data hiding method in binary images based on block masking for key authentication

9
Data hiding method in binary images based on block masking for key authentication Ki-Hyun Jung a,1 , Kee-Young Yoo b,a School of Computer Information, Yeungjin College, 35 Bokhyun-Ro, Buk-Gu, Daegu 702-721, Republic of Korea b School of Computer Science and Engineering, Kyungpook National University, 80 Daehak-Ro, Buk-Gu, Daegu 702-701, Republic of Korea article info Article history: Received 3 March 2008 Received in revised form 1 June 2011 Accepted 4 February 2014 Available online xxxx Keywords: Data hiding Block masking Digital watermark Key authentication abstract This paper proposes a new data hiding method for binary images that relies on block masking to distribute keys to two parts and then authenticates the right authorized part. The proposed method divides a cover image into small sub-blocks and designs key pairs that determine both where the bit is to be embedded and whether it is possible to embed it there. Furthermore, the key pairs are required to extract the secret data from the stego-image. Experimental results demonstrate a higher capacity and less distortion compared with previous methods since almost all data are hidden in the edge areas. Ó 2014 Elsevier Inc. All rights reserved. 1. Introduction Digital watermarking and data hiding techniques have attracted considerable attention from the viewpoint of applications to copyright protection, copy control, annotation, and authentication in order to keep up with the continued proliferation of digital media, including images, audio, and video. A general data hiding system does not require any information is required to initiate the communication process. The security of such a system thus depends entirely on its secrecy. Fundamentally, then this is not a truly secure practice because it violates Kerckhoffs’ principle [1]. Therefore, the security of a data hiding system should rely on some secret information exchanged between the sender and receiver, such as a stego-key. Without knowledge of this key, no one should be able to extract secret data from a stego-image. Many data hiding techniques have been proposed for applications to digital color and grayscale images. Most such techniques are based on either least significant bit (LSB) substitution or pixel-value differencing in the spatial domain. LSB substitution is a common and well-known technique for hiding data in grayscale and color images. In these methods, secret data are embedded into an image by replacing either a fixed or a variable length of bits. On the other hand, the pixel-value differencing method can be used to embed a large length of bits, and it offers good security. However, both these types of methods cannot be directly applied to binary images. This is because embedding different bits leads to changes in pixel values, and such changes lead to irregularities that are particularly noticeable in a binary. Therefore, hiding data in binary images is more challenging as compared to doing so in other types of images [2]. Generally, only some of the above mentioned types of techniques can be directly applied to binary images. http://dx.doi.org/10.1016/j.ins.2014.02.016 0020-0255/Ó 2014 Elsevier Inc. All rights reserved. Corresponding author. Tel.: +82 53 950 5553; fax: +82 53 957 4846. E-mail addresses: [email protected], [email protected] (K.-H. Jung), [email protected] (K.-Y. Yoo). 1 Tel.: +82 53 940 5527; fax: +82 53 940 5299. Information Sciences xxx (2014) xxx–xxx Contents lists available at ScienceDirect Information Sciences journal homepage: www.elsevier.com/locate/ins Please cite this article in press as: K.-H. Jung, K.-Y. Yoo, Data hiding method in binary images based on block masking for key authentica- tion, Inform. Sci. (2014), http://dx.doi.org/10.1016/j.ins.2014.02.016

Upload: kee-young

Post on 23-Dec-2016

215 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: Data hiding method in binary images based on block masking for key authentication

Information Sciences xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Information Sciences

journal homepage: www.elsevier .com/locate / ins

Data hiding method in binary images based on block maskingfor key authentication

http://dx.doi.org/10.1016/j.ins.2014.02.0160020-0255/� 2014 Elsevier Inc. All rights reserved.

⇑ Corresponding author. Tel.: +82 53 950 5553; fax: +82 53 957 4846.E-mail addresses: [email protected], [email protected] (K.-H. Jung), [email protected] (K.-Y. Yoo).

