Conference PaperPDF Available

Highly Efficient Novel Text Steganography Algorithms

May 2015

DOI:10.1109/LISAT.2015.7160209

Conference: IEEE Long Island Systems, Applications and Technology LISAT2015
At: Long Island, NY

Authors:

Ammar Odeh

Princess Sumaya University for Technology

Khaled Elleithy

University of Bridgeport

Miad Faezipour

Purdue University West Lafayette

Eman Abdelfattah

University of Bridgeport

This paper investigates eight novel Steganography algorithms employing text file as a carrier file. The proposed model hides secret data in the text file by manipulating the font format or inserting special symbols in the text file. Furthermore, the suggested algorithms can be applied to both Unicode and ASCII code languages, regardless of the text file format. In addition, a merging capability among the techniques is introduced, which allows alternatives for users based on the system requirements. The proposed algorithms achieve a high degree of optimized Steganography attributes such as hidden ratio, robustness, and transparency.

ENCRYPTION GATES

…

. ZKA SIMULATION OUTPUT

…

. SAMPLE OF HIDDEN BITS BY USING WORD SYMBOLS

…

. MERGED SIMULATION RESULTS OF A1, A2, A4, A7

…

. COMPARISON BETWEEN EIGHT DIFFERENT ALGORITHMS

…

Figures - uploaded by Khaled Elleithy

Content may be subject to copyright.

Content uploaded by Khaled Elleithy

Content may be subject to copyright.

Ammar Odeh, Khaled Elleithy, Miad Faezipour

Department of Computer Science and Engineering

University of Bridgeport

Bridgeport, CT

aodeh@my.bridgeport.edu, elleithy@bridgeport.edu,

mfaezipo@bridgeport.edu

Eman Abdelfattah

School of Engineering & Computing Sciences

Texas A&M University

Corpus Christi, TX

Eman.Abdelfattah@tamucc.edu

Abstract—This paper investigates eight novel Steganography

algorithms employing text file as a carrier file. The proposed

model hides secret data in the text file by manipulating the font

format or inserting special symbols in the text file.

Furthermore, the suggested algorithms can be applied to both

Unicode and ASCII code languages, regardless of the text file

format. In addition, a merging capability among the techniques

is introduced, which allows alternatives for users based on the

system requirements. The proposed algorithms achieve a high

degree of optimized Steganography attributes such as hidden

ratio, robustness, and transparency.

Index Terms— HTML, Kashida, Multipoint Letter, Left

Remark, Right Remark, Steganography, Zero Width.

I. INTRODUCTION

Different strategies are used to protect sensitive data

during transmission over unsecure channels. Some

algorithms suggest changing plain text into cipher text,

which is called Cryptography [1]. On the other hand, some

algorithms are presented to protect secret information by

hiding its existence [2]. Steganography is a security

mechanism that is used to hide data inside a carrier file such

as image, sound, video, or text [3]. Secret information can be

inserted inside a carrier file in different strategies, where

each one of them has advantages and drawbacks.

The first steganography technique is injection where the

secret data is injected inside the carrier file, which increases

the carrier file size, and sometimes changes the carrier file

format [4]. The second suggested technique is substitution

where sensitive information is replaced by other data from

the carrier file [5]. The substitution method searches for bits

that have the lowest effect in the carrier file to apply the

exchange operation on them. The main advantage of this

approach is that the Stego object size is the same as the

carrier file. In addition, this approach avoids the possibility

of attacker’s suspicions. In all Steganography approaches,

the main idea is to hide the data inside the carrier file and

then to place the Stego file in some transport media. After

the data is transmitted, a Stego-analyzer might analyze the

data if there is any suspicion about the carrier file [6].

II. TECHNICAL SOLUTION AND ALGORITHMS DETAILS

This section presents eight new text Steganography

algorithms: (1) Multipoint, (2) ZKA, (3) Diacritics, (4)

KVA, (5) ZWC, and (6) Remarks (7) HTML code (8) MS

Word symbols. The proposed algorithms enhance the hidden

capacity ratio, the system’s transparency and robustness. The

diversity of the algorithms enables the users to pass their

sensitive data in an authenticated and secure manner. Each

algorithm has a different strategy that allows the user to hide

the data inside the text file. Some of these algorithms

employ the text’s font format to embed. Other algorithms

use symbol insertion technique to pass sensitive data over

the public channels. The presented algorithms can be

classified into two categories (1) Unicode languages and (2)

Multiple languages.

1. Multipoint Algorithm (A1)

Multipoint algorithm is presented and applied in Unicode

language (Arabic, Persian). These languages have some

letters called multipoint shaped letters, which contain more

than one point, unlike the English shaped letters (i, j).

