2 Professor, Dept of CSE, Shadan College of Engineering & Technology, Hyderabad, Telangana, India,

Transcription

1 ISSN Vol.07,Issue.06, July-2015, Pages: A Novel Image Broadcasting Method using Secret-Fragment-Visible Mosaic Images through Inverse Color Transformations RAAFIA FATIMA 1, DR. MD. ATEEQ UR RAHMAN 2 1 PG Scholar, Dept of CSE, Shadan College of Engineering & Technology, Hyderabad, Telangana, India, syedasubbuhisiddiqua@gmail.com. 2 Professor, Dept of CSE, Shadan College of Engineering & Technology, Hyderabad, Telangana, India, mail_to_ateeq@yahoo.com. Abstract: Color Images from different sources are regularly utilized and are transmitted through the web for different purposes, for example, private endeavor chronicles, report stockpiling frameworks, medicinal imaging frameworks, and military picture databases. These pictures may contain mystery or classified data since it ought to be shielded from spillage amid transmissions. A methodology for secure picture transmission is required, which is to change a mystery picture into an important Secret Fragment Mosaic picture with size practically same and appearing to be like the preselected target picture. The mosaic picture is the result of masterminding of the piece parts of a mystery picture in a manner to camouflage the other picture called the objective picture. The mosaic picture, which appears to be like an arbitrarily chosen target picture, which is utilized for stowing away of the mystery picture by shading changing their qualities like the pieces of the objective picture. Such system is essential so for the lossless recuperation of the transmitted mystery picture. The suitable data is implanted into the mosaic picture for the recuperation of the transmitted mystery picture. Great trial results demonstrate the possibility of the proposed technique. Keywords: Color Transformation, Data Hiding, Image Encryption, Mosaic Image, Secure Image Transmission. I. INTRODUCTION These days, images from different sources are regularly utilized and are transmitted through the web for different applications, for example, classified endeavor documents, report stockpiling frameworks, therapeutic imaging frameworks, and military picture databases. These pictures generally contain private or secret data so that they ought to be shielded from spillages amid transmissions. As of late, numerous routines have been proposed for securing picture transmission, for which two regular methodologies are picture encryption and information covering up. Encryption of picture is a method that make utilization of the normal property of a picture, for example, high repetition and solid spatial relationship, to get a scrambled picture. The scrambled picture is good for nothing and this may stir the outsiders consideration because of its 2015 IJATIR. All rights reserved. arbitrariness in structure amid transmission. Another technique for secure picture transmission is information concealing that shrouds a mystery element into a spread picture so that an outsider can't discovered the vicinity of the mystery substance. The issue of information covering up is the trouble in installing substantial volume of mystery element into a solitary picture. In the event that anybody needs to shroud a mystery element into a spread picture, the mystery substance must be exceptionally compacted before. Amid recovery this will bring about twisting of the mystery element. In this paper, we propose an approach for secure image transmission is needed, which is to transform a secret image into a meaningful Secret Fragment Mosaic image with size almost same and looking similar to the preselected target image. The mosaic image is the outcome of arranging of the block fragments of a secret image in a way so as to disguise the other image called the target image. The mosaic image, which looks similar to a randomly selected target image, which is used for hiding of the secret image by color transforming their characteristics [5] similar to the blocks of the target image. Such technique is necessary so for the lossless recovery of the transmitted secret image. The encoded picture is a commotion picture so that nobody can acquire the mystery picture from it unless he/she has the right key. Be that as it may, the encoded picture is a trivial record, which can't give extra data before decoding and may stimulate an aggressor's consideration amid transmission because of its irregularity in structure. A distinct option for keep away from this issue is information concealing [8] [18] that shrouds a mystery message into a spread picture so that nobody can understand the presence of the mystery information, in which the information sort of the mystery message explored in this paper is a picture. Existing information concealing systems predominantly use the strategies of LSB substitution [8], histogram moving[9], distinction development[10] [11], expectation slip extension[12] [13], recursive histogram alteration[14],and discrete cosine/wavelet changes[15]-[18]. Notwithstanding, with a specific end goal to lessen the contortion of the subsequent picture, an upper headed for the mutilation worth is normally situated on the payload of the