1 Tel.: +82 53 940 5527; fax: +82 53 940 5299.

Please cite this article in press as: K.-H. Jung, K.-Y. Yoo, Data hiding method in binary images based on block masking for key authtion, Inform. Sci. (2014), http://dx.doi.org/10.1016/j.ins.2014.02.016

Ki-Hyun Jung a,1, Kee-Young Yoo b,⇑a School of Computer Information, Yeungjin College, 35 Bokhyun-Ro, Buk-Gu, Daegu 702-721, Republic of Koreab School of Computer Science and Engineering, Kyungpook National University, 80 Daehak-Ro, Buk-Gu, Daegu 702-701, Republic of Korea

a r t i c l e i n f o a b s t r a c t

Article history:Received 3 March 2008Received in revised form 1 June 2011Accepted 4 February 2014Available online xxxx

Keywords:Data hidingBlock maskingDigital watermarkKey authentication

This paper proposes a new data hiding method for binary images that relies on blockmasking to distribute keys to two parts and then authenticates the right authorized part.The proposed method divides a cover image into small sub-blocks and designs key pairsthat determine both where the bit is to be embedded and whether it is possible to embedit there. Furthermore, the key pairs are required to extract the secret data from thestego-image. Experimental results demonstrate a higher capacity and less distortioncompared with previous methods since almost all data are hidden in the edge areas.

� 2014 Elsevier Inc. All rights reserved.

1. Introduction

Digital watermarking and data hiding techniques have attracted considerable attention from the viewpoint ofapplications to copyright protection, copy control, annotation, and authentication in order to keep up with the continuedproliferation of digital media, including images, audio, and video. A general data hiding system does not require anyinformation is required to initiate the communication process. The security of such a system thus depends entirely on itssecrecy. Fundamentally, then this is not a truly secure practice because it violates Kerckhoffs’ principle [1]. Therefore, thesecurity of a data hiding system should rely on some secret information exchanged between the sender and receiver, suchas a stego-key. Without knowledge of this key, no one should be able to extract secret data from a stego-image.

Many data hiding techniques have been proposed for applications to digital color and grayscale images. Most suchtechniques are based on either least significant bit (LSB) substitution or pixel-value differencing in the spatial domain.LSB substitution is a common and well-known technique for hiding data in grayscale and color images. In these methods,secret data are embedded into an image by replacing either a fixed or a variable length of bits. On the other hand, thepixel-value differencing method can be used to embed a large length of bits, and it offers good security. However, both thesetypes of methods cannot be directly applied to binary images. This is because embedding different bits leads to changes inpixel values, and such changes lead to irregularities that are particularly noticeable in a binary. Therefore, hiding data inbinary images is more challenging as compared to doing so in other types of images [2]. Generally, only some of the abovementioned types of techniques can be directly applied to binary images.

entica-

Page 2: Data hiding method in binary images based on block masking for key authentication

