JPEG Compression Stegenography

8/2/2019 JPEG Compression Stegenography

1/5


2/5

I142 JOURNAL OF ADVANCES IN INFORMATION TECHNOLOGY. VOL. I. NO. 3. AUGUST 2010

though, keeps the originaldigital imageintact without thechance of lost, although is does not compress the imageto such a small file size.To compress an image into JPEG format, the RGB

colour representation is first converted to a YUVrepresentation. In this representation the Y componentcorresponds to the luminance (or brightness) and the Uand V components stand for chrominance (or color).According to research the human eye is more sensitive tochanges in the brightness (luminance) of a pixel than tochanges in its color. This fact is exploited by the JPEGcompression by down sampling the color data to reducethe size of the file. The color components (U and V) arehalved in horizontal and vertical directions, thusdecreasing the file size by a factor of 2.The next step isthe actual transformationof the image.

For ]PEG [18], the Discrete Cosine Transform (DCT)[18] is used, but similar transforms are for example theDiscrete Fourier Transform (DIT). These mathematicaltransformsconvert thepixels in such a way as to give theeffect of "spreading" the locationof the pixel values overpart of the image. The DCT transforms [l8] a signal froman image representation into a frequency representation,by grouping the pixels into 8 x 8 pixel blocks andtransforming the pixel blocks into 64 DCT coefficientseach. A modification of a single DCT coefficient willaffect all 64 image pixels inthat block.The next step is the quantization [18] phase of the

compression. Here another biological property of thehuman eye is exploited: The human eye is fairly good atspotting small differences in brightness over a relativelylarge area, but not so good as to distinguish betweendifferent strengths in high frequency brightness. Thismeans that the strength of higher frequencies can bediminished, without changing the appearance of theimage. JPEG does this by dividing all the values in ablock by a quantization coefficient. The results arerounded to integer values and the coefficients areencoded usingHuffman coding to further reduce the size.Originally it was thought that steganography would

not be possible to use with JPEG images, since they uselossy compression [3] which results inparts of the imagedata being altered. One of the major characteristics ofsteganography is the fact that information ishidden in theredundant bits of an object and since redundant bits areleft out when using JPEG it was feared that the hiddenmessagewould be destroyed. Even if one could somehowkeep the message intact it would be difficult to embed themessage ~ ithout the changes being noticeable because ofthe harsh compression applied. However, properties ofthe compression algorithm have been exploited in orderto developa steganographicalgorithmfor JPEGs.One of these properties of JPEG is exploited to make

the changes to the image invisible to the human eye.During the DCT transformationphase of the compressionalgorithm, rounding errors occur in the coefficient datathat are not noticeable. Although this property is whatclassifies the algorithm as being lossy, this property canalso be used to hide messages

C 2010 ACADEMY PUBLISHER

It is neither feasible nor possible to embed informationin an image that uses lossy compression, since thecompression would destroy all information in theprocess. Thus it is important to recognize that the JPEGcompression algorithm is actually divided into lossy andlossless stages [3]. The DCf and the quantization phaseform part of the lossy stage, while the Huffman encodingused to further compress the data is lossless.Steganography can take place between these two stages.Using the same principles of LSB insertionthe messagecan be embedded into the least significant bits of thecoefficients before applying the Huffman encoding. Byembedding the information at this stage, in the transformdomain, it is extremely difficult to detect, since it is not inthe visual domain.

III. PROPOSEDSYSTEMWe propose a framework for hiding large volumes of

data in images while incurring minimal perceptualdegradation. The embedded data can be recoveredsuccessfully, without any errors, after operations such asdecompression,additive noise, and image tampering. Theproposed methods can be employed for applications thatrequire high-volume embedding with robustness againstcertain non-malicious attacks. The hiding methods wepropose are guided by the growing literature on theinformationtheory of data hiding [22].The key novelty of our approach is that our coding

framework permits the use of local criteria to decidewhere to embed data. In order to robustly hide largevolumes of data in images without causing significantperceptual degradation, hiding techniques must adapt tolocal characteristics within an image. The mainingredients of our embedding methodology are asfollows.(a) As is well accepted, data embedding is done in thetransform domain, with a set of transform coefficients inthe low and mid frequency bauds selected as possiblecandidates for embedding. (These are preserved betterunder compression attacks than high frequencycoefficients)(b) A novel feature of our method is that, from the

candidate set of transform coefficients, the encoderemploys local criteria to select which subset ofcoefficients it will actually embed data in. In exampleimages, the use of local criteria for deciding where toembed isfoundto be crucial to maintainingimage qualityunder high volume embedding.(c) For each of the selected coefficients, the data to be

embedded indexes the choice of a scalar quantizer forthat coefficient. We motivate this by informationtheoreticanalysis.(d) The decoder does not have explicit knowledge of

the locations where data is hidden, but employs the samecriteria as the encoder to guess these locations. Thedistortion due to attacks may now lead to insertion errors(the decoder guessing that a coefficient has embeddeddata, when it actually does not) and deletion errors (thedecoder guessing that a coefficient does not have


3/5


4/5

144 JOURNAL OF ADVANCES IN INFORMA nON TECHNOLOGY, VOL. L NO.3. AUGUST 2010

IV. RESULT ANALYSISAll steganographic algorithms have to comply with afew basic requirements. The requirements are:Invisibility. Payload capacity, Robustness against

statistical attacks, Robustness against imagemanipulation. Independent of file format andUnsuspicious files. The following table compares leastsignificant bit (LSB) insertion in BMP and in GIF files,JPEG compression steganography, the patchworkapproach and spread spectrum techniques, according tothe above requirements:TABLEr.

