ISSN:2249-5789 Sunita Barve et al, International Journal of Computer Science & Communication Networks,Vol 1(1),September-October 2011

Efficient and Secure Biometric Image Stegnography using Discrete Wavelet Transform
Sunita Barve, Uma Nagaraj and Rohit Gulabani Department of Computer Engineering Maharashtra Academy of Engineering, Alandi

Abstract
Steganography is the science of concealing the existence of data in another transmission medium. It does not replace cryptography but rather boosts the security using its obscurity features. As proposed method is Biometric Steganography, here the Biometric feature used to implement Steganography is Skin tone region of images. Proposed method introduces a new method of embedding secret data within the skin portion of the image of a person, as it is not that much sensitive to HVS (Human Visual System). Instead of embedding secret data anywhere in image, it will be embedded in only skin tone region. This skin region provides excellent secure location for data hiding. So, firstly skin detection is performed in cover images and then Secret data embedding will be performed in DWT domain as DWT gives better performance than DCT while compression. This biometric method of Steganography enhances robustness than existing methods. downgrading and contamination which is now known as Steganography. Cyberterrorism, as coined recently, is believed to benefit from this digital revolution.

I. INTRODUCTION
Steganography is defined as science or art of hiding (embedding) data in transmission medium. Steganography is a type of hidden communication that literally means covered writing (from the Greek words stegano or covered and graphos or to write). The first use of the term steganography was recorded in 1499 by Johannes Trithemius in his Steganographia, a dissertation on cryptography and steganography disguised as a book on magic. The goal of steganography is to hide an information message inside harmless cover medium in such a way that it is not possible even to detect that there is a secret message. Oftentimes throughout history, encrypted messages have been intercepted but have not been decoded. While this protects the information hidden in the cipher, the interception of the message can be just as damaging because it tells an opponent or enemy that someone is communicating with someone else. Steganography takes the opposite approach and attempts to hide all evidence that communication is taking place. Figure 1 elaborates the idea of steganography messages by depicting a scenario where two prison inmates Bob and Alice try to communicate via the warden Wendy. Inspired by the notion that Steganography can be embedded as part of the normal printing process, Japanese firm Fujitsu is pushing technology to encode data into a printed picture that is invisible to the human eye (i.e., data) but can be decoded by a mobile phone with a camera. Figure 1: The prisoners problem

III. IMAGE STEGANOGRAPHY II. LITERATURE SURVEY

The earliest recordings of Steganography were by the Greek historian Herodotus in his chronicles known as "Histories" and date back to around 440 BC. In the 15th and 16th century, Romans used invisible inks, which were based on natural substances such as fruit juices and milk. During the times of WWI and WWII, significant advances in Steganography took place. Concepts such as null ciphers (taking the 3rd letter from each word in a harmless message to create a hidden message, etc), image substitution and microdot (taking data such as pictures and reducing it to the size of a large period Piece of paper) were introduced and embraced as great Steganographic techniques. With the boost of computer power, the internet and with the development of Digital Signal Processing (DSP), Information Theory and Coding Theory, Steganography went Digital. In the realm of this digital world Steganography has created an atmosphere of corporate vigilance that has spawned various interesting applications of the science. Contemporary information hiding was first discussed in the article The prisoners Problem and the Subliminal Channel [The prisoners problem]. More recently Kurak and McHugh carried out work which resembled embedding into the 4LSBs (Least Significant Bits). They discussed image Image steganography techniques can be divided into two groups: those in the Image Domain and those in the Transform Domain. Image also known as spatial domain techniques embed messages in the intensity of the pixels directly, while for transform also known as frequency domain, images are first transformed and then the message is embedded in the image. Image domain techniques encompass bit-wise methods that apply bit insertion and noise manipulation and are sometimes characterized as simple systems. The image formats that are most suitable for image domain steganography are lossless and the techniques are typically dependent on the image format. Steganography in the transform domain involves the manipulation of algorithms and image transforms. These methods hide messages in more significant areas of the cover image, making it more robust. Many transform Given the proliferation of digital images, especially on the Internet, and given the large amount of redundant bits present in the digital representation of an image, images are the most popular cover objects for steganography. In the domain of digital images, many different image file formats exist, most of them for specific applications. For these different image file formats, different steganographic algorithms exist. A. Image and Transform Domain

Available online @ www.ijcscn.com

96

ISSN:2249-5789 Sunita Barve et al, International Journal of Computer Science & Communication Networks,Vol 1(1),September-October 2011 domain methods are independent of the image format and the embedded message may survive conversion between lossy and lossless compression. In the next sections steganographic algorithms will be explained in categories according to image file formats and the domain in which they are performed. segmentation. However, several undesirable features of these colorspaces can be pointed out.

