In order to protect data privacy and to allow restrict
access, sensitive images need to be encrypted before being
uploaded to cloud server. However, if users want to retrieve
images from server, encrypted image need to be decrypted first,
then retrieval can be operated on plaintext, which makes the
sensitive information being exposed to servers which breaks
privacy and hence is not desired. Therefore it is important to
develop technologies of image retrieval over encrypted domain.
Currently the information retrieval on encrypted domain
mainly focuses on text document. Song et al. [1] proposed a
ciphertext scanning method based on streaming cipher to make
sure whether the search term is existed in the ciphertext. Boneh
et al. [2] proposed a keyword search method based on
public-key encryption, where the server can identify whether
messages encrypted by user’s public key contain some specific
keyword, but learn nothing else. Swaminathan et al. [3]
explored techniques to securely rank-order the documents and
extracted the most relevant document(s) from an encrypted
collection based on the encrypted search queries. Wang el al.
[4] utilized a new crypto primitive called the Order-Preserving
Symmetric Encryption (OPSE) to achieve both security and
privacy-preserving, although the guarantee to security could be
weakened by it. Cao et al. [5] proposed privacy-preserving
Yanyan Xu is with the LIESMARS, Wuhan University, Wuhan, China,
430079. (86-27-68771665, e-mail: xuyy@whu.edu.cn).
Lizhi Xiong and Zhengquan Xu are with the LIESMARS, Wuhan
University, Wuhan, China, 430079. (e-mail: xlzwhucs@gmail.com,
xuzq@whu.edu.cn).
Li Jiang is with the School of Information Engineering, Zhengzhou
University, Zhengzhou, China, 450001(e-mail: yigutong@163.com).
multi-keyword ranked search over encrypted data.
However, these techniques cannot be applied to
content-based image retrieval directly, because effective image
retrieval typically relies on comparing the distance of image
features, but encrypted data fails to preserve the distance
between feature vectors if the employed cryptographic
primitive are not designed especially for intended goals [6].
Only recently, secure text document search in the encrypted
domain has been extended to secure image retrieval research.
Lu et al. [7] proposed three schemes to solve the problem of
image retrieval over encrypted domain, including bit-plane
randomization, random projection, and randomized unary
encoding. Although it is efficient, the security has been
compromised. Karthik et al. [8] presented a transparent privacy
preserving hashing scheme tailored to preserve the DCT-AC
coefficient distributions. But the search engine is insensitive to
shapes and descriptions of both natural and artificial objects
due to missing space-frequency information. In addition, its
security is compromised because the constrained shuffling is
used on part of AC coefficients. Hsu et al. [9] proposed a
homomorphic encryption-based secure SIFT for image feature
extraction, but the size of cipher-text is expanded and the
computing is laborious. It should be noted that the above
research results are all relying on specific encryption methods,
such as shuffling, homomorphic encryption, etc., which
preserve the distance of image features after images are
encrypted and make the image retrieval in encrypted domain
possible. However, these schemes limit the universality of the
method. For example, in some situations that have high
requirements to security, these methods are not suitable.
Aiming at solving these problems, a secure image retrieval
method based on orthogonal decomposition is proposed in this
paper. The image is divided into encryption field and feature
extraction field by orthogonal decomposition, where
encryption operation and feature extraction can be executed
separately. Servers can get image features of encrypted image
directly without decrypt it, and compare with features of
queried image; the one with the closest distance is the retrieved
image. Different from other methods, the proposed method has
no specific requirements of cipher algorithms. Experimental
results show that the proposed scheme has good encryption
security and can achieve better retrieval precision.
The organization of this paper is as follows: Section II
discusses the related research, and Section III proposes our
scheme. Section IV provides experimental results and a
performance analysis, and Section V presents conclusions.
Secure Image Retrieval Based On Orthogonal
Decomposition under Cloud Environment
World Academy of Science, Engineering and Technology
International Journal of Computer, Control, Quantum and Information Engineering Vol:9, No:5, 2015
743International Scholarly and Scientific Research & Innovation 9(5) 2015
International Science Index Vol:9, No:5, 2015 waset.org/Publication/10001191