Title: Visual vocabulary tree-based partial-duplicate image retrieval for coverless image steganography

Authors: Yan Mu; Zhili Zhou

Addresses: School of Computer and Software and Jiangsu Engineering Center of Network Monitoring, Nanjing University of Information Science and Technology, Nanjing 210044, China ' School of Computer and Software and Jiangsu Engineering Center of Network Monitoring, Nanjing University of Information Science and Technology, Nanjing 210044, China

Abstract: The traditional image steganographic approaches embed the secret message into covers by modifying their contents. Therefore, the modification traces left in the cover will cause some damages to the cover, especially embedding more messages in the cover. More importantly, the modification traces make successful steganalysis possible. In this paper, visual vocabulary tree-based partial-duplicate image retrieval for coverless image steganography is proposed to embed the secret messages without any modification. The main idea of our method is to retrieve a set of duplicates of a given secret image as stego-images from a natural image database. The images in the database will be divided into a number of image patches, and then indexed by the features extracted from the image patches. We search for the duplicates of the secret image in the image database to obtain the stego-images. Each of these stego-images shares one similar image patch with the secret image. When receiver obtains those stego-images, our method can recover the secret image approximately by using the designed protocols. Experimental results show that our method not only resists the existing steganalysis tools, but also has high capacity.

Keywords: coverless image steganography; robust hashing algorithm; vocabulary tree; image retrieval; stego-image; image database; high capacity.

DOI: 10.1504/IJHPCN.2019.102133

International Journal of High Performance Computing and Networking, 2019 Vol.14 No.3, pp.333 - 341

Received: 22 Aug 2017
Accepted: 02 Dec 2017
Published online: 09 Sep 2019 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article