2 spread picture. An examination on this rate mutilation issue can be found in [19]. Hence, a primary issue of the systems for concealing information in pictures is the trouble to implant a lot of message information into a solitary picture. In particular, if one needs to conceal a mystery picture into a spread picture with the same size, the mystery picture must be exceptionally packed ahead of time. Case in point, for an information concealing system with an installing rate of 0.5 bits for each pixel, a mystery picture with 8 bits for every pixel must be packed at a rate of no less than 93.75% already keeping in mind the end goal to be covered up into a spread picture as shown in Fig.1. Be that as it may, for some applications, for example, keeping or transmitting therapeutic pictures, military pictures, authoritative reports, and so on., that are profitable with no recompense of genuine twists, such information pressure operations are generally unfeasible. Also, most picture pressure routines, for example, JPEG pressure, are not suitable for line drawings and printed design, in which sharp differences between neighboring pixels are regularly destructed to wind up perceptible ancient rarities [20]. Fig. 1. Encryption system diagram. With the always expanding heightening of media applications, security is a critical concern in regards to correspondence and capacity of pictures, and encryption is one the approaches to ensure security. Picture encryption gone through in change unique picture to another picture that is hard to comprehend; to keep the picture private between clients, in other word, it is fundamental that no one could procure the substance without a key for decoding. As per Kirchhoff's rule of secure cryptosystem, the insurance ought to rely on upon the mystery of the key, not the mystery of the encryption/decoding calculation that was utilized. As such, it is assumed that the calculation is freely known, yet decoding of message is infeasible on the premise of the figure message notwithstanding colleague of the calculation. Each figure content piece is influenced by numerous plaintext pieces. Normally, it is likewise unfeasible to scan for relationship between the keys without the information of some included data. In particular, after an objective picture is chosen selfassertively, the given mystery picture is initially partitioned into rectangular parts called tile pictures, which then are fit into comparable squares in the objective picture, called target pieces, as per a likeness rule in light of shading varieties. Next, the shading normal for every tile picture is changed to be that of the comparing target hinder in the objective picture, bringing about a mosaic picture which RAAFIA FATIMA, MD ATEEQ UR RAHMAN resembles the objective picture. Significant plans are additionally proposed to direct almost lossless recuperation of the first mystery picture from the subsequent mosaic picture. The proposed technique is new in that an important mosaic picture is made, conversely with the picture encryption strategy that just makes insignificant commotion pictures. Likewise, the proposed technique can change a mystery picture into a masking mosaic picture without pressure, while an information concealing strategy must shroud an exceptionally compacted rendition of the mystery picture into a spread picture when the mystery picture and the spread picture have the same information volume. II. IDEAS OF THE PROPOSED METHOD In this paper, another method for secure picture transmission is proposed, which changes a mystery picture into a significant mosaic picture with the same size and resembling a preselected target picture. The change procedure is controlled by a mystery key, and just with the key can a man recoup the mystery picture almost losslessly from the mosaic picture. The proposed strategy is propelled by Lai and Tsai [21], in which once again sort of PC workmanship picture, called mystery part noticeable mosaic picture, was proposed. The mosaic picture is the consequence of modification of the parts of a mystery picture in camouflage of another picture called the objective picture preselected from a database. Be that as it may, a conspicuous shortcoming of Lai and Tsai [21] is the prerequisite of an extensive picture database so that the created mosaic picture can be adequately like the chose target picture. Utilizing their strategy, the client is not permitted to choose uninhibitedly his/her most loved picture for utilization as the objective picture. It is in this way coveted in this study to uproot this shortcoming of the system while keeping its legitimacy, that is, it is meant to outline another strategy that can change a mystery picture into a mystery part noticeable mosaic picture of the same size that has the visual appearance of any unreservedly chose target picture without the need of a database. Fig. 2. Flow diagram of the proposed method. The proposed system incorporates two primary stages as demonstrated by the stream outline of Fig. 2: 1) mosaic picture creation and2) mystery picture recovery. In the first stage, a mosaic picture is yielded, which comprises of the parts of an information mystery picture with shading revisions as per a likeness paradigm in light of shading variations. The stage incorporates four stages: 1) fitting the tile pictures of the mystery picture into the objective pieces of a preselected