2 K.-H. Jung, K.-Y. Yoo / Information Sciences xxx (2014) xxx–xxx

On the other hand, very little research has been done on data hiding in binary images. The binary image is common andoften appears in newspapers, faxes, and magazines. Hiding is difficult for binary images since their black or white pixels re-quire only one-bit representation. There are two primary methods of data hiding for these images: sub-block modificationand single-pixel manipulation. The first method modifies sub-blocks, which is divided into a group of pixels. Matsui and Ta-naka embedded secret data in dithered images by manipulating the dithering patterns; they also embedded data in faximages, by manipulating the run lengths [3]. Low et al. changed the line spacing and character spacing to embed secret datain textual images, for bulk electronic publications [4,5]. These methods are used for some special types of binary images. Thesecond approach modifies a single pixel, from black to white or vice versa: some special single pixels in the image are chan-ged to embed the secret data. Koch and Zhao proposed a data hiding method by forcing the ratio of black and white pixels ina block to be larger or smaller than one [6]. However, there is some difficulty with this approach. Only a limited number ofbits can be embedded, since the enforcing method has trouble dealing with blocks that have a significantly low or high per-centage of black pixels. Wu et al. embedded bits in image blocks, selected by calculating a characteristic value and finding apattern [7]. Liu et al. partitioned the binary image into blocks of 2 � 2 pixels and embedded a bit 0 or 1 in the block. Thismethod can hide one bit per block by modifying 0.5 pixels on average [8]. Wu and Liu manipulated the flappable pixelsto embed secret data into shuffled blocks. The shuffling of the blocks before embedding ensures the equalization of theembedding capacity from region to region without causing noticeable visual effects [9]. Venkatesan et al. proposed usingthe parity of blocks. The cover image is partitioned into small blocks, in which one bit information is stored. Unfortunately,if all of the pixel values belong to 0 or 1, a secret bit cannot be hidden [10]. Pan et al. proposed a data hiding method bypartitioning into 4 � 4 blocks, where each block was repartitioned into overlapping sub-blocks [11]. Most of them do notinclude a stego-key, so this is less safe than other methods that adopt a stego-key [12–14]. Some methods are performedto authenticate for binary images with small distortion and image recovery [15–18].

In this paper we propose a new data hiding method in binary images using block masking to distribute the stego-key totwo parts and then authenticates the right authorized part. By determining the location of embedding and selecting the edgeareas, the image quality of the stego-image can be maintained at a high level quality with relatively low computationalcomplexity.

This paper is organized as follows. In Section 2, our proposed data hiding method is described in more detail. In Section 3,our experimental results are presented and discussed. Our conclusions are presented in Section 4.

2. The proposed method

In this section, we present a detailed consideration of how data are embedded and extracted. To begin with, a binary im-age is made up of black and white pixels. Only a single bit is used to represent each pixel, say 0 or 1.

Let C be a cover image of W � H pixels and S be the s-bit secret data. For each p(i, j) pixel value in image C, the new pixelvalue is p0(i, j).

2.1. Data embedding

Fig. 1 presents a block diagram of the embedding and distribution process. The cover image is divided into sub-blocks andthen key pairs are generated. These key pairs decide the embedding position and determine whether it is possible to embed asecret bit that position. The generated stego-image and key pairs can be distributed across other parts. In this situation, therecan be various applications. For example, an encryption algorithm can be used to generate the key pairs and send key pairsmore than two parts.

A given cover image is partitioned into M � N blocks. Data hiding is achieved by modifying some bits in the sub-blocks.The total number of sub-blocks, T generated

Pleasetion, I

T ¼W � HM � N

: ð1Þ

Here, the embedding capacity, E is less than or equal to T bits.

Divide into blocks

Cover imageC

Generatekey pairs

Determineembeddable location

Dataembedding

Secret dataS

Stego-imageC'

Keydistribution

Fig. 1. Block diagram of embedding and distribution process.

cite this article in press as: K.-H. Jung, K.-Y. Yoo, Data hiding method in binary images based on block masking for key authentica-nform. Sci. (2014), http://dx.doi.org/10.1016/j.ins.2014.02.016

Page 3: Data hiding method in binary images based on block masking for key authentication

K.-H. Jung, K.-Y. Yoo / Information Sciences xxx (2014) xxx–xxx 3

Next, we design key pairs. Let Kx and Ky be the key values for an M � N block. There are two constraints on the design ofeach key value. First, the generated key value has to satisfy the condition given in Eq. (2). Key values are determined underthe condition that the sum of each key is zero. This requirement is based on the idea that it can be determined whether asub-block belongs to an edge or a smooth area. If it belongs to an edge area, it can host more embedded secret data withoutdistortion visible to the human eye.

