Abstract. The phenomenal growth in Internet applications in the past few years has led to a genuine need,

Transcription

1 Abstract The phenomenal growth in Internet applications in the past few years has led to a genuine need, as well as a sense of urgency, for both small office and home office and corporate users to protect their data transactions through the Internet. These data transactions may include sensitive document transfer, digital signature authentication, digital watermarking for copyright protection, and digital data storage and linkage. For years, people have devised different tools for sing the data secretly. Sing the data secretly is needed in medical, military and other fields. Steganography is a technique that allows one to hide binary data within a medium such as text, image, audio and video while adding few noticeable changes to the original medium. Technological advancements over the past decade or so have brought terms like mp3, jpeg, and mpeg into our everyday vocabulary. These lossy compression techniques l themselves perfectly for hiding data. Steganography exploits the use of a host data or message (also known as a container) to hide or embed another data or message into it. Unlike encryption, the host data or container used in steganography is not scrambled or hidden during the communication process. Another key advantage of the lossless algorithm is the option to select any digital data file from a webpage on the Internet. As the algorithm does not corrupt or overwrite the container file, multimedia data posted on any webpage, such as images (JPEG, GIF), video clips (AVI, MPEG) or audio files (WAV, MIDI), can be selected as the container file. Furthermore, customized container files, such as the voices and images of the ser captured via video conferencing, can be generated very easily. Therefore, the probability of knowing which container file used during encoding is infinitesimally small. It is almost like "finding a needle in a haystack." A multimedia container file is first chosen from the PC hard disk or from a webpage on the Internet. The knowledge of this container file must be pre-determined and communicated 1

2 securely between the ser and receiver. The algorithm generates a hash file or stegfile from the inputs of the container file and the hidden text. The stegfile contains random data based on a number of mathematical operations between the two input files. The random data bears no data resemblance to either the container or the hidden file. In this project audio and image steganography are comparatively analyzed and it was found that it is better to hide in least significant bits for an image and it is better to hide in peaks for audio. There are different techniques for embedding the data but it is felt that embedding in peaks and least significant bits best suits for these media files so that we can reduce the noise in case of audio and the hidden image will not be revealed in case of image. Different parameters like capacity, robustness, speed and easiness are discussed for both of the techniques and analyzed. 2

3 Chapter 1: Introduction and Importance of Steganography In this chapter, steganography is introduced and its importance when compared to other techniques of embedding and its history are briefly described. 1.1 Introduction To human eyes data usually contains known forms like images, s, sounds and text. Most Internet data usually contains gratuitous headers too. Internet users frequently need to store, s or receive private information. The most common way to do this is to transform the data into a different form. The resulting data can be understood only by those who know how to return it to its original form. This method of protecting information is known as encryption. The major drawback to encryption is that the existence of data is not hidden. Data that has been encrypted, although unreadable, still exists as data. If enough time is given, someone could eventually decrypt the data. A solution to this problem is steganography. Digital steganography is the art of inconspicuously hiding data within an image, audio and video. In the past, people used hidden tattoos or invisible ink to convey steganographic content. Today, computer and network technologies provide easy-to-use communication channels for steganography. Essentially, the information-hiding process in a steganographic system starts by identifying a cover medium s redundant bits (those that can be modified without destroying that medium s integrity). The embedding process creates a stego medium by replacing these redundant bits with data from the hidden message. Modern steganography s goal is to keep its presence undetectable, but steganographic systems because of their invasive nature leave behind detectable traces in the cover medium. 3

4 1.2 Definition Steganography is the art and science of hiding information by embedding a message within another message. The term steganography is derived from the Greek words steganos meaning covered or hidden and graphy meaning writing. Steganography has expanded to include covert inclusion of data in any other data source including audio, text and video. The difference is where cryptography is designed to protect the content of the message, steganography is designed to hide the fact that a message even exists. Steganalysis is the science of detecting the presence of a steganographic message. The goal of steganography is to hide messages inside other harmless messages in a way that does not allow anyone to even detect that there is a second secret message present. Steganography includes a vast array of techniques for hiding messages in a variety of media. Among these methods are invisible inks, microdots, digital signatures, covert channels and spread spectrum communications. Today, steganography is used on text, images, audio signals and more. It should be noted that the hidden message does not need to be encrypted to qualify as steganography. The message itself can be in plain English and still be a hidden message. 1.3 Steganography Terms Carrier File A file which has hidden information inside of it. Steganalysis The process of detecting hidden information inside of a file. Stego-Medium The medium in which the information is hidden. Redundant Bits Pieces of information inside a file which can be overwritten or altered without damaging the file. Payload The information which is to be concealed. 4