COMPARISON OF IMAGE STEGANOGRAPHY ALGORITHMSLSB LSBin JPEG Patch Spreadin GlF compressi work spectrumBMP on

Invisibility High" Mediu High High Highm*

Payload High Mediu Medium Low Mediumcapacity mRobustnes Low Low Medium High Highs againststatisticalattacksRobustnes Low Low Medium High Mediums againstimagemanipulationlndepende Low Low Low High Highnt of fileformatUnsuspici Low Low High High Highous files

- Depends on cover lfiJUge usedThe levels at which the algorithms satisfy therequirements are defined as high, medium and low. Ahigh level means that the algorithm completely satisfiesthe requirement, while a low level indicates that thealgorithm has a weakness in this requirement A mediumlevel indicates that the requirement depends on outsideinfluences, for example the cover image used. LSB inGIF images has the potential of hiding a large message,but only when the most suitable cover image has been

chosen.The ideal, in other words a perfect; steganographicalgorithm would have a high level in every requirement.Unfortunately in the algorithms that are evaluated here,there is not one algorithm that satisfies all of therequirements. Thus a trade-off will exist in most cases,depending on which requirements are more important furthe specific application.TIle process of embedding information during JPEGcompression results in a stego image with a high level ofinvisibility, since the embedding takes place in thetransform domain. JPEG is the most popular image fileformat on the Internet and the image sizes are smallbecause of the compression, thus making it the leastsuspicious algorithm to use. However, the process of the

02010 ACADEMY PUBLISHER

compression is a very mathematical process, making itmore difficult to implement. The JPEG file format can beused for most applications of steganography, but isespecially suitable for images that have to becommunicated over an opeu systems environment likethe Internet.

V. CONCLUSION AND SCOPE FOR FUTUREWORKThe meaning of Steganography is hiding informationand the related technologies. There is a principaldifference between Steganography and Encryption;however they can meet at some points too. They can beapplied together, i.e. encrypted information can be hiddenin addition. To hide something a covering medium isalways needed. (Picture, sound track, text or even thestructure of a file system, etc.) The covering mediummust be redundant; otherwise the hidden information

could be detected easily. The technology of hiding shouldmatch the nature of the medium. The hidden informationshould not be lost, if the carrying medium is edited,modified, formatted, re-sized, compressed or printed.That's a difficult task to realize.The application isprimarily intended to be used to inconspicuously hideconfidential and proprietary information by anyoneseeking to hide information. This software has anadvantage over other information security systemsbecause the hidden text are inthe form of image, which isnot obvious text information carriers.Because of its user-friendly interface, the applicationcan also be used by anyone who wants to securelytransmit private information. The main advantage of thisprogram for individuals is that they do not have to haveany knowledge about steganography or encryption. Thevisual way to encode the text, plus the visual key makesit easy for average users to navigate within the program.Digital Image Steganography system allows anaverage user to securely transfer text messages by hidingthem in a digital image file. A combination ofSteganography and encryption algorithms provides astrong backbone for its security. Digital ImageSteganography system features innovative techniques forhiding text in a digital image file or even using it as a keyto the encryption.Digital Image Steganography [2J system allows a userto securely transfer a text message by hiding it in a digitalimage file. 128 bit AES encryption is used to protect thecontent of the text message even if its presence were tobe detected. Currently, no methods are known forbreaking this kind of encryption within a reasonableperiod of time (i.e., a couple of years). Additionally,compression is used to maximize the space available inan image.To send a message, a source text, an image in whichthe text should be embedded, and a key are needed. Thekey is used to aid in encryption and to decide where theinformation should be hidden in the image. A short textcan be used as a key. To receive a message, a sourceimage containing the information and the corresponding


5/5

IJOURNAL OF ADVANCES IN rNFORMA nON TECHNOLOOY, VOL. l. NO.3, AUGUST 2010

key are both required. The result will appear in the textlab after decoding.

The common Internet-friendly format is offered. It isinherently more difficult to hide information in a JPEGimage because that is exactly what the designers of JPEGwanted to avoid: the transmission of extra informationthat doesn't affect the appearance of the image.

ACKNOWLEDGEMENTThe work on this paper was supported by the Bharati

Vidyapeeth University & 'College of Engineering, Pune.The views and conclusions contained herein are those ofthe authors and the paper contains the original work ofthe authors. We took help from many books, papers andother materials.