IV. SKIN DETECTION Skin color has proven to be a useful and robust cue for face detection, localization and tracking. Image content filtering, content aware video compression and image color balancing applications can also benefit from automatic detection of skin in images. Face detection and tracking has been the topic of an extensive research for several decades. Many heuristic and pattern recognition based strategies have been proposed for achieving robust and accurate solution. Among feature-based face detection methods, the ones using skin color as a detection cue have gained strong popularity. Color allows fast processing and is highly robust to geometric variations of the face pattern. Also, the experience suggests that human skin has a characteristic color, which is easily recognized by humans. When building a system, that uses skin color as a feature for face detection, the researcher usually faces three main problems. First, what color space to choose, second, how exactly the skin color distribution should be modeled, and finally, what will be the way of processing of color segmentation results for face detection. Figure 3:- RGB to YCbCr Transformation D. Skin Modelling:The final goal of skin color detection is to build a decision rule that will discriminate between skin and non-skin pixels. This is usually accomplished by introducing a metric, which measures distance (in general sense) of the pixel color to skin tone. The type of this metric is defined by the skin color modeling method. One method to build a skin classifier is to define explicitly (through a number of rules) the boundaries skin cluster in some color space. For example:-

V. COLOUR SPACES Colorimetry, computer graphics and video signal transmission standards have given birth to many color spaces with different properties. A wide variety of them have been applied to the problem of skin color modeling. A. RGB RGB is a color space originated from CRT (or similar) display applications, when it was convenient to describe color as a combination of three colored rays (red, green and blue). It is one of the most widely used color spaces for processing and storing of digital image data. However, high correlation between channels, significant perceptual non-uniformity, mixing of chrominance and luminance data makes RGB not a very favorable choice for color analysis and color based recognition algorithms. B. YCbCr YCbCr is an encoded nonlinear RGB signal, commonly used by European television studios and for image compression work. Color is represented by luma (which is luminance, computed from nonlinear RGB), constructed as a weighted sum of the RGB values, and two color difference values Cr(Chrominance red) and Cb(Chrominance blue) that are formed by subtracting luma from RGB red & blue components. The transformation simplicity and explicit separation of Luminance and chrominance components make this color space attractive for skin color modeling. The simplicity of this method has attracted (and still does) many researchers. The obvious advantage of this method is simplicity of skin detection rules that leads to construction of a very rapid classifier. The main difficulty achieving high recognition rates with this method is the need to find both good color space and adequate decision rules empirically. Recently, there have been proposed a method that uses machine learning algorithms to find both suitable color space and a simple decision rule that achieve high recognition rates. The authors start with a normalized RGB space and then apply a constructive induction algorithm to create a number of new sets of three attributes being a superposition of r, g, b and a constant 1/3, constructed by basic arithmetic operations. A decision rule, which achieves the best possible recognition, is estimated for each set of attributes. The authors prohibit construction of too complex rules, which helps avoiding data over-fitting that is possible in case of lack of training set representativeness. They have achieved results that outperform Bayes skin probability map classifier in RGB space for their dataset. E. Masking and filtering:Masking and filtering techniques usually restricted to 24 bits or grayscale images for hiding a message. These methods are similar to paper watermarks, creating markings in an image. This is achieved for example by modifying the luminance of parts of the image. While masking changes the visible properties of an image, it can be done in such a way that the human eye will not notice the anomalies. Generally masking uses visible aspects of the image; also it is more robust than LSB modification with respect to compression, cropping and different kinds of image processing. Although the information is not hidden at the noise level , rather than it is inside the visible part of the image, which makes it more suitable than LSB modifications in case a lossy compression algorithm like JPEG is being used. In skin detection algorithms masking basically means covering the non Figure 2:- RGB to YCbCr Transformation C. HSV (Hue, Saturation, Value) Hue-saturation based colorspaces were introduced when there was a need for the user to specify color properties numerically. They describe color with intuitive values, based on the artists idea of tint, saturation and tone. Hue defines the dominant color (such as red, green, purple and yellow) of an area; saturation measures the colorfulness of an area in proportion to its brightness. The intensity, lightness or value is related to the color luminance. The intuitiveness of the color space components and explicit discrimination between luminance and chrominance properties made these colorspaces popular in the works on skin color skin region with a black mask. Filtering is replacing the white region (representing the skin portion in the binary image) with the original skin portion in the cover image. The masking and filtering operation is shown in Fig 2 below:Figure 4:- Masking & Filtering Operation

