Steganography -- Information Hiding In ImagesOlena Vikariy and Andrew CavanaughCMPT 495-01Computer and Data SecurityMontclair State UniversityDecember 13, 2004The word Steganography means “covered writing” and comes from the Greek words steganos, which means covered or secret and graphy, which means writing or drawing. Steganography is the “cousin” of cryptography and is the art of covert communications. Its roots go all the way back to the ancient Greeks when the fellow named Histiaeus, a prisoner of a rival king, needed to pass a secret message to his army. To accomplish that he shaved the head of the trusted slave, tattooed the message on it andsent him to deliver the message in person when his hair grew back(Cole). Other early forms of Steganography include writing on wood under the wax on the wax tablet, using invisible ink made out of juice or milk which shows when heated and using the method called “microdots” which reduces a text to a size of a single dot and inserts in a documentas punctuation. Today, steganography is usually referred to as hiding data in digital media(Cole). For example, a text document can be hidden in an image, sound or another text file. This paper concentrates specifically on data hiding in images. First, it presents some terminology and background information about images and some of their formats. Next, different methods for Steganography will be discussed which are publicly availableand widely used today. Following methods, specific tools that implement those methods will be described along with the results of tests performed on them to verify their quality. Concluding section will sum up what was discussed and give a look ahead to where the field is headed.There are some terms that will be used throughout the paper and here they are defined for easier understanding of the subject. The image in which the data is going to be hidden is called the cover image(Johnson, Jajodia), also called carrier image, overt or host file(Cole). After the data, which is referred to as covert data(Cole), has been hidden in an image, that image is called stego-image(Johnson, Jajodia), also sometimes called stego-carrier or stego-medium(Johnson, Jajodia). Stego-key(Johnson, Jajodia) can be used to make the stego-image more secure; it is the password that is stored in the carrier file along with the covert file without which data cannot be extracted from the stego-image. Looking for possible information in images and decoding it is called steganalysis.To understand how it is possible to hide files in other images it is important to know what digital images are comprised of. As other types of digital media, digital images are nothing but an array of numbers, representing light intensities at various picture elements or pixels(Cole). Each pixel is represented by some number of bits, now usually 8 or 24. With 8-bits, there is a total of 256 colors available. Every image using this representation has a color palette saved with the image which contains every color available to the image(Kessler). The common digital image format using 8-bit colors is Graphics Interchange Format (GIF). With 24-bits, color is derived from the percentage combination of the three primary colors – red, green and blue, 8 bits for each for a total of265*256*256 different colors. They can be represented as hex, decimal or binary. For example, the color white can be represented as such: 11111111, 11111111, 11111111 – 100% for red, 100% for blue and 100% for green.(Johnson, Jajodia) The common formats that use it are BMP(Bitmap) and TIFF(Tagged Image File Format). There also are grayscale images that use 256 shades of gray and pure black and white which use only 2 colors where each pixel can be represented with 1-bit. The size of the image file depends on the number of bits used to represent the colors and the number of pixels in theimage. The higher the both numbers, the higher the resolution of the image and the more space is needed to store such a file. For example, a 24-bit image of 1,024 by 768 pixels requires more than 2 MB of space(Johnson, Jajodia). This is still a lot to transmit over the internet, thus different compression methods have been invented. There are two categories of compression: lossless and lossy. Lossless formats use mathematical algorithms to manipulate data and compress images such that the original image can be recovered exactly when decompressing. It reduces the storage space needed by about 50%, which might still not be enough. Its mostly used to compress text and other data files where every piece of information is important and image file formats GIF and BMP. Lossy compression, on the other hand can reduce storage required by 90% without noticeable to human eye changes and down by 99% with noticeable changes when size ofthe file matters more than quality. The most popular format that uses it is JPEG(Joint Photographics Experts Group), which cuts out the non-significant data from images. The original file cannot be recovered exactly when using lossy compression, that’s why it is best to use lossless when embedding files into images so that if the stego-image is recompressed the embedded data will not be lost. Next, there are many different methods for hiding information in digital files. They are categorized in three groups: insertion-based, substitution-based and generation-based. Insertion-based techniques hide information in fields of the file which are not used by the program. It is used mostly to hide data in a text document with EOF(end-of-file) markers. Most text editors ignore any information after that marker, so anything of any size can be inserted there, however, those insertions do change the file size and if the editor only displays 2 lines of text and the size of the file is big, then it might raise suspicion. Another technique is substitution-based and involves substituting covert data in place of insignificant data in the overt file. The resulting stego-carrier is the same size as the overt file, but the space for putting covert data is limited. Last one is generation-based. This technique does not need the overt file, it only needs the secret file. The covert file is used to determine the angle, length, and color of each line that will comprisethe newly generated image by using mathematical-based manipulation. Such image will most likely not be of a known object(for example a house or a