Pleasetion, I

SðKxÞ ¼XM�1

i¼0

XN�1

j¼0

Kxði; jÞ ¼ 0; SðKyÞ ¼XM�1

i¼0

XN�1

j¼0

Kyði; jÞ ¼ 0: ð2Þ

Second, there exists P that satisfies Eq. (3), where the secret bit is embedded. This condition determines the location forembedding.

P ¼ fði; jÞjKxði; jÞ ¼ Kyði; jÞ ¼ 0; 0 6 i 6 M � 1; 0 6 j 6 N � 1g: ð3Þ

For example, let M = 3 and N = 3, and let a sub-block be as given in Fig. 2.To satisfy these two constraints, assume that Kx and Ky are as given in Fig. 3. Here, S(Kx) = (�1) + 0 + 1 + (�2) +

0 + 2 + (�1) + 0 + 1 = 0, S(Ky) = 1 + 2 + 1 + 0 + 0 + 0 + (�1) + (�2) + (�1) = 0, and for P = (1,1), Kx(1,1) = Ky(1,1) = 0. Thus, thekey pairs are well designed to embed secret data, since these two conditions are satisfied.

Next, we determine whether the selected sub-block can hide a secret bit. For the sake of simplicity, let Kx = x(i, j)x(i, j + 1). . . x(i, j + M � 1)x(i + 1, j) . . . x(i + N � 1, j + M � 1) and Ky = y(i, j)y(i, j + 1) . . . y(i, j + M � 1)y(i + 1, j) . . . y(i + N � 1, j + M � 1) insequence. Calculate G of each pixel for the sub-block, using Eq. (4).

G ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiGx2 þ Gy2

q: ð4Þ

Each Gx and Gy is calculated by

Gx ¼XM�1

m¼0

XN�1

n¼0

xðiþm; jþ nÞpðiþm; jþ nÞ; Gy ¼XM�1

m¼0

XN�1

n¼0

yðiþm; jþ nÞpðiþm; jþ nÞ: ð5Þ

The value of G is an important factor in determining whether or not to embed in the selected sub-block. Both Gx and Gyare the sum of key value and pixel for each location. For a sub-block, if any value of G is larger than the predefined threshold,that sub-block is where the secret bit is embedded.

Fig. 3. An example of key pairs.

Divide into blocks

Stego-imageC'

Authenticatekey pairs

Findembedded location

Dataextracting

Secret dataS

Fig. 4. Block diagram of the extraction process.

pij pi(j+1) pi(j+2)

p(i+1)j p(i+1)(j+1) p(i+1)(j+2)

p(i+2)j p(i+2)(j+1) p(i+2)(j+2)

Fig. 2. A sub-block of M = 3 and N = 3.

cite this article in press as: K.-H. Jung, K.-Y. Yoo, Data hiding method in binary images based on block masking for key authentica-nform. Sci. (2014), http://dx.doi.org/10.1016/j.ins.2014.02.016

Page 4: Data hiding method in binary images based on block masking for key authentication

(a) Baboon (b) Airplane

(c) Lena (d) Gatbawi

(e) Peppers (f) Epitaph

Fig. 5. Six cover images.

Table 1Capacity and visual quality of the proposed method for the 3 � 3 block size.

Cover image Embeddable bits (bits) Hidden bits (bits) Visual quality (dB)

Baboon 11,453 11,396 18.37Airplane 3788 3440 23.52Lena 4629 4415 22.50Gatbawi 9618 9616 19.19Peppers 3181 3181 24.01Epitaph 10,228 9887 19.02

4 K.-H. Jung, K.-Y. Yoo / Information Sciences xxx (2014) xxx–xxx

Please cite this article in press as: K.-H. Jung, K.-Y. Yoo, Data hiding method in binary images based on block masking for key authentica-tion, Inform. Sci. (2014), http://dx.doi.org/10.1016/j.ins.2014.02.016

Page 5: Data hiding method in binary images based on block masking for key authentication

K.-H. Jung, K.-Y. Yoo / Information Sciences xxx (2014) xxx–xxx 5

2.2. Data extracting

Fig. 4 presents a block diagram of the extraction process. Divide a stego-image into M � N sub-blocks. Receivers get keysat the same time, from which the location that the data are embedded can be determined. If this location is known the secretbit can be directly extracted from the sub-block. Finally, accumulate in order for all of the secret bits together that extractedfrom each step.