5 1.4 Digital Steganography We now live in the information age; more precisely we now live in the age of digital information. Digital information is peculiar compared to analog information in that there is a definite state to represent any data that comprises information. Data is raw, and generally a computer system is a two state machine, it is either on or off, 1 or 0. Data is stored in either a 1 or 0 states and by combining data and context we get information. It is the context and the abstraction layers around the information that digital steganography utilizes. The data of the original message is encoded in 1s or 0s but it is part of a cover message that will generally be made up of more data than the actual message. So, the cover message is important. Most information flows over the Internet, this traffic nowadays is in the form of TCP/IP communications, and files play a large part in this communication as well. Files of a raw text nature perhaps offer us less chance to create a steganography message, though it is still possible. Image and sound files are ubiquitous on the Internet mainly because of the web, and its multimedia approach to information presentation. Image and sound files are also transmitted on other protocols than just http and https, ftp, and other application specific protocols are also used to transmit them. People also store photography and sound for their own pleasure, personal photography is very common perhaps more so than personal sound generation. It is in this mix of interpretative data that digital Steganography can exist. 1.5 History of Steganography The Greek historian Herodotus [2] documented the earliest records of steganography. The writing medium of the time was indeed text. He describes how a man named Harpagus killed a hare and hid a message inside its belly. Then he sent the hare with a messenger disguised as a 5

6 hunter. Quite another clever way of hiding data was also demonstrated. When the king Darius in Susa held Histiaeus as a prisoner during the 5th century BCE, he had to s a secret message to his son-in-law Aristagoras in Miletus[2]. Histiaeus shaved the head of slave and tattooed a message on his scalp. When the slave's hair had grown long enough he was dispatched to Miletus. When the slave reached his destination, he was shaved and the message recovered. The writing medium of the time was text, written on wax-covered tablets. Demeratus, a Greek, needed to notify Sparta that Xerxes inted to invade Greece. To avoid capture, he scraped the wax off of the tablets and wrote the message on the underlying wood. Then he covered the tablets with wax again. The tablets appeared to be blank and unused so they passed inspection. Invisible inks have always been a popular method of steganography. Ancient Romans used to write between lines using invisible inks based on readily available substances such as fruit juices, urine and milk. When heated, the invisible inks would darken, and become legible. Invisible inks were used as recently as World War II. With the improvement of technology and the ease as to the decoding of these invisible inks, more sophisticated inks were developed which react to various chemicals. Some messages had to be developed much as photographs are developed with a number of chemicals in processing labs. Null ciphers (unencrypted messages) were also used. The real message is camouflaged in an innocent sounding message. Due to the sound of many open coded messages, the suspect s communication was detected by mail filters. However innocent messages were allowed to flow through. An example of a message containing such a null cipher is Apparently neutral s protest is thoroughly discounted and ignored. Isman hard hit blockade issue affects pretext for embargo on by-products, ejecting suet s and vegetable oils. 6

7 Taking the second letter in each word the following message emerges: Pershing sails from NY June 1. Steganography has been a loyal cousin to cryptography for a very long period of time. With the computer age, steganography has been given a marvelous boost. We are sure to see a great expansion of steganographical techniques in the coming years. 1.6 Importance of Steganography Since the rise of the Internet one of the most important factors of information technology and communication has been the security of information. Cryptography was created as a technique for securing the secrecy of communication and many different methods have been developed to encrypt and decrypt data in order to keep the message secret. Unfortunately it is sometimes not enough to keep the contents of a message secret, it may also be necessary to keep the existence of the message secret. The technique used to implement this, is called steganography. Steganography differs from cryptography in the sense that where cryptography focuses on keeping the contents of a message secret, steganography focuses on keeping the existence of a message secret. Steganography and cryptography are both ways to protect information from unwanted parties but neither technology alone is perfect and can be compromised. Once the presence of hidden information is revealed or even suspected, the purpose of steganography is partly defeated. The strength of steganography can thus be amplified by combining it with cryptography. 7

8 1.7 Principles of Embedding In today s society the most practical implementation of steganography is used in the world of computers. Data is the heart of computer communication and over the year a lot of methods have been created to accomplish the goal of using steganography to hide data. The trick is to embed the hidden object into a significantly larger object so the change is undetectable by the human eye. The best object up to this writing is probably a digital image. Digital images have the benefit of containing massive amounts of bytes to designate pixel color for the photo. It is important to understand the process by which digital steganography takes place, and to make sure your cover image is large enough to support the byte manipulation. The basics of embedding data rely on three different facets. These three items are capacity, security, and robustness. Capacity means the amount of data that can be hidden in the cover image. Security is the interceptor s ability to decipher the data hidden inside the cover image. Finally, robustness means the amount of manipulation a cover image can handle before it is obvious a change has taken place. Steganography closely resembles encryption in the fact that it requires the receiver to know the secret which is called the secret key. In order for steganography to remain secure the initial unmodified cover image must also be kept secret. It would be easy to detect hidden data in an image if one was to have the original image side by side with the steganography cover image. 1.8 Applications of steganography 1. Military people can exchange their data in the form of images so that no one can detect what is hidden in the images. 8