3 A Novel Image Broadcasting Method using Secret-Fragment-Visible Mosaic Images Through Inverse Color Transformations target picture; 2) changing the shading normal for every Furthermore, we have to embed into the created mosaic tile picture in the mystery picture to turn into that of the image sufficient information about the new tile image for relating target hinder in the objective picture; 3) pivoting every tile picture into a bearing with the base RMSE esteem regarding its comparing target square; and 4) installing pertinent data into the made mosaic picture for future recuperation of the mystery picture. In the second stage, the installed data is separated to recuperate almost losslessly the mystery picture from the produced mosaic picture. The stage incorporates two stages: 1) extricating the installed data for mystery picture recuperation from the mosaic picture, and 2) recouping the mystery picture utilizing the separated data. III. LOGIC OF MOSAIC IMAGE GENERATION Problems encountered in generating mosaic images are detailed in this section with the proposed solutions. A. Color Transformations between Blocks In the first period of the proposed plan, every tile picture T in the given mystery picture is fit into an objective square B in a preselected target picture. Since the shading attributes of T and B are not the same as one another, how to change their shading circulations to make them resemble the other alike is the primary issue here. Rein hard et al. [22] proposed a shading move plot in this viewpoint, which changes over the shading normal for a picture to be that of another in the Iαβ shading space. This thought is a response to the issue and is received in this paper, with the exception of that the RGB shading space rather than the Iαβ one is utilized to decrease the volume of the obliged data for recuperation of the first mystery image. More particularly, let T and B be portrayed as pixel and respectively. Let the color of each be denoted by and that of each by. At first, we compute the means and standard deviations of T and B, respectively, in each of the three color channels R, G, and B by the following formulas: (2) in which c i and denote the C-channel values of pixels p i and, respectively, with c = r, g, or b and C=R, G, or B (1) use in the later stage of recovering the original secret image. For this, theoretically we can use (4) to compute the original pixel value of p i. However, the involved mean and standard deviation values in the formula are all real numbers, and it is impractical to embed real numbers, each with many digits, in the generated mosaic image. Therefore, we limit the numbers of bits used to represent relevant parameter values in (3) and (4). Specifically, for each color channel we allow each of the means of T and B to have 8 bits with its value in the range of 0 to 255, and the standard deviation quotient q c in (3) to have 7 bits with its value in the range of 0.1 to That is, each mean is changed to be the closest value in the range of0 to 255, and each q c is changed to be the closest value in the range of 0.1 to We do not allow qc to be 0 because otherwise the original pixel value cannot be recovered back by(4) for the reason that 1/q c in (4) is not defined when q c =0. B. Choosing Appropriate Target Blocks and Rotating Blocks to Fit Better with Smaller RMSE Value In transforming the color characteristic of a tile image T to be that of a corresponding target block B as described above, how to choose an appropriate B for each T is an issue. For this, we use the standard deviation of the colors in the block as a measure to select the most similar B for each T. Specially, we sort all the tile images to form a sequence, Stile, and all the target blocks to form another, Star get, according to the average values of the standard deviations of the three color channels. Then, we fit the first in Stile into the first in Starget, fit the second in Stile into the second in Starget, and so on. Additionally, after a target block B is chosen to fit a tile image T and after the color characteristic of T is transformed, we conduct a further improvement on the color similarity between the resulting tile image and the target block B by rotating into one of the four directions, 0 o, 90 o, 180 o, and 270 o, which yields a rotated version of with the minimum root mean square error (RMSE) value with respect to B among the four directions for final use to fit T into B. C. Handling Overflows/Underflows in Color Transformation After the color transformation process is conducted as described previously, some pixel values in the new tile image might have overflows or underflows. To deal with this problem, we convert such values to be non-overflow or non-under flow ones and record the value differences as residuals for use in later recovery. Specifically, we convert all.next, we compute new color values for each p i the transformed pixel values in not smaller than 255 to be in T by 255,and all those not larger than 0 to be 0. Next, we compute the differences between the original pixel values and the (3) converted ones as the residuals and record them as part of the in which is the standard deviation quotient and c information associated with. Accordingly, the pixel values, = r,g, or b. It can be verified easily that the new color mean which are just on the bound of 255 or 0, however, cannot be and variance of the resulting tile image are equal to distinguished from those with overflow/underflow values those of B, respectively. To compute the original color during later recovery since all the pixel values with overflows/underflows are converted to be 255 or 0 now. To values (r i, g i,b i ) of pi from the new ones, we overcome this, we set the residuals of those pixel values use the following formula which is the inverse of (3): which are on the sure to be 0 and record them also. Notwithstanding, as can be seen from (3), the scopes of (4)