(c) Lena (d) Gatbawi

(e) Peppers (f) Epitaph

(a) Baboon (b) Airplane

Fig. 6. Embeddable bits for cover images.

Please cite this article in press as: K.-H. Jung, K.-Y. Yoo, Data hiding method in binary images based on block masking for key authentica-tion, Inform. Sci. (2014), http://dx.doi.org/10.1016/j.ins.2014.02.016

Page 6: Data hiding method in binary images based on block masking for key authentication

(a) EC=11,396bits, PSNR=18.37dB (b) EC=3,440bits, PSNR=23.52dB

(c) EC=4,415bits, PSNR=22.50dB (d) EC=9,616bits, PSNR=19.19dB

(e) EC=3,181bits, PSNR=24.01dB (f) EC=9,887bits, PSNR=19.02dB

Fig. 7. Six stego-images.

6 K.-H. Jung, K.-Y. Yoo / Information Sciences xxx (2014) xxx–xxx

For example, key pairs are given, as in Fig. 3. At the position p(i + 1, j + 1), a secret bit is embedded, so we extract the valueand invert the bit value, since binary images are used as cover images.

3. Experimental results

In our experiments, the six 512 � 512 binary images shown in Fig. 5 were used as cover images. The secret data werepseudo-random numbers. In this study, we adopted the peak signal-to-noise ratio (PSNR) to extend the 1-bit binary value

Please cite this article in press as: K.-H. Jung, K.-Y. Yoo, Data hiding method in binary images based on block masking for key authentica-tion, Inform. Sci. (2014), http://dx.doi.org/10.1016/j.ins.2014.02.016

Page 7: Data hiding method in binary images based on block masking for key authentication

K.-H. Jung, K.-Y. Yoo / Information Sciences xxx (2014) xxx–xxx 7

to 8-bit, and calculated the capacity in terms of the amount of data that can be embedded. Our experiments employed anextended 8-bit value in which the white is extended to 0 � FF and black to 0 � 00.

Table 2Compar

Cove

BaboAirpLenaGatbPeppEpita

Pleasetion, I

PSNR ¼ 10� log102552=MSE; ð6Þ

where MSE is the mean square error, which is defined as

MSE ¼XW�1

i¼0

XH�1

j¼0

ðpði; jÞ � p0ði; jÞÞ2=W � H: ð7Þ

Table 1 shows the experimentally determined capacity and visual quality of the proposed method with the keyKx = (�1)01(�2)02(�1)01, Ky = 121000(�1)(�2)(�1), and threshold Th = 255 for M = N = 3. The table shows that the pro-posed method can hide more secret data when the cover image has rapidly changing pixel values. For example, the Baboon,Gatbawi, and Epitaph image have a higher capacity than the other images because they contain many blocks to be hidden sothat all of the sub-block’s values are scattered uniformly and randomly. Each sub-block can host an embedded secret bit,except for a full black or white block where a secret bit is embedded with a different bit value, or where it can be easily de-tected. The number of hidden bits is less than or equal to the number of embeddable bits for each cover image. This meansthat a sub-block can hide one or more secret bits.

For the selected key pairs, the embeddable blocks for each cover image are shown in Fig. 6. These embeddable bits deter-mine how much secret data can be hidden. In the Baboon image, 11,396 blocks can be used to hide bits since the values of thepixels are uniformly distributed. In the Peppers image, many areas are filled with black or white spaces, so it cannot hide asmuch secret data. Secret data can be hidden around each character in the Epitaph image.