Available online @ www.ijcscn.com

97

ISSN:2249-5789 Sunita Barve et al, International Journal of Computer Science & Communication Networks,Vol 1(1),September-October 2011

VI. 2D HAAR DWT

The frequency domain transform we applied in this research is Haar-DWT, the simplest DWT. A 2-dimensional Haar- DWT consists of two operations: One is the horizontal operation and the other is the vertical one. Detailed procedures of a 2-D Haar-DWT are described as follows: Step 1: At first, scan the pixels from left to right in horizontal direction. Then, perform the addition and subtraction operations on neighboring pixels. Store the sum on the left and the difference on the right as illustrated in Figure 4. Repeat this operation until all the rows are processed. The pixel sums represent the low frequency part (denoted as symbol L) while the pixel differences represent the high frequency part of the original image (denoted as symbol H).

1. Select the message to be embedded, preprocess it. 2. Select the cover image, with optimum number of skin pixels 3. Segment out the skin pixels. 4. Select the particular region of skin, where to embed the message. (Hereafter referred as Skin ROI) 5. Take 2D wavelet transform of the skin ROI.

Figure 5:- Horizontal Operation on the first row

Step 2: Secondly, scan the pixels from top to bottom in vertical direction. Perform the addition and subtraction operations on neighboring pixels and then store the sum on the top and the difference on the bottom as illustrated in Figure 5. Repeat this operation until all the columns are processed. Finally we will obtain 4 sub-bands denoted as LL, HL, LH, and HH respectively. The LL sub-band is the low frequency portion and hence looks very similar to the original image.

Figure 8:- Proposed Encoder

6. Select the particular sub band to be used for embedding. 7. Perform the payload check. 8. Embed the preprocessed data in particular sub band. 9. Perform Inverse DWT

Figure 6:- Vertical Operation

The whole procedure which has been described above is called the first-order 2-D Haar-DWT. The first-order 2-D Haar- DWT applied on the image Lena is illustrated in Figure 6.

10. We obtain the stego image.

Fig 9. Region of Interest The decoder end has been purposefully not been kept exactly inverse of the encoder. Keeping in mind the general tendency of assuming the decoder to be the Figure 7:- (a) Original image-Lena, (b) Result after the first-order 2-D HaarDWT inverse of encoder, we have implemented the IDWT on the encoder end and no inverse transform on the decoder end. This acts as a very important security feature and increases the robustness of our algorithm.

VII. PROPOSED METHOD FOR DIGITAL IMAGE STEGNOGRAPHY

We are proposing a Digital Image Steganography technique which explores the Biometric feature of Skin tone region in images. This system should be capable of embedding secret images using the Discrete Wavelet Transform. The primary goal driving us is building a robust and secure Steganography system with additional security features as compared to the existing systems. Proposed method creating high quality stego image their by increasing security.

The following are the steps to be implemented while decoding the stego-image. 1. 2. 3. 4. 5. Apply key 1 to the stego image Perform DWT on the selected sub band Carry out LSB extraction to get a distorted image Reduce noise components in the image Result is the hidden image

VIII. ADDITIONAL FEATURE

Whenever security enhancements have been proposed till date, the first question that arises in the mind of a third person is to try breaking it. History is witness to this fact, and hence as an outcome we have ever advancing fields of

The detailed algorithm for the encoder is discussed below:-

Available online @ www.ijcscn.com

98

ISSN:2249-5789 Sunita Barve et al, International Journal of Computer Science & Communication Networks,Vol 1(1),September-October 2011 Cryptography, Steganography and so forth. Our proposed framework already encompasses three-tier robustness with the existence of two keys and IDWT on the encoder end. We know that there is no bound to the thoughts of a negative mind. Yet considering a scenario in which a Hacker manages to obtain the stego image along with the two keys. In such a case too he shall be unsuccessful in decoding the message. As mentioned earlier, our proposed decoder is not the exact inverse of the encoder. Thus if the Hacker tries to reverse the encoder upon the stego image and if he tries to fiddle with the bits, all he obtains is a completely distorted image. This challenge is a proof to the robustness in steganography that were proposing. Fig 11. Comparison of Cover and Stego Image

6.

X. CONCLUSION
Digital Steganography is a fascinating scientific area which falls under the umbrella of security systems. Proposed framework is based on steganography that uses Biometric feature i.e. skin tone region. Skin tone detection plays a very important role in Biometrics and can be considered as secure location for data hiding. Secret data embedding is performed in DWT domain than the DCT as DWT outperforms than DCT. Using Biometrics resulting stego image is more tolerant to attacks and more robust than existing methods.