4 conceivable lingering qualities are obscure, and this causes an issue of choosing what number of bits ought to be utilized to record a remaining. To solve this problem, we record the residual values in the untransformed color space rather than in the transformed one. That is, by using the following two formulas, we compute first the smallest possible color value cs (with c = r, g, or b) in T that becomes larger than 255, as well as the largest possible value c L in T that becomes smaller than 0, respectively, after the color transformation process has been conducted: (5) Next, for an untransformed value ci which yields an over flow after the color transformation, we compute its residual as c i - c S ; and for c i which yields an underflow, we compute its residual as c L - c i. Then, the possible values of the residuals of c i will all lie in the range of 0 to 255 as can be verified. Consequently, we can simply record each of them with 8bits. And finally, because the residual values are centralized around zero, we use further in this study the Huffman encoding scheme to encode the residuals in order to reduce the number of required bits to represent them. D. Embedding Information for Secret Image Recovery In order to recover the secret image from the mosaic image, we have to embed relevant recovery information into the mosaic image. For this, we adopt a technique proposed by Coltuy and Chassery [24] and apply it to the least significant bits of the pixels in the created mosaic image to conduct data embedding. Unlike the classical LSB replacement methods [8],[25], [26], which substitute LSBs with message bits directly, the reversible contrast mapping method applies simple integer transformations to pairs of pixel values. Specifically, the method conducts forward and backward integer transformations as follows, respectively, in which (x, y) are a pair of pixel values and (x, y ) are the transformed ones (7) The technique results in high data embedding capacities which is close to the highest bit rates and has the lowest complexity reported so far. The information required to recover a tile image T which is mapped to a target block B includes: 1) the index of B; 2) the optimal rotation angle of T; 3) the truncated means of T and B and the standard deviation quotients, of all color channels and 4) the overflow/underflow residuals. These data items fore covering a tile image T are integrated as a five-component bit stream of the form In which the bit segments represent the values of the index of B, the rotation angle of T, the means of T and B, the standard deviation quotients, and the residuals, respectively. RAAFIA FATIMA, MD ATEEQ UR RAHMAN (6) In more detail, the numbers of required bits for the five data items in M are discussed below: 1) the index of B needs m bits to represent, with m computed by (8) in which WS and HS are respectively the width and height of the secret image S, and NT is the size of the target image T; 2) it needs two bits to represent the rotation angle of T because there are four possible rotation directions; 3) 48 bits are required to represent the means of T and B because we use eight bits to represent a mean value in each color channel; 4) it needs 21 bits to represent the quotients of T over B in the three color channels with each channel requiring 7 bits; and 5) the total number k of required bits for representing all the residuals depends on the number of overflows or underflows in T then, the above-defined bit streams of all the tile images are concatenated in order further into a total bit stream M t for the entire secret image. Moreover, in order to protect M t from being attacked, we encrypt it with a secret key to obtain an encrypted bit stream M, which is finally embedded into the pixel pairs in the mosaic image using the method of Coltucand Chassery [24] described above. It may require more than one iteration in the encoding process since the length of may be larger than the number of pixel pairs available in an iteration. A plot of the statistics of the numbers of required bits for secret image recovery is shown in Fig. 8(b).Moreover, we have to embed as well some related information about the mosaic image generation process into the mosaic image for use in the secret image recovery process. Such information, described as a bit stream I like M mentioned previously, includes the following data items: 1) the number of iterations conducted in the process for embedding the bit stream ; 2) the total number of used pixel pairs in the last iteration for embedding ; and 3) the Huffman table for encoding the residuals. IV. ALGORITHMS OF THE PROPOSED METHOD The detailed algorithms for mosaic image creation and secret image recovery may now be described in Algorithms 1 and 2 respectively. Algorithm1: Mosaic image creation T-target image, S- secret image, F-mosaic image Stage1: Fitting blocks of secret images into blocks of target blocks If the size of T is different from S, change the size. Divide S and T into n blocks of same size. Compute the means and the standard deviations (SD) of each tile [1]. Compute the average SD. Sort the tile images in S and T. Map tile between S and T. Create F. Stage2: Transforming color characteristics of blocks of secret image similar to target image For each mapping from secret to target calculate the mean and SD.