Fig. 7 shows the stego-images obtained after secret data were embedded in each 3 � 3 sub-block. In these experiments,the proposed method produces less distortion of the cover images, since almost all data are hidden in the edge areas, whichare difficult for the human eye to detect. For the stego-images shown in Fig. 7(b, c, and e), secret data can be embedded lessthan easily than in the others, since many sub-blocks in these images have an area that is filled 0 or 1 and was thus rejectedfor embedding.

Table 2 shows a comparison of our proposed method and other methods. As can be seen in the table, the proposed meth-od can hide more secret data than the other methods, though the PSNR value is higher for each image. For all the coverimages used, the proposed method demonstrated a high capacity and good visual quality. Compared with the method ofVenkatesan et al., the proposed method can hide more from 179 bits to 2343 bits and produces good quality from1.79 dB to 2.42 dB. Our proposed method can hide 6989 bits and 25.00 dB on average. Compared to the method of Tsenget al., our method can hide 119 bits more and produces visual quality that is 6 dB better.

We now consider the other key pairs, Kx = 00(�1)010000, Ky = (�1)00010000 in terms of improvement of visual quality.Table 3 shows that the proposed method maintains a good level of visual quality rather than a high capacity. Compared withthe results in Table 2, these new key pairs keep a good level of quality from 1.70 dB to 2.22 dB. This indicates that the capac-ity and visual quality of the proposed method depends on the generated key pairs.

Fig. 8 shows the difference between the stego-images and the cover images; in the cover images, most of the distortionsoccurred on the edges. As the results show, all of the secret data are embedded in the edge areas, which make them lessvisible. We can see that most of the distortions occurred on the edges of the cover images.

Table 3Capacity and visual quality of proposed method for the improvement of PSNR.

Cover image Embeddable bits (bits) Hidden bits (bits) Visual quality (dB)

Baboon 7764 7707 20.05Airplane 2579 2239 25.33Lena 3131 2917 24.20Gatbawi 5827 5825 21.41Peppers 1981 1981 26.04Epitaph 6356 6015 21.08

ison of embedding results.

r image Venkatesan et al.’s method [10] Tseng et al.’s method [14] Proposed method

Capacity (bits) PSNR (dB) Capacity (bits) PSNR (dB) Capacity (bits) PSNR (dB)

on 9441 16.36 11,806 16.83 11,396 18.37lane 3293 21.25 3292 21.74 3440 23.52

3657 20.45 4198 20.70 4415 22.50awi 7273 17.13 9551 17.11 9616 19.19ers 2880 21.59 3015 20.50 3181 24.01ph 8953 17.23 9360 17.34 9887 19.02

cite this article in press as: K.-H. Jung, K.-Y. Yoo, Data hiding method in binary images based on block masking for key authentica-nform. Sci. (2014), http://dx.doi.org/10.1016/j.ins.2014.02.016

Page 8: Data hiding method in binary images based on block masking for key authentication

(a) Baboon (b) Airplane

(c) Lena (d) Gatbawi

(e) Peppers (f) Epitaph

Fig. 8. Differences between cover and stego-images.

8 K.-H. Jung, K.-Y. Yoo / Information Sciences xxx (2014) xxx–xxx

4. Conclusions

In this paper, we have proposed a data hiding method that used block masking to distribute keys to two parts and thatauthenticated the correct authorized part. No reference to the original cover image was required during extraction of theembedded secret data from the stego-image. This method manipulated blocks that were sub-divided into a small M � N-sizeblock, and it then designed design key pairs which determined both the location of embedding and whether or not it waspossible to embed. For the generated key pairs, the embedding capacity and visual quality were also determined. In addition,key pairs were required in order to extract secret data from the stego-image. Each key value checked upon an authorizedpart.

