Text Component Authentication

The pattern authentication technique mentioned in Section 2.3.6 can also be applied to the stego-text component of stego-HTML documents by adopting another data hiding technique for the text components of HTML documents. Here, the technique of hiding authentication signals of share data in between-word spaces of a stego-text component, which was proposed in Section 2.3.3 and applied in 4.2.2, is exploited again with modifications.

By specifying the relative positions between the authentication signals and the share data both in a between-word space, a data hiding technique for authentication of the share data in the stego-text component of each stego-HTML document is described in the following.

Let Hj be a stego-HTML document, K be the key of Hj, N be the specified number of bits of authentication signals hidden in a space, AHj be the share derived from Hj, AUTHEN_GEN(s, h, k) denote the procedure, proposed in Section 2.3.6, of authentication signal generation with three input parameters share data s, space size h for hiding authentication signals and a key k, Rk() be the random number generator of key k, L(x) denote the length of string x and N(x) denote the number of inter-word, interline and inter-sentence spaces in string x. The process of embedding authentication signals for authentication of share data in the stego-text component of a stego HTML document is as follows.

Algorithm 5.3 Authentication signal embedding for stego-text components.

Input: Hj, K.

Output: AHj. Steps.

1. Parse and extract stego-text component segments S1,j, …, Sm,j between two successive tags in Hj, where m is the total number of segments of Hj’s stego-text component.

2. Extract strings E1,j, …, Em,j of translated share data from S1,j, …, Sm,j, respectively.

3. Translate E1,j, …, Em,j back into share data SD1,j, …, SDm,j.

4. Concatenate SD1,j, …, SDm,j together to form a synthetic share data SDj. 5. Set a key k = RK().

6. Set authentication signals TAj = AUTHEN_GEN(SDj, m×N, k).

7. Divide TAj into m strings TA1,j, …, TAm,j. 8. For i = 1 to m, set TASDi,j=TAi,jSDi,j.

9. Translate TASD1,j, …, TASDm,j into strings AE1,j, …, AEm,j, which are made up of tab symbols, new-line symbols, and space symbols.

10. For i = 1 to m, embed

⎣

⎦

11. Set AHj=Hj.

12. For i = 1 to m, replace S1,j, …, Sm,j in AHj with AS1,j, …, ASm,j.

The corresponding authentication signal extraction and verification process is described in the following.

Algorithm 5.4 Verification of stego--text components.

Input: AHj, K.

Output: VAHj. Steps.

1. Parse and extract m stego-text component segments AS1,j, …, ASm,j between two successive tags in AHj, where m is the total number of segments of Hj’s stego-text component.

2. Extract translated strings AE1,j, …, AEm,j of share data from AS1,j, …, ASm,j, respectively.

3. Translate AE1,j, …, AEm,j back into strings TASD1,j, …, TASDm,j. 4. For i = 1 to m, separate TASDi,j into TAi,j and SDi,j.

5. Concatenate SD1,j, …, SDm,j together to form a synthetic share data SDj. 6. Set a key k = RK().

7. Set the computed authentication signals CTAj to be AUTHEN_GEN(SDj, m×N, k).

8. Concatenate TA1,j, …, TAm,j together to form the extracted authentication signals ETAj.

9. If ETAj and CTAj are not identical, mark the share data hidden in AS1,j, …, ASm,j with question marks.

Experimental Results

Some experimental results are described in this section. Figure 5.3 in the following is a stego-HTML document, in which two stego-text component segments are located in the top and the bottom, respectively, a stego-image component is in the middle-left side, a stego-video component is at the center, and a stego-flash component is in the middle-right side. Figure 5.4 is a share of the document and the display of it on browsers is identical to that of the document itself. Suppose that the share data in the share is modified. The authentication result is shown as Figure 5.5.

The source path, a dynamic link, of the image is replaced with that of the pre-determinate image and the strings of symbols, which are translated from the

corresponding share data and located at between-word spaces, are marked as a string of question marks.

Discussions and Summary

In order to ensure the fidelity and integrity of the share data of the components of stego-HTML documents, authentication signals of the share data are embedded in each stego-HTML document. For the share data in the stego-text component, authentication signals are distributed and hidden into between-word spaces by controlling the number of authentication signals hidden in per space. As for share data in the stego-non-text components, the authentication signals are embedded in the tags of each stego-HTML document by controlling the case of the letters of the tag titles and the attribute names in the tags. The process of authenticating the share data is conducted component by component.

Figure 5.3 A stego-HTML document.

Figure 5.4 A share of the HTML document in Figure 5.3.

Figure 5.5 A verified share of a tampered HTML share.

Chapter 6

Hierarchical Secret Sharing with Steganographic Effects for E-mail Documents

Introduction

Because of the population of exchanging messages via e-mail, most of users use e-mail for communication frequently. In addition, properties of e-mail for carrying any kinds of files by e-mails make it possible and convenient to put a bundle of related files in an e-mail. Therefore, applying secret sharing to e-mails that contain secret files of various types is worthwhile. And the hierarchical secret sharing is adopted and applied to secret e-mails.

In the remainder of this section, properties of e-mail are first investigated and then an overview of sharing and recovery processes is proposed. In the following sections, the detailed processes for sharing secret e-mails are described in Section 6.2, experimental results of sharing secret e-mail documents are shown and illustrated in Section 6.3, and, finally, discussions and summary are proposed.