9 2. For business users, steganography has two important applications namely enhancing the privacy of personal communication and protecting copyrights on digital materials such as images, software, video and audio. 3. Current steganographic applications with audio media are primarily limited to providing proof of copyright and assurance of content integrity. 4. Watermarking: Although the concept of watermarking is not necessarily steganography, there are several steganographic techniques that are being used to store watermarks in data. The main difference is on intent, while the purpose of steganography is hiding information, watermarking is merely exting the cover source with extra information. The kind of information hidden in objects when using watermarking is usually a signature to signify origin or ownership for the purpose of copyright Protection. 5. Fingerprinting: With fingerprinting on the other hand, different, unique marks are embedded in distinct copies of the carrier object that are supplied to different customers. This enables the intellectual property owner to identify customers who break their licensing agreement by supplying the property to third parties. 9

10 Chapter 2: Data Hiding Techniques for Text This chapter briefly describes how a text message can be embedded in binary images, color images and in audio files. First it discusses about the discrete cosine transform technique of embedding followed by least significant bit embedding in images and then how text can be embedded in audio peaks. 2.1 Binary Images Discrete Cosine Transform (DCT) Steganography 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 domain methods are indepent of the image format and the embedded message may survive conversion between lossy and lossless compression Approach A two-dimensional DCT is performed on small blocks (8 pixels by 8 lines) of each component of the picture to produce blocks of DCT coefficients [1]. The magnitude of each DCT coefficient indicates the contribution of a particular combination of horizontal and vertical spatial frequencies to the original picture block. The coefficient corresponding to zero horizontal and vertical frequency is called the DC coefficient. 10

11 In the formulas, F (u,v) is the two-dimensional NxN DCT. u,v,x,y = 0,1,2,...N-1 x,y are spatial coordinates in the sample domain. u,v are frequency coordinates in the transform domain. C(u), C(v) = 1/(square root (2)) for u, v = 0. C(u), C(v) = 1 otherwise. The two-dimensional DCT is just a one-dimensional DCT applied twice, once in the x direction, and again in the y direction. When you apply the DCT to an input image, it yields a matrix of weighted values corresponding to how much of each basis function is present in the original image. For most images, much of the signal energy lies at low frequencies; these appear in the upper-left corner of the DCT. The lower-right values represent higher frequencies, and are often small - small enough to be neglected with little visible distortion. The DCT is often used in signal and image processing, especially for lossy data compression, because it has a strong "energy compaction" property: most of the signal information ts to be concentrated in a few low-frequency components of the DCT [2]. 11

12 Figure 1. Process of Image Steganography Design Constraints The technique used for hiding data in the image file should be in such a way that there should be no noticeable change in the image file after hiding the data. The technique also relies on what type of image we are selecting as a cover. If the image we selected is a plain one with more of white and black background there are noticeable changes for the stego object. As shown in Figure1 we can observe that the cover image last four bits are changed with the message image. When more bits are changed in the cover image the observer can just analyze that there exists some type of data hidden but he can t retrieve the message until and unless he knows the retrieving algorithm. But if we try to choose an image as a cover with more information in it (i.e., 12

13 without white background) we can successfully embed the message in the cover since changing the bits of the color does not change the whole color. Suppose changing the last bits of the green color results in other type of green but the whole color does not change to other color Security 1. The user should not be able to reveal the data without the key. 2. The user shouldn t even know that the file contains any hidden data Algorithm for Binary Gray Scale Images 1. Read an image and message. 2. Convert the message, which is in text format into binary format. 3. Divide the image into blocks of size 16*16 4.Calculate the DCT for individual block and obtain the 256 DCT coefficients for each block. 5.Replace the last DCT coefficient of each block with the binary formatted hidden message. 6. Find Inverse Discrete Cosine Transform (IDCT). 7. Deliver the image to the destination. 8. Divide the image into blocks of size 16*16 9. Calculate the DCT for individual block and obtain the 256 DCT coefficients for each block. 10. Read the last DCT coefficient of each block which will be a binary formatted message. 11. Convert this binary formatted message in text message. 13

14 2.1.5 Disadvantages We cannot embed more information since the algorithm is only restricted to binary images. As binary images cannot store more information data hiding capacity is low. As we are using transformation technique the data can be embedded only at certain regions of the image. Stego image becomes visible if we try to embed more information Least Significant Bit (LSB) in Bitmap (BMP) Gray Scale Images When embedding a message in a raw image that has not been changed with compression, such as a BMP, there exists a trade-off between the invisibility of the message and the amount of information that can be embedded. A BMP is capable of hiding quite a large message, but the fact that more bits are altered results in a larger possibility that the altered bits can be seen with the human eye [3]. The main disadvantage regarding LSB in BMP images is surely the suspicion that might arise from a very large BMP image being transmitted between parties, since BMP is not widely used anymore. As shown in Figure 2 we can clearly distinguish that some hidden data exists in the cover image but we cannot identify what is that message. Results are sown for two different images. The cover image with letter can be easily identified when compared to the ocean image. LSB in BMP is most suitable for applications where the focus is on the amount of information to be transmitted and not on the secrecy of that information. 2.2 Basic Idea Of Least Significant Bit Color (RGB) Image Steganography The concept of LSB embedding is simple. It exploits the fact that the level of precision in many image formats is far greater than that perceivable by average human vision. Therefore, an altered image with slight variations in its colors will be indistinguishable from the original by a 14

