Search

The shuffle index has been recently proposed for organizing and accessing data in outsourcing scenarios while protecting the confidentiality of the data as well as of the accesses to them. In this paper, we extend the shuffle index to the use of multiple servers for storing data, introducing a new protection technique (shadow) and enriching the original ones by operating in a distributed scenario. Our distributed shuffle index produces a significant increase in the protection of the system, with no additional costs.

In this paper we discuss the problem of collusion secure fingerprinting. In the first part of our contribution we prove the existence of equidistant codes that can be used as fingerprinting codes. Then we show that by giving algebraic structure to the equidistant code, the tracing process can be accomplished by passing a modified version of the Viterbi algorithm through the trellis representing the code.

Development of digital communications systems significantly extended possibility to perform covert communications (steganography). This recalls an emerging demand in highly efficient counter-measures, i.e. steganalysis methods. Modern steganography is presented by a broad spectrum of various data-hiding techniques. Therefore development of corresponding steganalysis methods is rather a complex problem and challenging task. Moreover, in many practical steganalysis tasks second Kerckhoff's principle is not applicable because of absence of information about the used steganography method. This motivates to use blind steganalysis, which can be applied to the certain techniques where one can specify at least statistics of the hidden data. This paper focuses on the class of supervised steganalysis techniques developed for the additive steganography, which can be described as y = f(x, s, K) = x + g(s, K), where stego image y is obtained from the cover image x by adding a low-amplitude cover image independent ((1 embedding also known as LSB matching) or cover image dependent (LSB embedding) stego signals that may be also depended on secret stego key K and the secret data s. The function g(.) represents the embedding rule.

In this paper, we analyze a possibility of reversibility of data-hiding techniques based on random binning from multimedia perspectives. We demonstrate the capabilities of unauthorized users to perform hidden data removal using solely a signal processing approach based on optimal estimation as well as consider reversibility on the side of authorized users who have the knowledge of key used for the message hiding.