REFERENCES
1. Abbas Cheddad, Joan Condell, Kevin Curran and Paul Mc Kevitt, Biometric Inspired Digital Image Steganography, 15th Annual IEEE International Conference and Workshop on the Engineering of Computer Based Systems, 978-0-7695-3141-0/08 $25.00 2008 IEEE DOI 10.1109/ECBS.2008.11.159) Abbas Cheddad, Joan Condell, Kevin Curran and Paul Mc Kevitt, A Skin Tone Detection Algorithm for an Adaptive Approach to Steganography, Faculty of Computing and Engineering, University of Ulster, BT48 7JL, Londonderry, Northern Ireland, United Kingdom. Po- Yueh Chen and Hung-Ju Lin, A DWT Based Approach for Image Steganography, Department of Computer Science and Information Engineering, National Changhua University of Education, No. 2 Shi-Da Road, Changhua City 500, Taiwan, R.O.C. Vladimir Vezhnevets and Vassili Sazonov, A Survey on Pixel-Based Skin Color Detection Techniques, Alla Andreeva Graphics and Media Laboratory, Faculty of Computational Mathematics and Cybernetics Moscow State University, Moscow, Russia. Neil F. Johnson and Sushil Jajodia, Steganalysis: The Investigation of Hidden Information, IEEE conference on Information Technology, pp. 113-116, 1998. Lisa M.Marvel and Charles T. Retter, A Methodlogy for Data Hiding using Images, IEEE conference on Military communication, vol. 3, Issue. 18-21, pp. 1044-1047, 1998. Giuseppe Mastronardi, Marcello Castellano, Francescomaria Marino, Steganography Effects in Various Formats of Images. A Preliminary Study, International Workshop on Intelligent data Acquisition and Advanced Computing Systems: Technology and Applications, pp. 116119, 2001. LIU Tong, QIU Zheng-ding A DWT-based color Images steganography Scheme IEEE International Conference on Signal Processing, vol. 2, pp.1568-1571, 2002. Jessica Fridrich, Miroslav Goijan and David Soukal, Higher-order statistical steganalysis of palette images Proceeding of SPIE, Electronic Imaging, Security, Steganography, and Watermarking of a. Multimedia ContentsV, vol. 5020, pp. 178-190, 2003. Jessica Fridrich and David Soukal, Matrix Embedding for Large Payloads SPIE, Electronic Imaging, Security, Steganography, and Watermarking of Multimedia Contents , vol. 6072, pp. 727-738. 2006. Yuan-Yu Tsai, Chung-Ming Wang A novel data hiding scheme for color images using a BSP tree Journal of systems and software, vol.80, pp. 429437, 2007. Jun Zhang, Ingemar J. Cox and Gwenael Doerr.G Steganalysis for LSB Matching in Images With High-frequency Noise IEEE Workshop on Multimedia Signal Processing, issue 1-3, pp.385- 388, 2007. M. Mahdavi, Sh. Samavi, N. Zaker and M. Modarres-Hashemi, Steganalysis Method for LSB Replacement Based on Local Gradient of Image Histogram, Journal of Electrical and Electronic Engineering, vol. 4, no. 3, pp. 59-70, 2008.

2.

3.

Figure 10:- Proposed Decoder 4.

IX. ANALYSIS
To establish an objective criterion for digital image quality, a parameter named PSNR (Peak Signal to Noise Ratio) is defined as follows: 5.

6. MSE (Mean Square Error) stands for the mean-squared difference between the cover-image and the stego-image. The mathematical definition for MSE is: 7.

Where aij means the pixel value at position (i, j) in the cover image and bij means the pixel value at the same position in the corresponding stego-image. The calculated PSNR usually adopts dB value for quality judgment. The larger PSNR is, the higher the image quality is (which means there is only little difference between the cover-image and the stego-image). On the contrary, a small dB value of PSNR means there is great distortion between the cover-image and the stego-image. For color images, the reconstruction of all three color spaces must be considered in the PSNR calculation. The MSE is calculated for the reconstruction of each color space. The average of these three MSEs is used to generate the PSNR of the reconstructed RGB image (as compared to the original 24-bit RGB image). The color PSNR equations are as follows:-

8.

9.

10.

11.

12.

13.

MSE red (or green or blue) is similar to the main MSE equation for each color space.

Available online @ www.ijcscn.com

99