REFERENCES[IJ N. Provos, "Defending Against Statistical Steganography,"

Proc lOth USENEX Security Symposium 2005.[2J N Provos and P Honeyman, Hide and Seek: All

introduction to Steganography," IEEE Security & PrivacyJoumal2003.

[3J Steven W Smith , The Scientist and Engineer's Guide toDigital Signal Processing

[4J Katzenbeisser and Petitcolas , "Information HidingTechniques for Stenography and Digital watermaking"Artech House, NOIWood. MA. 2000 .

(SJ L. Reyzen And S. Russell, More efficient provably secureSteganography" 2007.

(6] S.Lyu and H. Farid, "Steganography using higher orderimage statistics. " IEEE Trans. Inf. Forens. Secur. 2006.

[7J Venkatraman, s, Abraham, A . & Paprzycki M."Significance of Steganography on Data Security ,Proceedings of the International Conference onInformation Technology: Coding and computing, 2004.

1'8) Fridrich,.r., Goljan M., and Hogea , D ; New Methodologyfor Breaking stenographic Techniques for JPEGs. "Electronic Imaging 2003".

(9J http:! aakash.ece.ucsb.edu.z data hiding !stegdemo.aspx.Ucsb data hiding online demonstration.Released on Mar .09.2005.

[10] Mitsugu Iwamnoto and Hirosuke Yamamoto, "TheOptimal n-out-of-n Visual Secret Sharing Scheme forGray Scale Images", IEICE Trans. Fundamentals, voLE85-A, No. 10, October 2002, pp. 2238-2247.

[II] Doron Shaked, Nur Arad, Andrew Fitzhugh, Irwin Sobel."Color Diffusion: Error Diffusion for Color Halftones",I-:!PLaboratories Israel, May 1999.

[I2J ZZhou, G.R.Arce, and G.Di Crescenzo, "Halftone VisualCryptography". IEEE Tans. On Image Processing,voLl5,No.8, August 2006, pp. 2441-2453.

[I3J M.Naor and A.Shamir, "Visual Cryptography", inProceedings of Eurocrypt 1994, lecture notes in computerscience, 1994, vo1.9S0, pp. J -12.

(14] Robert Ulichney, "The void-and-cluster method for ditherarray generation", IS&T/SPIE Symposium on ElectronicTmaging and Science, San Jose, CA, 1993, voI.1913,pp.332-343.

[I S] ERVerheul and RC.A. Van Tilborg, "Constructions andproperties of k out of n visual secret sharing scheme",Designs, Codes, and Cryptography, vol.I, no.2, 1997,pp.179-1%.

020 t o ACADEMY PUBLISHER

145

(16J Daniel L.Lau, Robert Ulichney, Gonzalo RArce,"Fundamental Characteristics of Halftone Textures: Blue-Noise and Green-Noise", Image Systems Laboratory, HPLaboratories Cambridge, March 2003.

[17] C. Yang and C.Laih, "New colored visual secret sharingschemes", Designs, Codes and Cryptography, vo1.20,2000, pp.32S-33S.[18] Jain, Anil K., "Fundamentals of Digital ImageProcessing", Prentice-Hall ofIndia, 1989

[19J C.Chang, C.Tsai. and TChen, "A new scheme for sharingsecret color images in computer network", in Proc. ofInternational Conference on Parallel and DistributedSystems, 2000, pp. 21-27.

(20) R.L.Alder, B.P.Kitchens, M.Martens, "The mathematics ofhalftoning", IBM J. Res. & Dev. Vo1.47 No.1, Jan. 2003,pp.5-15.

[2IJ R.Lukac, K.N.Plantaniotis, B.Smolka, "A new approach tocolor image secret sharing", EUSIPCO 2004, pp. 1493-1496.

[22J H.Allcin, Anoop KBhattacharjya, Joseph Shu, Improvingvoid-and-cluster for better halftoneunifonnity",lnternational Conference on Digital PrintingTechnoogies,

[23J D. Hankerson, P. D. Johnson, and G. A. Harris,"Introduction to Information Theory and DataCompression".

(24] Ranjan Bose, "Information Theory Coding andCryptography".

Meenu Kumari- Completed B.E. in Information Technologyfrom Sanjivani Educational Society & College of Engineering,Kopargaon, Pune University in 2005. Persuing M. Tech. IT fromBharati Vidyapeeth University CoUege of Engineering, Pune.Presented one national comference on Image Compression.Published research paper in one e-journal & one internationaljournals. Submitted research paper in other national &international journals for publication.

Prof. A. Khare- Completed B.E. and M.E. from Bhopal.Currently working as Assistant Profes ..sor, in Bharati VidyapeethUniversity College of Engineering, Information TechnologyDepartment, Pune. Presented many national & internationalconferences & journals.

PaUa"i Khare- Research student of SSS[ST, E&TCDepartment, Bhopal, India.

JPEG Compression Stegenography

Documents