5 A Novel Image Broadcasting Method using Secret-Fragment-Visible Mosaic Images Through Inverse Color Transformations Each pi in each block of F with color value c i, transform c i into a new value using c i =q c (c i -μ c ) + μ c. If ci is not less than 255 or if it is not greater than0, then change to be 255 or 0. Stage 3: Rotating secret image blocks in the direction with minimum RMSE value Compute the RMSE values. Rotate tile into the optimal direction with the smallest RMSE value. Stage 4: Embed information for recovery purpose For each tile image in F, construct a bit stream M for recovering T Index, rotation angle θ, means and the SD quotients. Generate a bit stream Mt by K. Embed Mt into F. Fig.4. Target image &Secret image. Algorithm 2: Secret image recovery T-target image, S- secret image, F-mosaic image Stage 1: Extracting the embedded information. Extract the bit stream Mt by K. Decompose Mt into n bit streams. Decode M for each tile image to obtain the data items. Index, rotation angle θ, means and SD quotients Stage 2: Recovering the secret image. 1. Recover tile images by the following steps. 2. Rotate tile in the reverse direction and fit the resulting. block content into T to form an initial tile image Use the extracted means and related SD quotients compute the original pixel value. 3. Scan T to find out pixels with values 255 or Take the results as the final pixel values. 5. Compose all the final tile images to form the desired secret image S. V. RESULTS Results of this paper is shown in bellow Figs.3 to 10. Fig. 5. Image into blocks or tiles. Fig. 6. Lab transformed image. Fig. 3. Target image. Fig.7. Secret image with attack like noises.

6 Fig.8. Generated Mosaic image with target & secret images. Fig. 9. Resultant target image with psnr value. Fig. 10. Recovered secret image with psnr value. RAAFIA FATIMA, MD ATEEQ UR RAHMAN V. CONCLUSION Images from different sources are transmitted through the internet for various applications. These images usually contain private or secret data so that they should be protected from leakages during transmissions. A method is proposed to securely transmit a secret image that create mosaic images which also can transform a secret image into a mosaic tile image with the same size of data for concealing the secret image. This is done by the use of proper color transformations pixel by pixel in mosaic tile images with large color similarities. The original secret image can be reconstructed nearly lossless from the created mosaic images. A new secure image transmission method has been proposed,which not only can create meaningful mosaic images but also can transform a secret image into a mosaic one with the same data size for use as a camouflage of the secret image. By the use of proper pixel color transformations as well as a skillful scheme for handling overflows and underflows in the converted values of the pixels colors, secret-fragment visible mosaic images with very high visual similarities to arbitrarily-selected target images can be created with no need of a target image database. Also, the original secret images can be recovered nearly losslessly from the created mosaic images. Good experimental results have shown the feasibility of the proposed method. VI. REFERENCES [1] J. Fridrich, Symmetric ciphers based on two-dimensional chaoticmaps, Int. J. Bifurcat. Chaos, vol. 8, no. 6, pp , [2] G. Chen,Y. Mao, and C. K. Chui, A symmetric image encryptionscheme based on 3D chaotic cat maps, Chaos Solit. Fract., vol. 21,no. 3, pp , [3] L. H. Zhang, X. F. Liao, and X. B. Wang, An image encryptionapproach based on chaotic maps, Chaos Solit. Fract., vol. 24, no. 3,pp , [4] H. S. Kwok and W. K. S. Tang, A fast image encryption system basedon chaotic maps with finite precision representation, Chaos Solit. Fract.,vol. 32, no. 4, pp , [5] S. Behnia, A. Akhshani, H. Mahmodi, and A. Akhavan, A novelalgorithm for image encryption based on mixture of chaotic maps, Chaos Solit. Fract., vol. 35, no. 2, pp , [6] D. Xiao, X. Liao, and P. Wei, Analysis and improvement of a chaosbasedimage encryption algorithm, Chaos Solit. Fract., vol. 40, no. 5,pp , [7] V. Patidar, N. K. Pareek, G. Purohit, and K. K. Sud, A robustand secure chaotic standard map based pseudorandom permutationsubstitutionscheme for image encryption, Opt. Commun., vol. 284,no. 19, pp , [8] C. K. Chan and L. M. Cheng, Hiding data in images by simple LSB substitution, Pattern Recognit.., vol. 37, pp , Mar [9] Z. Ni, Y. Q. Shi, N. Ansari, and W. Su, Reversible data hiding, IEEE Trans. Circuits Syst. Video Technol., vol. 16, no. 3, pp , Mar [10] J. Tian, Reversible data embedding using a difference expansion, IEEE Trans. Circuits Syst. Video Technol., vol. 13, no. 8, pp ,Aug