15 human being, just by looking at it. By using the least significant bits of the pixels color data to store the hidden message, the image itself will seem unaltered. Observations Figure 2. Stego Objects for Two Different Cover Images Picking a Good Medium As important as the steganographic technique is, equally important is the choice of the cover image. In LSB embedding, a poor choice of cover image can lead to a stego-image that is easily differentiable from the original. Current image formats can be divided into two broad categories, lossy and lossless. Lossy images are those formats, which loses some of the image s data when stored. An example would be JPEG. The plus side of lossy images, in particular JPEG, is that it achieves extremely high compression, while maintaining fairly good quality. However, due to 15

16 the very nature of lossy formats, it is not suitable for LSB Embedding. Since LSB Embedding spreads the hidden message throughout the image s data, the loss of the image s data by compression would lead to the lost of parts of the hidden message. On the other hand, lossless images are suitable for LSB Embedding, since the integrity of the image data is preserved. However, they do not have the high compression ratio that lossy formats do. Not all lossless images are good candidates as a cover image. 24-bit bitmaps, as well as grayscale images and other color images with small variations in its palette are good candidates as cover images The Simple Case 24-bit Bitmaps Perhaps the simplest implementation of LSB embedding is that using 24-bit bitmaps. The structure of a 24-bit bitmap is a bitmap header, followed by the pixels data. Each pixel is represented by three bytes, representing the red, green and blue color values for that pixel. The higher the number, the more intense that color is for that pixel. For example, if the data for a pixel p a were FF FF FF 16, that pixel would contain the most of all three primary colors and thus be white. LSB uses the fact that changing the LSB of these bytes would produce only a minute insignificant change to the color value (Johnson & Jojodia, 1998). For example, changing the color values for p a to FE FE FE 16 would make the color darker by a factor of 1/256. This change would be imperceptible to the human eye. The idea then is to simply encode one bit of the hidden message in the LSB of each byte of pixel data [4]. Thus, we can embed <number of bytes per pixel> * <number of pixels in image> bits of secret information in any particular cover image. In the implementation however, one should be aware of a particular detail. In 24-bit bitmaps, the number of bytes per row is always -padded with zeros to be a multiple of four. Although initially one may think to use these extra bytes to store hide additional information that would be unwise. Since these bytes are supposed to contain zeros, any alteration would be easily 16

17 detectable. Thus, in order for the image to remain inconspicuous, only the LSBs of the actual pixel data should be altered. If we are using 24-bit color, the amount of change will be minimal and indiscernible to the human eye. As an example, suppose that we have three adjacent pixels (nine bytes) with the following RGB encoding: Now suppose we want to "hide" the following 9 bits of data (the hidden data is usually compressed prior to being hidden): If we overlay these 9 bits over the LSB of the 9 bytes above, we get the following (where bits in bold have been changed): Note that we have successfully hidden 9 bits but at a cost of only changing 4, or roughly 50%, of the LSBs. This is clearly evident from figure3 as we cannot make out which is original image and which one is stego. 17

18 Cover Image Stego Image Figure 3. Color image Stego result Other Variations of LSB Variations of the basic LSB technique have been developed in order to make it more robust. So far, the techniques that have been described are called sequential LSB. That is, the message is laid out across the image data sequentially. One variation would be random LSB, in which the secret data are spread out among the image data in a seemingly random manner. This can be achieved if both the ser and receiver share a secret key. They can use this key to generate pseudorandom numbers, which will identify where, and in what order the hidden message is laid out. The advantage of this method is that it incorporates some cryptography in that scattering in different ways is applied to the secret message. However, it goes beyond just making it difficult for an attacker knows that there is a secret message to figure out the message. It also makes it harder to determine that there was a secret message in the first place. The reason is because the randomness makes the embedded message seem more like noise statistically than in the sequential method. 18

19 Another variation to regular LSB is to repeat the message multiple times across the image data. This way, if the message is relatively small compared to the cover image, the message may survive any image manipulation, such as cropping. However in this case, a redundant pattern may be easier to discern, and will seem less like random noise. Thus the image may be more susceptible to statistical steganalysis. 2.3 Digital Audio Steganography Digital audio differs from traditional analog sound in that it is a discrete rather than continuous signal. A discrete signal is created by sampling a continuous analog signal at a specified rate as shown in figure4. Figure 4. Process for Converting Signals from Analog to Digital Digital audio is stored on a computer as a sequence of 0's and 1's. With the right tools, it is possible to change the individual bits that make up a digital audio file. Such precise control allows changes to be made to the binary sequence that are not discernible to the human ear. In a computer-based audio steganography system, secret messages are embedded in digital sound. The secret message is embedded by slightly altering the binary sequence of a sound file. Existing audio steganography software can embed messages in WAV, AU, and even MP3 sound files. Embedding secret messages in digital sound is usually a more difficult process than embedding 19

20 messages in other media, such as digital images. In order to conceal secret messages successfully, a variety of methods for embedding information in digital audio have been introduced. These methods range from rather simple algorithms that insert information in the form of signal noise to more powerful methods that exploit sophisticated signal processing techniques to hide information. But our technique makes use of the peaks of the music file and embeds the information in the peak bytes of the music file [6]. Our human ear cannot distinguish the sounds at the peak points since more sound exists at that peaks. If we sequentially change the bits of the audio there will be more disturbances in the audio file and we can clearly determine that some form of disturbance or noise exists in the music file. Instead if we choose the peak values we cannot determine the noise of the music file. To select the peaks we take some threshold value of the music file and we try to embed the message beyond the threshold value. Threshold value differs from one music file to other music file. And it also deps on the length of the message. If the length of the message is short the number of peaks would be less and thus threshold will be large. If the length of the message is large we require more number of peaks and thus the value of the threshold would be small. The retrieval of the message will purely dep upon the threshold value [7]. If the threshold value is not chosen correctly one cannot retrieve the message completely. The threshold value is chosen between 0 and 1.If the threshold value is chosen less than the original value all the peaks are not selected. Some of the peaks would cut off thereby some of the embedded message in the peaks is missed. Figure 5 and Figure 7 show the audio stego file for audio files Clearasil and Apple jacks. Figure 6 and Figure 8 show the amount of data retrieved in kilobytes for different threshold values. 20

21 Figure 5. Audio Results for Music File Clearasil 21

22 Figure 6. Audio Result For Music File Clearasil 22

23 Figure 7. Audio Results for Music File Apple Jacks 23

24 Figure 8. Audio Results for Music File Apple Jacks 24

25 Chapter 3: Data Hiding Techniques of Image This chapter briefly describes how an image can be converted into audio file and how an image can be hidden in other images. 3.1 LSB Hiding Image in Image In multilevel security systems such as the one used by the army one wants sometimes to declassify some information from, say, top secret to confidential or even public. Unfortunately this is not as easy as it seems, especially if you want to downgrade images. Indeed information could have been covertly hidden in a top secret image for later retrieval when the image is declassified. We are aware of the fact that when we divide the image into eight bit planes most of the information about the image lies in the most significant bit planes. So we concentrate on the least significant bit planes of the cover image. T he message image is also divided into eight bit planes. The most significant bit planes of the message image are replaced with the least significant bit planes of the covered image. Here the replaced significant bit planes (n) of the cover image become the significant factor. As the number of replaced bit planes increases the cover image quality becomes degraded. So we should try to choose the value of n such that stego image doesn t reveal the actual message image. From my observation it is found that at n=4 i.e., replacing the least 4 significant bit planes of cover to hide the message image would be the better one since we cannot make out the secret image. The images from Figure 9 to Figure 15 in the next seven pages describe the steganographic process for different n values. The first two top images in each figure are the cover and hidden images respectively. The bottom two images in each figure are stego and extracted images respectively. 25

26 For n=1.the least significant bit plane of arctic hare is replaced with most significant bit plane of F15 image. Cover image Hidden Image Stego Image Extracted Image Figure 9. Image Steganography for bit plane n=1 26

27 For n=2.the two least significant bit planes of arctic hare are replaced with two most significant bit planes of F15 image. Cover Image Hidden Image Stego Image Extracted Image Figure 10. Image Steganography for bit plane n=2 27

28 For n=3.the three least significant bit planes of arctic hare are replaced with three most significant bit planes of F15 image. Cover Image Hidden Image Stego Image Extracted Image Figure 11. Image Steganography for bit plane n=3 28

29 For n=4.the four least significant bit planes of arctic hare are replaced with four most significant bit planes of F15 image. Cover Image Hidden Image Stego Image Extracted Image Figure 12. Image Steganography for bit plane n=4 29

30 For n=5. The five least significant bit planes of arctic hare are replaced with five most significant bit planes of F15 image. Cover Image Hidden Image Stego Image Extracted Image Figure 13. Image Steganography for bit plane n=5 30

31 For n=6. The six least significant bit planes of arctic hare are replaced with six most significant bit planes of F15 image. Cover Image Hidden Image Stego Image Extracted Image Figure 14. Image Steganography for bit plane n=6 31

32 For n=7. The seven least significant bit planes of arctic hare are replaced with seven most significant bit planes of F15 image. Cover Image Hidden Image Stego Image Extracted Image Figure 15. Image Steganography for bit plane n=7 32

33 3.1.1 The Ups and Downs of LSB As presented, LSB embedding has the advantage that it is simple to implement. This is especially true in the 24-bit bitmap case. It also allows for a relatively high payload, carrying one bit of the secret message per byte of pixel data. In addition, it is also seemingly undetectable by the average human if done right. However, the assumption has been that the stego image is indistinguishable from the original cover image by the human eye. There have been many statistical techniques developed to determine if an image has been subjected to LSB embedding. In addition to being vulnerable to detection techniques, LSB is extremely vulnerable to corruption. That is, the integrity of the hidden message can easily be destroyed. All the attacker must do is to randomize the LSBs of the image. The attacker may not even know that it is a stego-image, but such actions would destroy the secret message. 3.2 Conversion of image to.wav file Here we use the wavwrite function in matlab to convert the image into sound file. In matlab stereo data should be specified as a matrix with two columns. The amplitude values of the sound file should be in the range of -1 to 1. So our image values through some manipulations should be arranged in a matrix of two columns with all of the values ranging in [-1 to 1]. First of all we need to read the image file and should observe all the values of the image. We need to determine the size of the image. Image pixel values may range in any form of decimals from 0 to 255 since image pixel values are fixed in that range. Image pixel values are separated. If one of the pixel value is [567] it separates the value as [5 6 7].All the values are divided by 10. All the values are subtracted with 127.5(half of 255) and divided with By doing this all the pixel values are ranged from -1 to 1.All the values may not be in two columns. So we use the reshape predefined function to arrange all these values in a matrix of two columns with amplitude values -1 to 33

34 1.Figure 16 describes conversion to audio files for different images. Note that the image size is not equal to converted audio size since we added the size of the image (in terms of 1 s and 0 s) to the image values. The reverse process is implemented for converting sound file to image. The length of the audio file deps on the size of the image. The image size and audio size may not be equal since in the conversion some junk values are introduced. The time taken for converting an image into sound (wav) file deps on the image size and the data it acquired. When compared to conversion of image to wav file, conversion of wav file to image file takes longer. As shown in Figure 17 we can clearly examine the time taken for converting audio to image files. As the size of the image grows data retrieval from audio to image takes longer. 34

35 Figure 16. Conversion of image to audio for different image sizes 35

36 Figure 17. Conversion of audio to image for converted audio files 36

37 Chapter 4: Comparison of Image and Audio Steganography In this chapter we briefly compared image and audio steganography and discussed the applications of steganography. 4.1 High data capacity is achieved through audio medium as cover Generally size of the audio files is greater than image files. So the number of bits to be replaced in the audio files will be more than the image files. Binary images cannot hide more data as shown. As binary images they don t have much information in the image, when we try to embed large text file we can clearly make out that the image has some information hidden in it. When it comes to the audio file no such problem exists. It only checks if the peak threshold is enough or not and later it goes with embedding. If the data file to be hidden is an image then it definitely matters. Images cannot embed the whole image. It can only embed certain number of bit planes. If the number of bit planes increases the cover image loses its visibility. 4.2 Data embedding is not easy with cover medium as audio Data embedding with images is very easy when compared to audio. In DCT image steganography we apply the transformation and we concentrate on the lower bits of the image to hide the information whereas in LSB steganography we simply change least significant bits of each data pixel and embed the data. So there is no selection procedure for hiding the bits. In the case of audio we cannot change each data pixel for hiding the data since it results noise in the stego file. So we are concentrating only on the peaks of the audio file. If the number of peaks chosen is less than the data hidden we cannot embed the whole information. So choosing the threshold peak value is difficult in the case of audio. 37

38 4.3 Speed of embedding and retrieving Data embedding and retrieving is fast with cover medium as image instead of audio. Data can be sequentially embedded and retrieved in each least significant of the pixel and it can be retrieved in the same manner. The algorithm checks for each and every least significant bit of the data and it embeds or retrieves the data. But in audio it first checks for the threshold value given by the user and then embeds or retrieves the data. Deping on the size of the data sometimes data embedding and retrieving may take more than 15 minutes. So image steganography is faster when compared to audio steganography. 4.4 Ease of Steganography Data embedding in images is easy when compared to audio. Data is sequentially embedded in images whereas in audio it is embedded above certain threshold value. If the number of peaks is not sufficient to hide the information then a compromise has to be made on the threshold value. But in the case of image the last significant bits are increased if the data to be hidden increases. 4.5 Robustness High robustness is achieved with the cover medium as audio. If we consider a data file to be hidden in audio no one can make out any difference from the original audio. But in the case of image as the data file increases image loses its visibility thereby showing the data hidden. 38

39 Table 1 : Comparison for image and audio steganography Parameters Audio Image Data Capacity Data capacity is high in audio files since the cover capacity is high Data capacity is low in image files since cover capacity is low when compared to audio. Ease of Data embedding is difficult Data embedding is easy since we Embedding since we embed the data in sequentially change the least significant peaks. bits of the data. Speed of Data is slowly embedded and Embedding and retrieving of data is faster Embedding. retrieved at the peaks after the since we sequentially concentrate on least threshold value. significant bits. Ease of Steganography with audio is Steganography with images is easy as we Steganography Robustness difficult since we concentrate on peak values to reduce the noise. High robustness is achieved since noise is less than 0.4%. From Figure 19 we observe that for different sizes of data the noise is not even greater than 0.5% of the audio. apply only transformation techniques and LSB techniques. Robustness is less when compared to audio since we can see the hidden image after crossing 4 bit planes. From Figure 18 we can clearly say that at n=4 the data becomes visible. 39

40 Figure 18. Data Visibility of Image Steganography Leg for Figure 18 : Orange Line-Indicates the data invisibility. Blue Line-Indicates data visibility. 40

41 Figure 19. Noise of Different Text Files From Figure 19 we can clearly observe as the data increases in the text file noise also increases in the audio file. 41

42 Chapter 5: Conclusion and Future work This chapter discusses about the conclusion and future work of the project. 5.1 Conclusion Some people might argue that encryption is superior to steganography. In some ways they are right, but in other ways they are very wrong. Although encryption is a great cryptographic tool to secure your data, there are reasons why steganography is superior. To start with, when you encrypt data, you are telling all third party (would be hacker) readers here is secret data I do not want you to read and I am encrypting it so you cannot read it. This leaves them knowing for sure there is something hidden from their eyes and they now have the challenge to hack into it. When you use steganography tools to secure data, you present the data in a way that looks like text, images, sounds and videos which appears to be unsecured and unimportant so that hackers will look at the data, think there is nothing particularly important about it, and are more than likely to leave it alone altogether. From our project we improved from binary to color images.later we came up with audio steganography in which we can hide more data when compared to color images. As we have seen from our observations one cannot make out that there is some information in the color images and wav files. So we can conclude that we can securely hide the information in images and wav files when compared to cryptographic tools. 5.2 Future Work From what I have read, in the articles is there are people who have argued there is nothing new for steganography. They seem to think everything that can be done with steganography has been done. This is absurd. It seems to me the sky is not even the limit. What other ways are there to hide information using computers? If we can hide text in pictures, is there a way to hide pictures 42

43 in text? Hidden messages are not even limited to pictures. Technology has advanced to where now we can hide messages in MP3 files and other audio formats like wav files. What future technologies are there to be invented? Maybe someday, someone will find a new way to hide messages in websites and videos, beyond hiding things in photos on websites. There are people always finding ways to overcome obstacles. People have been hiding messages through different means for thousands of years and now they have even more ways to do it thanks to the tools available, sometimes for free download, online. Watermarking tools, still image steganography and audio steganography are among the best ways to hide data today and will still be until even better tools are developed and released to the public. Due to time and computing limitations, we could not explore all facets of steganography and detection techniques. As we saw, we studied the power in our pictures and audio to test for hidden data. Another method which we were unable to explore was to analyze how data can be hidden in videos. Today more research work is going on video steganography. If there is a fair development in video steganography then data embedding capacity will be much higher and data detection also becomes very difficult for the intruders. Deping on the techniques we adopt we can make steganography more effective in future. From our observations we found that color image steganography is effective when compared to grayscale image steganography. In the same way from image to audio also we found that in certain parameters audio is best way to encrypt. Finally encryption through steganography when compared to cryptography and other techniques is the best one at present since we cannot predict future. Yet another research extension is that of using wavelet transforms and Hadamard transforms for embedding text in images. 43

44 References [1]: Rafael C. Gonzalez, Richard E. Woods Digital Image Processing Prentice Hall Inc, [2]: Oppenheim, V. Alan, Schafer, W. Ronald, Discrete-Time Signal Processing, Prentice Hall Signal Processing Series, [3] N.F. Johnson and S. Jajodia. Exploring steganography: Seeing the unseen. Computer, 31, no 2: pp 26-34, IEEE Computer, February [4] Sos Agaian, Benjamin Rodriguez, Glenn Dietricht, Steganalysis Using Modified Pixel Comparison and Complexity Measure, SPIE Conference, (2004). [5] Lossless Adaptive Digital Audio Steganography, Agaian, S.S.; Akopian, D.; Caglayan, O.; Dapos;Souza, IEEE pp NOV [6] Sos S. Agaian, David Akopian and Sunil A. D Souza Frequency Domain Based Secure Digital Audio Steganography Algorithms, IEEE SP/CAS, 2005 International Workshop on Spectral Methods and Multirate Signal Processing, SMMS, June pp 20-22, 2005, Riga, Latvia. [7] Bassia, P., Pitas, I., and Nikolaidis, N, Robust Audio Watermarking in the Time Domain, IEEE Trans. Multimedia, vol. 3 (2001) pp

45 APPENDIX A Gray_Embed.m (This file embeds text in binary gray scale images) clc; clear all; close all; I=double(imread('mail.bmp'))/255; [m,n]=size(i); if rem(m,16)~= 0 t1=16-rem(n,16); else t1=0; if rem(n,16)~= 0 t2=16-rem(n,16); else t2=0; m1=m+t1; n1=n+t2; IMG=zeros(m,n); IMG=double(imread('mail.bmp'))/255; IMG(m:m1,n:n1)=zeros; disp '*************WELCOME TO THE PROJECT************'; str=input('type a message you want to hide in the cover image:'); [l1,l2]=size(str); disp 'Length of the Message is:'; disp(l2); 45

46 lemes=length(str); w=lemes.*7; me=char2bin(str); me=me.*0.7; mes=me+((~me).*(-0.7)); I1=zeros(16,16); I2=zeros(16,16); I3=zeros(16,16); x=0; y=0; disp 'Cover image is:'; imshow(img); title('cover Image'); rows=(m1/16); cols=(n1/16); f=1; for x1=1:rows y=0; for y1=1:cols I1(:,:)=IMG(x+1:x+16,y+1:y+16); dcti1=dct2(i1); dcti2(:,:,f)=dcti1; f=f+1; y=y+16; x=x+16; 46