Please cite this article in press as: K.-H. Jung, K.-Y. Yoo, Data hiding method in binary images based on block masking for key authentica-tion, Inform. Sci. (2014), http://dx.doi.org/10.1016/j.ins.2014.02.016

Page 9: Data hiding method in binary images based on block masking for key authentication

K.-H. Jung, K.-Y. Yoo / Information Sciences xxx (2014) xxx–xxx 9

Our experimental results have shown that the proposed method provided a better way to hide more secret data com-pared with other methods, and did so without producing noticeable distortions. The proposed method produced less distor-tion since almost all the data were hidden in the edge areas and where they were difficult to detect by the human eye.According to the capacity or visual quality, various applications of this method can be implemented depending on thekey pairs.

References

[1] A. Kerckhoffs, La cryptographie militaire, J. Sci. Militaires 9 (1883) 5–38.[2] H. Liang, W. Ran, X. Nie, A secure and high capacity scheme for binary images, in: Proc. of the ICWAPR, 2007, pp. 224–229.[3] K. Matsui, K. Tanaka, Video-steganography: how to secretly embed a signature in a picture, in: Proc. IMA Intellectual Property Project, 1994, pp.187–

206.[4] S.H. Low, N.F. Maxemchuk, A.M. Lapone, Document identification for copyright protection using centroid detection, IEEE Trans. Commun. (1998) 372–

383.[5] J.T. Brassil, S.H. Low, N.F. Maxemchuk, Copyright protection for the electronic distribution of text documents, in: Proc. of IEEE, 1999, pp. 1181-1196.[6] E. Koch, J. Zhao, Embedding robust labels into images for copyright protection, in: Proc. of the International Congress on Intellectual Property Rights for

Specialized Information, Knowledge & New Technologies, 1995, pp. 242–251.[7] M. Wu, E. Tang, B. Liu, Data hiding in digital binary images, in: IEEE Inter. Conf. on Multimedia & Expo, 2000, pp. 393–396.[8] C. Liu, Y. Dai, Z. Wang, A novel information hiding method in binary images, J. Southeast Univ. (2003).[9] M. Wu, B. Liu, Data hiding in binary image for authentication and annotation, IEEE Trans. Multimedia (2004) 528–538.

[10] M. Venkatesan, P. Meenakshidevi, K. Duraiswamy, K. Thiagarajah, A new data hiding scheme with quality control for binary images using block parity,in: 3rd Inter. Symposium on Information Assurance and Security, 2007, pp. 468–471.

[11] G. Pan, Y.J. Wu, Z.H. Wu, A novel data hiding method for two-color images, LNCS (2001) 261–270.[12] H. Yang, A.C. Kot, Pattern-based data hiding for binary image authentication by connectivity-preserving, IEEE Trans. Multimedia (2007) 475–486.[13] H. Ajetrao, Kulkarni, N. Gaikwad, A novel scheme of data hiding in binary images, in: Proc. of IEEE, 2007, pp. 70–77.[14] H.W. Tseng, F.R. Wu, C.P. Hsieh, Data hiding for binary images using weight mechanism, IIHMSP (2007) 307–310.[15] J.M. Guo, S.C. Pei, H. Lee, Watermarking in halftone images with parity-matched error diffusion, Signal Process. 91 (2011) 126–135.[16] K.T. Lin, Hybrid encoding method by assembling the magic-matrix scrambling method and the binary encoding method in image hiding, Opt.

Commun. 284 (2011) 1778–1784.[17] Y. Lee, H. Kim, Y. Park, A new data hiding scheme for binary image authentication with small image distortion, Inform. Sci. 179 (2009) 3866–3884.[18] R. Chamlawi, A. Khan, Digital image authentication and recovery: employing integer transform based information embedding and extraction, Inform.

Sci. 180 (2010) 4909–4928.

Please cite this article in press as: K.-H. Jung, K.-Y. Yoo, Data hiding method in binary images based on block masking for key authentica-tion, Inform. Sci. (2014), http://dx.doi.org/10.1016/j.ins.2014.02.016