7 A Novel Image Broadcasting Method using Secret-Fragment-Visible Mosaic Images Through Inverse Color Transformations [11] Y. Hu, H.-K. Lee, K. Chen, and J. Li, Difference [27] Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P. expansion based reversible data hiding using two Simoncelli, Image quality assessment: From error visibility embedding directions, IEEE Trans. Multimedia, vol. 10, to structural similarity, IEEETrans. Image Process., vol. 13, no. 8, pp , Dec no. 4, pp , Apr [12] V. Sachnev, H. J. Kim, J. Nam, S. Suresh, and Y.- [28] T. R. Nielsen, P. Drewsen, and K. Hansen, Solving Q. Shi, Reversible watermarking algorithm using sorting jigsaw puzzles usingimage features, Pattern Recog. Letters, and prediction, IEEE Trans.Circuits Syst. Video vol. 29, no. 14, pp ,2008. Technol., vol. 19, no. 7, pp , Jul [29] T. S. Cho, S. Avidan, and W. T. Freeman, A [13] X. Li, B. Yang, and T. Zen, Efficient reversible probabilistic image jigsawpuzzle solver, in Proc. IEEE watermarking based on adaptive prediction-error expansion CVPR, 2010, pp and pixel selection, IEEE Trans. Image Process., vol. 20, [30] D. Pomeranz, M. Shemesh, and O. Ben-Shahar, A fully no. 12, pp , Dec automatedgreedy square jigsaw puzzle solver, in Proc. IEEE [14] W. Zhang, X. Hu, X. Li, and N. Yu, Recursive CVPR, 2011,pp histogram modification: Establishing equivalency between [31] E. Demaine and M. Demaine, Jigsaw puzzles, edge reversible data hiding and lossless data compression, matching, andpolyomino packing: Connections and IEEE Trans. Image Process., vol. 22, no. 7,pp , complexity, Graphs Combinator.,vol. 23, no. 1, pp , Jul [15] J. Fridrich, M. Goljan, and R. Du, Invertible [32] Related images of the experiments [Online]. Available: authentication, Proc.SPIE, vol. 3971, 2001, pp sfv.html [16] C. C. Chang, C. C. Lin, C. S. Tseng, and W. L. Tai, Reversible hiding in DCT-based compressed images, Inf. Sci., vol. 177, no. 13,pp , [17] S. Lee, C. D. Yoo, and T. Kalker, Reversible image watermarking basedon integer-to-integer wavelet transform, IEEE Trans. Inf. Forens. Secur.,vol. 2, no. 3, pp , Sep [18] W.-H. Lin, S.-J. Horng, T.-W. Kao, P. Fan, C.-L. Lee, and Y. Pan, An efficient watermarking method based on significant difference ofwavelet coefficient quantization, IEEE Trans. Multimedia, vol. 10, no. 5, pp , Aug [19] X. Hu, W. Zhang, X. Hu, N. Yu, X. Zhao, and F. Li, Fast estimationof optimal marked-signal distribution for reversible data hiding, IEEETrans. Inf. Forens. Secur., vol. 8, no. 5, pp , May [20] W. B. Pennebaker and J. L. Mitchell, JPEG: Still Image Data CompressionStandard. New York, NY, USA: Van Nostrand Reinhold, 1993,pp [21] I. J. Lai and W. H. Tsai, Secret-fragment-visible mosaic image A newcomputer art and its application to information hiding, IEEE Trans. Inf.Forens. Secur., vol. 6, no. 3, pp , Sep [22] E. Reinhard, M. Ashikhmin, B. Gooch, and P. Shirley, Color transferbetween images, IEEE Comput. Graph. Appl., vol. 21, no. 5, pp ,Sep. Oct [23] D. L. Ruderman, T. W. Cronin, and C. C. Chiao, Statistics of coneresponses to natural images: Implications for visual coding, J. Opt.Soc. Amer., vol. 15, no. 8, pp , [24]D.Coltuc and J.-M. Chassery, Very fast watermarking by reversiblecontrast mapping, IEEE Signal Process. Lett., vol. 14, no. 4,pp , Apr [25] R. Z. Wang, C. F. Lin, and J. C. Lin, Image hiding by optimal LSB substitution and genetic algorithm, Pattern Recog., vol. 34, no. 3,pp , [26] C. H. Yang, Inverted pattern approach to improve image quality of information hiding by LSB substitution, Pattern Recog., vol. 41, no. 8,pp , 2008.