Vertically or horizontally shifting techniques can be

employed to hide two bits of information in each character.

The proposed algorithm offers a highly hidden capacity ratio

compared to other algorithms. Each multipoint letter can

hide 2 bits, whereas, other algorithms can hide only one bit

per letter. The proposed algorithm enhances robustness by

changing the Stego-object to an image or a PDF file in order

to avoid the retype problem. In addition, the introduced

algorithm can be applied into other languages such as Pashto

and Urdu [7].

2. Diacritics Algorithm (A2)

Diacritics Algorithm is used to hide data inside Arabic

text [8]. This proposed algorithm uses Arabic diacritics

which are language characteristics represented by small

vowel letters. Diacritics are an optional property for any

Arabic text, and are not popularly used. Most of the Arabic

letters need diacritics to correct word pronunciation. The

proposed algorithm employs diacritics to hide two secret bits

from each diacritic, in order to offer a high hidden ratio

compared to other algorithms. The main drawback of the

Highly Efficient Novel Text Steganography

Algorithms

diacritics algorithm is that the use of diacritics is uncommon.

Thus, hackers might be more intrigued to analyze the file.

Table 1 shows a sample cover object before and after

inserting the secret data.

TABLE 1. DIACR ITICS ALGORITHM SIMULATION OUTPUT

3. Zero Width Character and Kashida Algorithm (ZKA)

(A3)

Zero Width character and Kashida are introduced through

the Unicode language (Arabic) by using special characters to

justify sentences. Kashida and Zero Width character hide

two bits between connected letters. In order to avoid any

attackers’ suspicion, a randomization algorithm utilizing

ZKA has been applied. By using this process, each message

applies different strategies to conceal secret messages. The

proposed algorithm can be extended to other Unicode

languages. Two weaknesses of ZKA are: (1) the retyping

problem, and (2) the clear format problem. These

deficiencies reduce the algorithm robustness [9]. Table 2

illustrates the changes in the carrier text and secret data

when applying the ZKA algorithm.

TABLE 2. ZKA SIMULATION OUTPUT

Cover

Object

ﺖﻧﺎآ ﻲﺘﻟا ﺔﻴﻧﺎﻣوﺮﻟا ﺔﻳرﻮﻃاﺮﺒﻣﻻاو ﺮﺼﻣ ءاﺬﻏ ﺔَﻠﺳ ﻲﻟﺎﻤﺸﻟا ﺮﺼﻣ ﻞﺣﺎﺳ نﺎآ

مﻼﺳﻹا ﻞﺒﻗ ﺮﺼﻣ ﻞﺘﺤﺗ . ﻰﻓ ﻞﻴﻨﻟا ﻩﺎﻴﻣ ﻰﻠﻋ نﻮﻳﺮﺼﻤﻟا ﺪﻤﺘﻋا ،ﻲﻟﺎﻌﻟا ﺪﺴﻟا ءﺎﻨﺑ ﻞﺒﻘﻓ

ﺔﻋارز ﻲﻓ رﺎﻄﻣﻷا ﻩﺎﻴﻣ ﻰﻠﻋ اوﺪﻤﺘﻋا ﺎﻤآ ﺎﺘﻟﺪﻟاو يداﻮﻟا ﻲﻓ ﺔﻴﻔﻴﺼﻟا تﺎﻋارﺰﻟا

ﺢﻤﻘﻟا

Stego

Object

ﻲﺘـﻟا ﺔـﻴـﻧﺎـﻣوﺮﻟا ﺔﻳرﻮـﻃاﺮﺒﻣﺎـﻟاو ﺮـﺼـﻣ ءاﺬﻏ ﺔَـﻠـﺳ ﻲﻟﺎـﻤﺸـﻟا ﺮﺼـﻣ ﻞﺣﺎﺳ نﺎـآ

مﺎـﻠـﺳﻹا ﻞـﺒـﻗ ﺮـﺼـﻣ ﻞـﺘـﺤـﺗ ﺖـﻧﺎآ . نﻮـﻳﺮﺼﻤـﻟا ﺪﻤﺘﻋا ،ﻲـﻟﺎـﻌﻟا ﺪـﺴـﻟا ءﺎﻨـﺑ ﻞﺒﻘـﻓ

اوﺪـﻤـﺘـﻋا ﺎـﻤآ ﺎـﺘﻟﺪﻟاو يداﻮـﻟا ﻲـﻓ ﺔـﻴـﻔـﻴـﺼﻟا تﺎﻋارﺰﻟا ﻰـﻓ ﻞﻴـﻨـﻟا ﻩﺎـﻴـﻣ ﻰـﻠـﻋ

ﻩﺎﻴـﻣ ﻰـﻠـﻋﺢـﻤـﻘﻟا ﺔﻋارز ﻲﻓ رﺎـﻄﻣﻷا

Secret

Data

100100100000111010110001011110110001111111001110111

110011111010000010010111000110000000101101000111111

1000000101001

4. Kashida Variation Algorithm (KVA) (A4)

Kashida variation algorithm (KVA) is presented in

Kashida according to the space position in order to hide two

bits between words. KVA is a simple application that,

typically, uses justified text over any word editors. The

proposed KVA introduces four different Kashida

information hidden scenarios. A specific scenario is applied

in each fragment to embed the secret data inside carrier file.

Furthermore, an aggregation is applied over message blocks

to reassemble the message that contains hidden information.

The benefits of the variation process within KVA is the

creation of an extra complex dimension, enhancing

robustness, and improving transparency [10]. Table 3

illustrates the cover file effect after embedding the secret

data. As demonstrated in Table 3, the semantics and the

syntax of the carrier file did not change.

TABLE 3. KVA SIMULATION OUTPUT

Cover

Object ﺖﻧﺎآ ﻲﺘﻟا ﺔﻴﻧﺎﻣوﺮﻟا ﺔﻳرﻮﻃاﺮﺒﻣﻻاو ﺮﺼﻣ ءاﺬﻏ ﺔَﻠﺳ ﻲﻟﺎﻤﺸﻟا ﺮﺼﻣ ﻞﺣﺎﺳ نﺎآ

مﻼﺳﻹا ﻞﺒﻗ ﺮﺼﻣ ﻞﺘﺤﺗ . ﻞﺒﻘﻓ ﻰﻓ ﻞﻴﻨﻟا ﻩﺎﻴﻣ ﻰﻠﻋ نﻮﻳﺮﺼﻤﻟا ﺪﻤﺘﻋا ،ﻲﻟﺎﻌﻟا ﺪﺴﻟا ءﺎﻨﺑ

ﺔﻋارز ﻲﻓ رﺎﻄﻣﻷا ﻩﺎﻴﻣ ﻰﻠﻋ اوﺪﻤﺘﻋا ﺎﻤآ ﺎﺘﻟﺪﻟاو يداﻮﻟا ﻲﻓ ﺔﻴﻔﻴﺼﻟا تﺎﻋارﺰﻟا

ﺢﻤﻘﻟا

Stego

Object ﻲﺘـﻟا ﺔـﻴـﻧﺎـﻣوﺮﻟا ﺔﻳرﻮـﻃاﺮﺒﻣﺎـﻟاو ﺮـﺼـﻣ ءاﺬﻏ ﺔَـﻠـﺳ ﻲﻟﺎـﻤﺸـﻟا ﺮﺼـﻣ ﻞﺣﺎﺳ نﺎـآ

ﻞـﺒـﻗ ﺮـﺼـﻣ ﻞـﺘـﺤـﺗ ﺖـﻧﺎآمﺎـﻠـﺳﻹا . نﻮـﻳﺮﺼﻤـﻟا ﺪﻤﺘﻋا ،ﻲـﻟﺎـﻌﻟا ﺪـﺴـﻟا ءﺎﻨـﺑ ﻞﺒﻘـﻓ

اوﺪـﻤـﺘـﻋا ﺎـﻤآ ﺎـﺘﻟﺪﻟاو يداﻮـﻟا ﻲـﻓ ﺔـﻴـﻔـﻴـﺼﻟا تﺎﻋارﺰﻟا ﻰـﻓ ﻞﻴـﻨـﻟا ﻩﺎـﻴـﻣ ﻰـﻠـﻋ

ﺢـﻤـﻘﻟا ﺔﻋارز ﻲﻓ رﺎـﻄﻣﻷا ﻩﺎﻴـﻣ ﻰـﻠـﻋ

Secret

Data 100100100000111010110001011110110001111111001110111

110011111010000010010111000110000000101101000111111

1000000101001

5. Zero Width Character Algorithm (ZWC) (A5)

Zero Width Character algorithm merges a space and Zero

Width Character to hide two bits between words inside any

document[11]. The main idea of ZWC is to introduce a

variation between two letters in order to hide data. An

advantage of ZWC is its ability to be applied to any

language. A disadvantage of ZWC is the addition of an extra

space that can be checked by most of the word editors. In

addition, retyping can destroy the robustness of the

algorithm. Table 4 exemplifies the Stego Object and the

secret data position.

TABLE 4. ZWC ALGORITHM OUTPUT

6. Remarks Steganography (A6)

Remarks Steganography Algorithm uses a text file as a

carrier to hide the data inside it[12]. The main goal of the

algorithm is to hide data inside a word file without any

changes in the file format. In Remarks Algorithm the Right-

to-Left Remark (U200F) symbol “ ” and the Left-to-Right

Remark (U200E) symbol “ ” are used to hide the bits inside

the carrier file. This method does not change the format of

the file and can also be applied to different languages,

regardless of UNICODE or ASCII coding. This method is

easily applied to the Microsoft Office Word files to hide

secret data.

7. Novel Steganography over HTML Code (A7)

In the HTML Steganography technique [13],

Cryptography and Steganography techniques are used to

Cover

Object ﱠﻞ َﺘ ْﺤ َﺗ ْﺖ َﻧ ﺎ َآ ﻲِﺘﱠﻟا َﺔ ﱠﻴ ِﻧ ﺎ ﻣ و ﱡﺮ ﻟ ا ﺔﻳرﻮﻃاﺮﺒﻣﻻاو ٍﺮ ْﺼ ِﻣ ُء ا َﺬ ِﻏ ﺔَﻠﺳ ِﻲ ِﻟ ﺎ َﻤ ﱠﺸ ﻟ ا ِﺮ ْﺼ ِﻣ ُﻞ ِﺣ ﺎ َﺳ َن ﺎ َآ

ِم ﻼ ْﺳ ِﺈ ْﻟ ا َﻞ ْ ﺒ َﻗ ًﺮ ْﺼ ِﻣ . ﻰﻓ ِﻞ ْﻴ ﱠﻨ ﻟ ا ِﻩ ﺎ َﻴ ِﻣ ﻰَﻠَﻋ َن ﻮ ﱡﻳ ِﺮ ْ ﺼ ِﻤ ْﻟ ا َﺪ َﻤ َﺘ ْﻋ ِا ، ﻲِﻟﺎَﻌْﻟا ﱢﺪ ﱠﺴ ﻟ ا ُء ﺎ َﻨ ِﺑ َﻞ ْﺒ َﻘ َﻓ

ِﺢ ْﻤ َﻘ ْ ﻟ ا ِﺔ َﻋ ا ر ِز ﻲِﻓ ِرﺎَﻄْﻣَﺄْﻟا ِﻩ ﺎ َﻴ ِﻣ ﻰَﻠَﻋ ا و ُﺪ َﻤ َﺘ ْﻋ ِا ﺎَﻤَآ ﺎﺘْﻟﱢﺪﻟاَو يِداَﻮْﻟا ﻲِﻓ ُﺔ ﱠﻴ ِﻔ ْ ﻴ ﱠﺼ ﻟ ا ُتﺎَﻋارﱢﺰﻟا

Stego

Object َﻤ ﱠﺸ ﻟ ا ِﺮ ْ ﺼ ِﻣ ُﻞ ِﺣ ﺎ َﺳ َن ﺎ َآ ﱠﻞ َﺘ ْﺤ َﺗ ْﺖ َﻧ ﺎ َآ ﻲِﺘﱠﻟا َﺔ ﱠﻴ ِﻧ ﺎ ﻣ و ﱡﺮ ﻟ ا ﺔﻳرﻮﻃاﺮﺒﻣﻻاو ٍﺮ ْﺼ ِﻣ ُء ا َﺬ ِﻏ ﺔَﻠﺳ ِﻲ ِﻟ ﺎ

ﺎﺘْﻟﱢﺪﻟاَو يِداَﻮْﻟا ﻲِﻓ ُﺔ ﱠﻴ ِﻔ ْ ﻴ ﱠﺼ ﻟ ا ُتﺎَﻋارﱢﺰﻟا ِﺢ ْﻤ َﻘ ْ ﻟ ا ِﺔ َﻋ ا ر ِز ﻲِﻓ ِرﺎَﻄْﻣَﺄْﻟا ِﻩ ﺎ َﻴ ِﻣ ﻰَﻠَﻋ ا و ُﺪ َﻤ َﺘ ْﻋ ِا ﺎَﻤَآ

Secret

Bits 1001001000001110101100010111101100011111110011101111

1001111101000001001011100011000000010110100011111110

00000101001

Cover

Object ﺖﻧﺎآ ﻲﺘﻟا ﺔﻴﻧﺎﻣوﺮﻟا ﺔﻳرﻮﻃاﺮﺒﻣﻻاو ﺮﺼﻣ ءاﺬﻏ ﺔَﻠﺳ ﻲﻟﺎﻤﺸﻟا ﺮﺼﻣ ﻞﺣﺎﺳ نﺎآ

مﻼﺳﻹا ﻞﺒﻗ ﺮﺼﻣ ﻞﺘﺤﺗ . ﻰﻓ ﻞﻴﻨﻟا ﻩﺎﻴﻣ ﻰﻠﻋ نﻮﻳﺮﺼﻤﻟا ﺪﻤﺘﻋا ،ﻲﻟﺎﻌﻟا ﺪﺴﻟا ءﺎﻨﺑ ﻞﺒﻘﻓ

ﺔﻋارز ﻲﻓ رﺎﻄﻣﻷا ﻩﺎﻴﻣ ﻰﻠﻋ اوﺪﻤﺘﻋا ﺎﻤآ ﺎﺘﻟﺪﻟاو يداﻮﻟا ﻲﻓ ﺔﻴﻔﻴﺼﻟا تﺎﻋارﺰﻟا

ﺢﻤﻘﻟا.

Stego

Object ﻞﺣﺎﺳ نﺎـآ ﻲﺘـﻟا ﺔـﻴـﻧﺎـﻣوﺮﻟا ﺔﻳرﻮـﻃاﺮﺒﻣﺎـﻟاو ﺮـﺼـﻣ ءاﺬﻏ ﺔَـﻠـﺳ ﻲﻟﺎـﻤﺸـﻟا ﺮﺼـﻣ

مﺎـﻠـﺳﻹا ﻞـﺒـﻗ ﺮـﺼـﻣ ﻞـﺘـﺤـﺗ ﺖـﻧﺎآ . نﻮـﻳﺮﺼﻤـﻟا ﺪﻤﺘﻋا ،ﻲـﻟﺎـﻌﻟا ﺪـﺴـﻟا ءﺎﻨـﺑ ﻞﺒﻘـﻓ

اوﺪـﻤـﺘـﻋا ﺎـﻤآ ﺎـﺘﻟﺪﻟاو يداﻮـﻟا ﻲـﻓ ﺔـﻴـﻔـﻴـﺼﻟا تﺎﻋارﺰﻟا ﻰـﻓ ﻞﻴـﻨـﻟا ﻩﺎـﻴـﻣ ﻰـﻠـﻋ

ﻓ رﺎـﻄﻣﻷا ﻩﺎﻴـﻣ ﻰـﻠـﻋﺢـﻤـﻘﻟا ﺔﻋارز ﻲ

Secret

Bits

Positions

pass secure information. Webpages are used as the carrier

for secret data while they are published over the Internet.

Authenticated users can access the hidden data. The

proposed algorithm consists of three main steps, where the

first and third steps represent inverse operations.

The Conceal operation consists of six stages:

1. The Statistical Stage consists of an array of 26

elements that count the characters' frequency. The

frequency array can be increased or decreased based

on a webpage’s language. Our experiments are based

on the English language.

2. The Character Representation Stage assigns “bits”

based on the usage frequency of the character. For

example, after the frequency array has been generated,

the lowest two frequently found characters could be

represented by one bit. If two characters have the

same frequency, the character order specifies which

one is zero. For example, if letter ‘X’ appears 10 times

and a letter ‘Z’ appears 6 times, then ‘Z’ is

represented as ‘0’, and ‘X’ as ‘1’. Moreover, if both

letters have the same occurrence number, then ‘X’ is

represented as ‘0’ and ‘Z’ is represented as ‘1’.

Similarly, the next four characters can be represented

by two bits.

3. The Embedding Stage occurs when the secret bits are

embedded after the character representation equals ‘8’

bits. In other words, if the first character

representation is ‘0’ and the secret bits are “0111011”,

the code will be “00111011”.

4. The Encryption Stage: consists of three simple binary

operations. First, the binary representation is

complemented. Then, “exclusive OR” (XOR) is

performed with the binary key. In order to produce the

output of XOR gate, shift left must be applied by one

bit and re-entered as XOR input. In addition, the

binary key code creation depends on the webpage

index where each page has a rear index. This

operation is repeated twice as shown in Figure 1.

FIGURE 1. ENCRYPTION GATES

A numerical example is presented below where the input

is (C) 01000011, and the key is 10001100:

Step 1:- Binary representation for C =>01000011

Step 2: - 1’s complement of C => 10111100

Step 3:- 10001100 XOR 10111100=>00110000

Step 4:-Shift left 00110000 => 01100000

Step 5:- 10001100 XOR 01100000=>11101100

Step 6:- Shift left 11101100=>11011001

5. The Decoding Stage converts the binary code to ASCII

code. Here, the numerical form is changed into text. In

the example, shown above “11011001” is decoded to

(Ù).

6. The Insertion Stage occurs when the text from the

decoding stage is produced. Then, the text is placed into

the webpage code as comment. The comments do not

appear in the webpage’s output view.

To reveal the secret message the following procedural

stages are applied:-

1. The Statistical Stage relies on 26 characters, similar to

those found in the “Conceal Operation” statistical stage.

2. The Reading Comments Stage allows the user to read

comments from the webpage.

3. The Encoding Stage converts any comment appearing in

the text into binary representation.

4. The Decryption Stage follows the same binary process

that exists in the “Conceal Operation” of the encryption

stage.

The Reveal Character Code Stage is performed by using

the frequency array created in the statistical stage, and

comparing it with the binary output of the decryption stage.

The embedded information is obtained by removing the

character representation.

8. Steganography in Text by Using MS Word Symbols (A8)

The MS Word Symbols Algorithm hides data inside a

Word file without altering the carrier file properties such as

file size, content and format[14]. The MS Word Symbols

Algorithm employs some non-printable symbols to hide four

bits between letters. This process improves the hidden

capacity ratio compared to other algorithms. Moreover, no

modification in the Word format file or letter shapes would

be made. Thus, the suggested algorithm avoids raising any

hackers’ suspicions.

Table 5 represents some of the hidden codes. For

example, if we insert all four-table variation symbols after

each letter, then the passing bits code is 0000. This insertion

enables us to hide four bits of secret data. The four symbols

applied are Right remark (200E), Left remark (200F), Zero

width joiner (200D), and Zero width non-joiner (200C). In

this technique, different variations can be used to represent

hidden bits for a total of 16 different codes.

Three inputs are used in the hiding process: the Stego key,

carrier file, and hidden data. The main purpose of Stego key

is to change the symbols’ bit representation. For example,

‘1’ represents a bit’s absence, and ‘0’ represents a bit’s

presence. In the next step, a symbols’ table is created

depending on the outcome of the Stego key.

TABLE 5. SAMPLE OF HIDDEN BITS BY USING WORD SYMBOLS

Right

Remark

Left

Remark

ZWJ ZWNJ Hidden

code

X X X X 0000

X X X 0001

X X 0101

X X 0011

X 0111

1111

X 1101

III. PERFORMANCE EVALUATION

Table 7 shows the proposed algorithms and their

classifications. Multipoint, Diacritics, ZKA, and KVA

algorithms are language dependent. These algorithms (A1,

A2, A3, and A4) have the potential to hide secret data inside

Unicode Languages such as Arabic, Persian, and Urdu.

Multipoint and Diacritics algorithms hide data by

substituting one letter with another modified shaped letter.

This approach works well, since there is no standard position

or distance between the letter and the points. Also, this

approach has proven that there is no standard position

between the letter and the Diacritics. The main advantage of

the A1 and A2 algorithms is the file’s size stability.

Moreover, A1 and A2 have the ability to combine with A3,

A4, A5, A6, A7, and A8 in order to improve the hidden ratio

capacity. A3 and A4 hide the data inside a carrier object by

inserting non-printed symbols. The main drawback of A3

and A4 is that the carrier file’s size increases with the

process of embedding the secret message.

A5, A6, and A8 hide the data by embedding non-printed

symbols. In addition, A5, A6, and A8 provide a highly

hidden ratio capacity. Moreover, these algorithms have the

ability to be combined with A1 and A2 in order to improve

the hidden ratio capacity. One disadvantage associated with

A5, A6, and A8 is that the carrier file size increases through

the insertion of non-printed symbols. HTML Code A7

algorithm hides the encrypted secret data as comments

inside the HTML Webpage. Limited hidden bits can be

inserted based on the webpage’s original language. The

hidden capacity of HTML Code A7 algorithm can be

enhanced by combing it with A1 and/or A2. Table 8 explains

the results of combining one algorithm with another

algorithm and the effect on the hidden ratio capacity and the

file size. It is notable that combing one algorithm with

another will increase the hidden ratio capacity at a risk of

increasing the file size. The only case that the file size does

not increase is when A1 and A2 are combined. Here, either

A1 or A2 did not increase the carrier file size. These

algorithms are only based on changing the points or the

Diacritics positions.

FIGURE 2. HIDDEN RATIO [BIT/KBYTE]

Figure 2 illustrates and compares the hidden ratio of eight

proposed algorithms and Line and Word Shift algorithms

introduced in literature. Equation 1, stated below, represents

the hidden ratio capacity (HR). Based on the simulation

results, A8 achieves the highest hidden ratio range. A6

achieves the second highest hidden ratio range. A7

represents a constant hidden ratio range. The average range

of the hidden ratio of A1 = 104 bits per Kbyte (KB); A2 =

144b/KB; A3 = 153b/KB; A4 =355b/KB; A5 = 153 b/KB;

A6 = 626b/KB; A7 = 127 b/KB; and A8 = 799 b/KB.

     

    1

FIGURE 3. FILE SIZE EFFECT AFTER INSERTING SECRET DATA

Figure 3 represents and compares the file size of the Stego

objects that increase the hidden ratio capacity of eight

proposed algorithms with Line and Word Shift algorithms.

Based on the simulation results for the hidden ratio range,

there is no change in the carrier file size when each A1, A2,

and A7 is applied. On other hand, the highest file increase

ratio appears in A8. A6 offers the second highest file size

increase.

200

400

600

800

1000

Line Shift

Word Shift

Aljazeera.net

BBC.COM

CNN

Alhayat

0.2

0.4

0.6

0.8

1.2

1.4

1.6

Line Shift

Word Shift

Aljazeera.net

BBC.COM

CNN

Alhayat

IV. MERGING ALGORITHMS

In this section, we discuss the possibility of merging more

than one algorithm. The proposed algorithms can be merged

collectively to further improve the hidden ratio range of the

carrier file. For example, A1 can be combined with either

A2, A3, A4, A5, A6, A7 or A8. Figure 4 illustrates the

hidden ratio output when A1 is combined with other

algorithms. Figure 5 represents the change in the carrier file

size, once other algorithms have been merged with A1. The

merging process will be done in sequence. For example, the

user can select A1 algorithm. The carrier file will be updated

based on the secret data, and then the user can get the output

file of the A1 algorithm and select the second suggested

algorithm.

FIGURE 4.HIDDEN RATIO MERGED WITH MULTIPOINT

FIGURE 5. FILE SIZE CHANGE MERGED WITH MULTIPOINT

Based on the simulation results, the best outcome will

depend on the individual user requirements. If the file size is

not allowed to change, then the user will select A1, A2, or

A7. Otherwise, if the carrier file size is allowed to increase,

and large amounts of secret data are being embedded in the

carrier file, then A8 will be the best algorithm choice. One

distinctive merging algorithm scenario is (A1, A2, A4, A7)

Multipoint, Diacritics, KVA, and HTML Code. This

scenario provides a highly hidden ratio while the size of the

carrier file will minimally increase, as shown in Table 6. The

results show that the merging scenario is very consistent

among different Websites in terms of hidden ratio capacity

and the file size change.

TABLE 6 . MERGED SIMU LATION RESULTS OF A1, A2, A4, A7

Webpage Hidden Ratio b/KB File size Change

Aljazeera 723 2%

CNN 709 1%

BBC 717 2%

Alhayat 771 2%

V. CONCLUSION

In this paper, we have discussed and compared the

performance of eight new text Steganography Algorithms.

Multipoint, ZKA, Diacritics, KVA, ZWC, Remarks, HTML

code, and MS Word Symbols Algorithms. These algorithms

provide additional and necessary protection of confidential

data. KVA, ZKA, Multipoint, and Diacritics Algorithms are

applied to specific Unicode based languages such as, Arabic,

Persian, Pashto, and Urdu. Hence, these particular

algorithms could potentially benefit, approximately, a

population of two billion people. Whereas, ZWC Remarks,

HTML code, and MS Word Symbols Algorithms are

language independent, and can be used to pass sensitive

information by inserting specific symbols.

We have presented and discussed the possibility and

effect of merging several algorithms together. The choice

between different merging scenarios depends on the user’s

constraints. The scenarios we provided work well in

situations where there is either a constraint or no constraint

placed on the entire file. Furthermore, we presented tradeoff

scenarios between the hidden ratio and the increase of the

file size.

The presented algorithms can be used to establish safe

communication, privacy enhancement, secure data sharing,

and additional protection of copyrighted products. For

example, our algorithms can be applied to multimedia and

publishing products. By using any of the eight referenced

algorithms or merging among them, users will have a highly,

secured data system in place to prevent and curtail hackers.

200

400

600

800

1000

1200

and

Hidden Ratio/ Merge

with Multipoint

Aljazeera.net

BBC.COM

CNN

Alhayat

50%

100%

150%

200%

and

File size change /Merge

with Multipoint

Aljazeera.net

BBC.COM

CNN

Alhayat

REFERENCES

[1] M. Alfred, V. O. Paul, and V. Scott, Handbook of applied

cryptography: CRC press, 2010.

[2] J. Neil and J. Sushil, "Exploring steganography: Seeing the

unseen," Computer, vol. 31, pp. 26-34, 1998.

[3] V. Potdar and E. Chang, "Visibly Invisible: Ciphertext as a

Steganographic Carrier," in Proceedings of the 4th

International Network Conference (INC2004), 2004, pp. 385-

391.

[4] E. Martiri, A. Baxhaku, and E. Barolli, "Steganographic

Algorithm Injection in Image Information Systems Used in

Healthcare Organizations," in 2011 Third International

Conference on Intelligent Networking and Collaborative

Systems (INCoS), 2011, pp. 408-411.

[5] M. Zamani, A. Manaf, and R. Ahmad, "Knots of Substitution

Techniques of Audio Steganography," in The 2009

International Conference on Telecom Technology and

Applications, Singapore, 2009, pp. 415-419.

[6] I. Avcibas, N. Memon, and B. Sankur, "Steganalysis using

image quality metrics," IEEE Transactions on Image

Processing, vol. 12, pp. 221-229, 2003.

[7] A. Odeh, A. Alzubi, Q. B. Hani, and K. Elleithy,

"Steganography by multipoint Arabic letters," in Systems,

Applications and Technology Conference (LISAT), 2012 IEEE

Long Island, Farmingdale State College - State University of

New York, 2012, pp. 1-7.

[8] A. Odeh and K. Elleithy, "Steganography in Arabic Text

Using Full Diacritics Text," presented at the 25th International

Conference on Computers and Their Applications in Industry

and Engineering (CAINE-2012), New Orleans, Louisiana,

USA, 2012.

[9] A. Odeh and K. Elleithy, "Steganography in Arabic Text

Using Zero Width and Kashidha Letters," International

Journal of Computer Science & Information Technology

(IJCSIT), vol. 4, 2012.

[10] A. Odeh, K. Elleithy, and M. Faezipour, "Steganography in

Arabic text using Kashida variation algorithm (KVA)," in

Systems, Applications and Technology Conference (LISAT),

2013 IEEE Long Island, 2013, pp. 1-6.

[11] A. Odeh and K. Elleithy, "Steganography in Text by Merge

ZWC and Space Character," in 28th International Conference

on Computers and Their Applications, CATA-2013, Honolulu,

Hawaii, USA, 2013, pp. 1-7.

[12] A. Odeh, K. Elleithy, and M. Faezipour, "Text Steganography

Using Language Remarks," in Prcoceeding of the American

Society of Engineering Education, pp. 1-7, 2013.

[13] A. Odeh, K. Elleithy, M. Faezipour, and E. Abdelfattah,

"Novel Steganography over HTML Code," in New Trends in

Networking, Computing, Informatics, Systems Sciences, and

Engineering, T. Sobh and E. Khaled, Eds., ed: Springer, 2014.

[14] A. Odeh, K. Elleithy, and M. Faezipour, "Steganography in

text by using MS word symbols," in 2014 Zone 1 Conference

of the American Society for Engineering Education (ASEE

Zone 1), CT, USA, 2014, pp. 1-5.

TABLE 7. COMPARISON BETWEEN EIGHT DIFFERENT ALGORITHMS

Algorithm Applicable Language General Categories Technique File size Ability to combine with other

Algorithm(s)

A1:Multipoint Unicode(Arabic, Persian, Urdu) Substitution Linguistic No Effect (A2,A3,A4,A5,A6)

A2:Diacritics Unicode(Arabic, Persian, Urdu) Substitution Linguistic No Effect (A1,A3,A4,A5,A6)

A3:ZKS Unicode(Arabic, Persian, Urdu) Injection Linguistic Increase Can be combined with (A1,A2)

A4:KVA Unicode(Arabic, Persian, Urdu) Injection Linguistic Increase Can be combined with (A2)

A5:ZWC Language Independent Injection Format Increase Can be combined with (A1,A2)

A6:Remarks Language Independent Injection Format Increase Can be combined with (A1,A2)

A7: HTML Code Language Independent Injection Format No Effect Can be combined with (A1,A2)

A8: MS Word Language Independent Injection Format Increase Can be combined with (A1,A2)

TABLE 8. OUTPUT OF THE COMBINING PROCESS BETWEEN PROPOSED ALGORITHMS

Algorithm Applicable Language General Categories Technique Hidden Capacity File Size

A1(A3,A4,A5,A6) Unicode(Arabic, Persian,