47 dcti2(16,16,1:w)=mes; res=zeros(m1,n1); p=0; q=0; count=0; pro=rows*cols; for z1=1:pro I2(:,:)=dcti2(:,:,z1); idct1=idct2(i2); res(p+1:p+16,q+1:q+16)=idct1; count=count+1; q=q+16; if count==cols p=p+16; q=0; count=0; disp('stego Object is:'); figure,imshow(res); title('stego Object'); diff=img-res; figure; imshow(diff); title('difference image'); disp 'CODE IN MATLAB TO RETRIEVE THE TEXT FROM AN IMAGE'; x=0; 47

48 y=0; f=1; for x1=1:rows y=0; for y1=1:cols I3(:,:)=res(x+1:x+16,y+1:y+16); dcti3=dct2(i3); dcti4(:,:,f)=dcti3; f=f+1; y=y+16; x=x+16; mess=dcti4(16,16,1:w); rmess=floor(mess+(0.7000)); restore=bin2char(rmess); disp(restore); Char2bin.m (This file converts stream of characters to binary format to embed in binary gray scale images) % Algorithm to convert string to binary value function[g]=char2bin(ch) xz=0; len=length(ch); for j=1:len ab=abs(ch(j)); 48

49 i=7; while ab>0 if (rem(ab,2)==0) d(i)=0; else d(i)=1; i=i-1; ab=floor(ab/2); g(xz+1:xz+7)=d(1:7); xz=xz+7; Bin2char.m (This file convert s stream of binary bits to character format at the receiver ) % Algorithm to convert Binary to String value function [character]=bin2char(a) le=length(a); let=le/7; xy=0; for j=1:let n=64; sum=0; for i=xy+1:xy+7 sum=sum+(a(i).*n); 49

50 n=n/2; character(j)=char(sum); xy=xy+7; 50

51 APPENDIX B Embedd.m (This file embeds text file in to color images) %%% Least Significant Bit Plane Based Embedding Concept %%% This funstion takes the location of the image that is used %%% as the cover image and the location of the text file that is need %%% to embedded %function out = Embed(pathimage, pathfile); cover = imread('boat.jpg'); info = file_readd('msg.txt'); k = 1; len = length(info); [row col depth] = size(cover); out = cover; for Z = 1:depth for i = 1:row for j = 1:col temp = double(cover(i,j,z)); if mod(temp,2) == info(1,k) k = k+1; else if temp == 0 temp = temp+1; else temp = temp-1; 51

52 k = k+1; out(i,j,z) = uint8(temp); break; if k > len break; if k > len break; if k > len figure;imshow(out); imwrite(out,'c:\users\jyotula Chakri\Desktop\Stego.bmp'); File_readd.m (This file reads each character from the text file and converts into stream of binary bits) %% This program reads data from a text file and convert it into stream of %% binary bits [1 0] that can be used for hiding information securely into %% binary cover images. function bin_out = file_readd(pathname) %%% Input: this function takes the path location of the text file %%%% opening the file and reading information 52

53 fid = fopen(pathname,'r'); fdata = fread(fid,'*char'); data = fdata'; fclose(fid); %%%% Converntion of data into binary stream based on the its corresponding %%%% ASCII values for each character array = [ ]; len = length(data); len1 = len; temp1 = zeros(1,20); %%%%% ---- key generation of sucessful reconstruction ---- for i = 20:-1:1 if len1 >= 2^(i-1) temp1(1,i) = 1; len1 = len1-2^(i-1); else temp1(1,i) = 0; temp2 = 1; for i = 1:len asc = int8(data(1,i)); con = 8; temp = zeros(1,8); while asc > 0 53

54 if array(1,con) > asc else con = con -1; asc = asc-array(1,con); temp(1,con) = 1; con = con-1; temp2 = [temp2 temp]; out = [temp1 temp2(1,2:length(temp2))]; bin_out = out; File_writee.m (This file reads binary data and converts into corresponding characters and writes into the text file) %% This program reads binaty data [1 0] and converts into corresponding %% ASCII characters and writes the information to a text file that has been %% created function data = file_writee(info); %% This Algorithm Reads any random data and converts into string of binary %% information that could be easily hidden into the signature or any binary %% data. %%%% obtaining the length of the text temp = zeros(1,20); for i = 1:20 temp(1,i) = 2^(i-1); 54

55 len_file = sum(info(1,1:20).*temp ); %%%% retriving the characters from the binary bit stream len = length(info); len1 = len-20; if len1 > (len_file*8) temp1 = info(1,21:20+(len_file*8)); else len2 = floor(len1/8); temp1 = info(1,21:20+(len2*8)); %%%% ******** %%%%% array = [ ]; char_temp = 'a'; for i = 1:(length(temp1)/8) ai = ((i-1)*8)+1; bi = i*8; temp2 = temp1(1,ai:bi); temp3 = sum(temp2.*array); char_temp = [char_temp char(temp3)]; data = char_temp(1,2:length(char_temp)); fid = fopen('c:\users\jyotula Chakri\Desktop\stego.txt', 'w'); fwrite(fid, data, 'char'); fclose(fid); 55