Journal of Applied Sciences ›› 2018, Vol. 36 ›› Issue (4): 628-634.doi: 10.3969/j.issn.0255-8297.2018.04.006

• Signal and Information Processing • Previous Articles     Next Articles

Encrypted JPEG Image Retrieval Based on Dynamic BoW Model

WEI Qiu-han, LIANG Hai-hua, ZHANG Xin-peng   

  1. School of Communication and Information Engineering, Shanghai University, Shanghai 200444, China
  • Received:2017-02-21 Revised:2017-06-03 Online:2018-07-31 Published:2018-07-31

Abstract: This paper presents an efficient and dynamic retrieval scheme for encrypted JPEG images, which is based on the histogram of AC coefficients. First, the robust bagof-word (BoW) model is constructed in the encrypted domain via the advanced k-means clustering. Second, the local features of images are transformed into the global statistical histograms by the dynamic BoW model. Finally, the cloud server uses the histograms of the encrypted images to judge similarity. Besides, dynamic update operations like insertion and deletion of images are also available, so the retrieval scheme is still robust and efficient. The results of experiments show that the proposed scheme can improve the retrieval efficiency and ensure the accuracy of retrieval at the same time, having practical application value.

Key words: encrypted JPEG image retrieval, adaptive k-means clustering, dynamic update, bag-of-word model

CLC Number: