XORSTEG: A New Model of Text

Steganography
Tapodhir Acharjee, Ashish Konwar, Rakesh Kumar Ram,Rahul Sharma, Dhrubajyoti Goswami
Department of Computer Science and Engineering
Assam University, Silchar, Assam, India
[email protected], [email protected], [email protected], [email protected], [email protected]

Abstract—Staganography is the practice of hiding secret starting and ending letter. Our proposed model overcomes this
information using texts, images and audio as the cover media. limitation which increases the efficiency of the approach.
But, using text as the target medium is not as easy as compared to
the other target media, because of the lack of available redundant The rest of this paper is organised as follows: Section II
information in a text file.This paper proposes a new model of
discusses the existing method and its limitations. Section III
key-based text steganography. Various methods in this type of
key-based steganography do not work when the cover file describes the proposed approach. Section IV shows the results
contained single letters and words which have the same starting and compares our proposed approach with the existing
and ending letter. The proposed approach works on this approach. The paper concludes with the conclusion in Section
limitation and conceals a message, without degrading cover, by V.
using XOR operation on the start and end letter of words of the
cover.
II. RELATED STUDY
Keywords—steganography; text steganography; key-based text In this section we discuss an existing text steganography
steganography; cover file method[4] which uses any paragraph and hides the secret
message in the paragraph (cover file) thus generating a stego
I. INTRODUCTION key; without degrading the originality of the paragraph .
Steganography is a word stemming from the ancient Greek
with the meaning of covered writing. Steganography uses the The model works by hiding a message using start and end
help of a cover file (text, image, audio,video)to hide a secret letter of the words of a cover file. It works on the binary value
message. The secret message is embedded with the cover of a character. After converting the cipher text to a stream of
message by using various methods and is then sent to the bits,each bit is hidden by picking a word from the cover file
intended recipient. In steganography the text to be concealed and using either the start or the end letter of that word
is called embedded data. An innocuous medium, such as a depending on the bit to be concealed.Bit 0 or 1 is hidden by
text, audio, image, etc. used to hide the message is called the reading a word, sequentially, from the cover file and including
cover file. The key which is optional and is used in the the starting letter or the end letter, respectively, of the word in
embedding process is called a stego key. A stego key is used the stego key. A word having same start and end letter is
to control the hiding process so as to restrict detection from skipped. Since no change is made to the cover, the cover file
third parties and/or recovery of embedded data to the parties and its corresponding stego file are exactly the same [4].
who know it [1, 4].

Steganography which uses text as a cover medium is termed

as text steganography. Text steganography can involve
A. THE HIDING ALGORITHM
anything from changing the formatting of an existing text, to
changing words within a text, to generating random character
sequences or using context-free grammars to generate readable 1. Get a cover file.
texts[5]. The formatting of a text can easily draw attention 2. Convert the input file to its binary equivalent (bin).
from third parties if the original plaintext is available [5]. 3. Read a bit (x) from the bin.
4. Read a word from the cover file and write it in stego file.
This paper presents an approach of text steganography which 5. If start and end letter of the word is same, then read the next
does not change the format of the cover text and hides the word of the cover file andwrite it in the stego file.
secret message in the cover file by generating a stego key. A 6. s = start letter of the word and e = end letter of the word.
method for this type of key based steganography is proposed 7. If x = 0, write s in the stego key.
in [4]. But it is limited in the way that it cannot hide the 8. Else if x = 1, write e in the stego key.
message when the cover file contains letters having same 9. Repeat steps 3 to 8 till the end of the bin file.
10. Send the stego file and the stego key to the receiver. is written in the key file else the smallest decimal value is
written. The steps are:
B. THE SEEKING ALGORITHM
1. Take the secret message and convert it into
equivalent binary.
1. Read a character (c) from the stego key. 2. Take a normal text file as the cover file.
2. Read a word from the stego file. 3. Read a line from the cover file and copy it to the
3. If start and end letter of the word is same, then skip that stego file.
word and read the next wordfrom the stego file. 4. Take the first and last word from the line.
4. Get the start letter (s) and end letter (e) of the word. 5. Take the first and last letter from the first word.
5. If c = s, then bit b = 0. 6. Convert the letters into their binary equivalent.
6. Else if c = e, then bit b = 1. 7. Take XOR between the values.
7. Write b in a file. 8. Perform steps (5)-(7) for the last word of that line.
8. Execute above steps repeatedly till the end of the stego key. 9. Compare both XOR results (If both are same then
9. Convert the file into its character equivalent. take complement oflast XOR result; else: Change
binary to equivalent decimal.)
The main limitation of this approach is that this method cannot 10. Take a bit from the converted binary message.
hide bits of a message in words which have the same starting 11. If bit is 1 then write greater decimal in the key file
and ending letter. This can greatly reduce the capacity of the else write the smallerdecimal.
cover file to contain a message, if the cover file contains many 12. Take the next words from both side of that line.
words which have the same starting and ending letters. Also, it 13. Perform steps (5)-(11) until the reading of the words
can be very easy for third parties to find patterns from the in that line is finished.
stego key as the stego key is only made up of the starting and 14. Perform steps (5)-(13) for the next lines of the stego
ending letters of the cover text. file.
III. XORSTEG: THE PROPOSED METHOD
After using this algorithm on the cover file, a stego file and a
Our new proposed method of text steganography key file are generated which are sent to the receiver side.
"XORSTEG" works on the same principle of hiding messages
in an innocent looking text paragraph. It works on the binary B. EXTRACTING ALGORITHM (RECEIVER SIDE)
value of the characters and the value of the XOR between the
first and the last words of a line in the paragraph. The main The extracting algorithm does the reverse process of the
demerit of the existing method of key based steganography [4] hiding algorithm and with the help of the stego file and the
is that the secret message has to be hidden in a paragraph key file which is send by the sender, extracts the hidden
which should not contain words which have the same starting message from the stego file.
and ending letter. In our proposed method the cover file that is
1. Read the first decimal as a word and store it in k.
used for hiding the message can contain any number of words
with the same starting and ending letters. 2. Read a line from stego file.
3. Take first word as F1 and last as L2 from that
A. THE HIDING ALGORITHM (SENDER SIDE) line.
4. Take the last and first letter from first word as F1l
Here, any paragraph is taken as the cover file where the and F2s and take last and first letter from last
message to be sent to the receiver is embedded. First the secret word as L1l and L1s.
message is converted into its equivalent binary form. The next 5. Take XOR(F1s and F1l) and XOR(L1s and L1l).
step consists of reading a line of the cover file and then
6. If results of XORs are same, take complement of
writing it into the stego file.
last XOR result.
After the creation of the stego file a line is taken from the 7. Change XOR result to decimal.
stego file and the first and last word of that line is taken. From 8. If k= max(R1,R2) then add 1 to binary else add 0
this first word and the last word, the first and last letters of the to binary.
respective words are taken and their binary equivalent is found
out. XOR is taken between these binary values and a 9. Perform steps (3)-(8) until the same word or
comparison is then made. When the XOR result of both the ending of the word isreached.
first and last word is same, complement of the value of the last 10. Read the next line of the stego file.
word is taken and the binary values are converted into 11. Perform steps (3)-(10) until the words of the key
equivalent decimal values. file are ended.

The key file is generated by taking a bit from the converted 12. Convert the binary to string which will be the
binary secret message. If the bit is 1, the greater decimal value required message.
 The results of hiding an example message of “This is a new
text steganography method” after implementing our proposed
IV. RESULTS AND DISCUSSIONS method in Python are given in Figure 1, 2 and 3.
Here, we compare and analyze the two methods, discussed in
the previous sections. To find the efficiency of the two
methods we calculate the capacity ratio. Capacity can be
defined as the total bytes of a secret message which can be
stored in a cover file. The capacity ratio is computed by
dividing the amount of hidden bytes over the size of the
cover text in bytes [6].

Capacity ratio = (Amount of hidden bytes) / (Size of the

cover text in bytes) Figure 1: The cover file

In the existing method and our proposed method, every bit of

the message is hidden by taking a word of the cover file. So, if
a word contains its starting and ending letter same, the existing
method skips that word. In the existing method if the number
of words in the cover file is less than the number of bytes in
the message then the message cannot be embedded in the
cover file.Some examples of cover files that hide the message
have been given below:
The samples of embedded data are: Figure 2: The generated stego-file

1. America (7 bytes)
2. Every man has two sides to him (29 bytes)
3. With spikes growing out of their skulls (39 bytes)
4. The name milky-way is derived from Greek
mythology (50 bytes)
5. Black was the colour of sorrow and pain, black was
the colour outside the train. (80 bytes)
The message that has to be hidden is taken to be the same as
the cover file. Table 1 shows the capacities of the two
methods: Figure 3: The generated key-file

From the above figures we see that the stego file and cover file
does not differ, unlike other format based methods. This stego
Sample 1 2 3 4 5
file and key file which is generated is then sent to the intended
Number
recipient.
Existing 0 0 0 0 1
Method V. CONCLUSIONS
This type of key based steganography is more efficient than
Proposed 1 1 1 1 1
format based methods which alter the visual properties of a
method
text to hide the message. But similar to cryptography, if the
key and the stego file gets into the hands of any third party,
Table 1: Capacity of the two methods he/she can easily extract the message.

Moreover, we see that the cover file which is used to hide the
In the samples 1, 2, 3 and 4 the existing method’s capacity message has to be larger than the message which is to be
ratio turns out to be 0, 1.21, 1.95, 1.16 but is written as 0 hidden. This can be thought of as another limitation of these
because it cannot hide the message in the cover file as the total key based steganography methods.
number of bytes in the message is greater than the number of
words in the cover file, as some words are neglected by the The future work of this study can be done mainly on
algorithm because they contain the same starting and ending increasing its efficiency by being able to hide a message into
letters. It is also seen that single letter words cannot be used in
any length of a cover file. Support for other languages other
the existing method unlike our proposed method.
than English can also be added. To further increase the
security, we can use both cryptography and steganography Computers, vol.7, no.6, pp. 735-745, 2008.
methods to hide and encipher the message and the key and
[4] Monika Agarwal,"Textsteganographic approaches: a comparison"
then transmit the key to the receiver. International Journal of Network Security Its Applications (IJNSA),
Vol.5, No.1, January 2013.

References [5] K. Benett, “Linguistic steganography- survey, analysis and robustness

concerns for hiding information in text” Purdue University, CERIAS
Tech. Report 2004-13, 2004.
[1]F. A. P. Petitcolas, R.J. Anderson, and M. G. Kuhn, “Information
hiding- a survey,” InProceedings of IEEE, vol.87, pp. 1062-1078, 1999. [6] F. A. Haidari, A. Gutub, K. A. Kahsah, and J. Hamodi, “Improving
security and capacity forArabic text steganography using
[2] L. Y. Por, and B. Delina, “Information hiding- a new approach in text “kashida”extensions” 2009 IEEE/ACS Int.Conf. on ComputerSystems
steganography”, 7th WSEAS Int. Conf. on Applied Computer and and Applications, 2009, pp. 396-399.
AppliedComputational Science, 2008, pp. 689-695.

[3] L. Y. Por, T. F. Ang, and B. Delina, “WhiteSteg- a new scheme in

information hiding using textsteganography,”WSEAS Transactions on