Digital Crime Investigation and Features of Steganography and Steganalytical Techniques - Coursework Example

London Metropolitan Steganography and Steganalytical Techniques by Liban Abokar ID: 13044708 10 December CC6003 -Digital Crime Investigation Abstract The growth in the usage of information, specifically in the public domain, such as the Internet, has in the past decade been staggering; it has significantly facilitated the sharing of information and the transfer of data. Nevertheless, the use of such open communication has greater susceptibility to security threats resulting in unapproved information access. Encryption has traditionally been used to realize security in information technology communication; but once protected information is decoded, it is no longer protected. Steganography is the science and art of communicating in way that the existence of the communication is hidden. In steganography, important information is firstly hidden in a host data such as text, digital video, audio, or image, and then secretly transmitted to the intended receiver. On the other hand, Steganalysis refers to the art and science of detecting the existence of or where steganography has been used. This paper critically reviews both Steganography and Steganalysis, and analyses the different techniques and approaches used to implement steganography using digital text, video, image or audio; various techniques of Steganalysis are also discussed. This paper also introduces some recent algorithms and shows how they have developed, and how currently used techniques are related to Computer Vision and Image Processing. Key Words: Steganography, Steganalysis, information hiding, digital image, text, audio or video, algorithms, image processing Contents I.Introduction 4 II.Discussion 5 LSB modification/insertion 7 DCTs and FFTs 8 Block Tweaking 8 X2 Analysis 9 III.Conclusion 10 IV.Bibliography 11 V.Appendices 12 I. Introduction Usually, Steganography-the art and science of secret communication-is confused with cryptography; they are related in many aspects including information hiding and communication security, but cryptography and steganography not the same. The main aim of steganography is to hide a secret message in a cover-media, in a way that the presence of the hidden message cannot be easily discerned by those without correct access rights; simply put, steganography refers to hiding one piece of information within another piece of information. As opposed to steganography, the presence of a cryptographic message can easily be discerned through causal observation, since messages are usually scrambled to hide it. Thus, the biggest advantage of steganography is the fact that, secret messages transmitted this way are literally non-observable. The downside of this is that, steganography require the use of a big magnitude of overhead just to hide a very small amount of information. However, in order to make a steganography system more secure and productive, encryption and compression techniques can always be applied. In fact, many current steganography applications utilize these systems to come up with a robust system; an important necessity given that a discovered or discerned steganography system is no longer useful (Bhattacharyya et al. 2011). Steganography applications can include a location feature-subcomponent in a data set, time-stamping, captioning, and tamper-proofing-for demonstrating that the original contents have not been tampered with. Thus, there is need to research and come up with ways of developing algorithms that are capable of detecting the presence of secret messages-steganography. The body of approaches and technique that are used for this purpose, the distinction between cover objects/non-stego, those that do not contain secret messages and those that have secret messages-stego-objects, is what is referred to as Steganalysis. Both steganography and Steganalysis have in the past decade received a lot of attention from all over the world; from both those who are interest in transmitting secure secret information, and those that are interested in detecting secret hidden information (Provos & Honeyman 2003). This paper critically reviews both Steganography and Steganalysis, and analyses the different techniques and approaches used to implement steganography using digital text, video, image or audio; various techniques of Steganalysis are also discussed. This paper also introduces some recent algorithms and shows how they have developed, and how currently used techniques are related to Computer Vision and Image Processing. II. Discussion A. The Main Concepts of Steganalysis and Steganography According to Por & Delina (2008), the general principle behind Information Hiding is the embedding of data, which is the message that is intended to be secretly sent. The embedded or secret data is more often hidden in a safe medium referred to as cover message. Apparently, different kinds of cover messages exist depending on the type of medium used to hide a message. For instance, when text is used to for information hiding, then, the cover message is referred to as cover text; or when an image is used for information hiding, then the cover message is cover image. The process of embedding data results in a stego object that contains the secret message. As had earlier been mentioned, a stego key can be used to control the process of embedding data; and also restrict recovery or detection by other parties without correct access rights; see Appendices-Figure IV-1, for a sample process followed to hide data using an image (Li et al. 2011). Three competing aspects must always be put into consideration when designing techniques for information hiding; these include robustness, capacity, and security. Security relates to the inability of an eavesdropper to detect, identify, and recover secret information; capacity is the amount of information that can be embedded in a cover object; while robustness is the magnitude of changes that a stego-object can overcome before the information it contains can be destroyed. Therefore, it suffices to say that steganography aims at ensuring high capacity, robustness and security. For an attack on steganography to be considered successfully, it must detect the secret data or hidden content, as well removal of copyright signature, such as is the case with watermarking. According to Provos & Honeyman (2003) the information hiding hierarchy is made up of covert channels-the use of secure and secret channels to communicate, steganography-classified as technical and linguistic, Anonymity-techniques used to secretly search the web, and copyright marking; see Appendices-Figure IV-2. B. Applications of Information Hiding Various ways in which Information Hiding can be put into use exists. For instance, information hiding can be utilized in advanced data structure where unplanned information can easily be concealed without interfering with the compatibility with old software. A good example is putting extra information about photos on the photos themselves; such information will be transmitted with the photos without interfering or disturbing old software that is unaware of the existence of such information (Rocha & Goldenstein 2006). Hospitals today also utilize information hiding with medical doctors putting together patient’s examination data, their information and imagery. Patient’s information is usually embedded on the image, when radiology is performed, which results in reduced fraud and misdiagnosis. Other areas in which information hiding is used include in watermarks, document authentication, radar systems, document tracking too, intelligence and in military agencies, electronic money and digital elections, remote sensing, and in general communication. C. Steganography Steganography is the science and art of hiding information so that it cannot easily be detected when it is transmitted. It is classified as either linguistic or technical; it is classified as technical when physical means, such as micro-dots or invisible links are use dot conceal information; while it is considered to be linguistic in the event that only linguistics properties of the cover object, such as changes in letter positions or image pixels, are used. This sections introduces and describes the most common digital images embedding techniques; in this case, digital images are used as cover objects, since, as Li et al. (2011) asserts, they are more related to and are widely used in image processing and computer vision. These techniques can also be extended to other types of digital media as cover object including audio, text, and video files. Algorithms for steganography are in general, reliant on the replacement with a pseudo-random embedded message, some noise component of a digital object. The most common noise component in digital images is the LSBs, the least significant bits. This is because alterations in the value of LSBs are imperceptible by the human eye; thus, LSBs are an important place in which information can be hidden without any significant changes to the cover object that can be perceived by humans. Changing some of the properties may result in damaging or loss of those properties given that originally, LSBs contain some statistical properties. This therefore requires that the message being embedded must mimic the cover bits’ characteristics. This can be achieved by utilizing a selection method in which a large number of cover messages are generated in the same way, and only the one with the secret message in it is chosen. This method is, according to Por & Delina (2008) is a method that is computationally expensive; plus, it only provides for very small embedding. Further, constructive method is yet another possibility that can be used, since it entails the building of a mimic function which is able to simulate the cover bits characteristics. Despite the fact that LSBs methods of embedding data often results in stego objects that the humans cannot perceive; the embeddings can however be destroyed very easily. Further, LSB embedding happens on noise making it very easy to access, modify, and destroy through filtering, further compression, or using a less perfect conversion size or format. LSB modification/insertion LSB insertion/modification is one of the still imagery steganography message embedding techniques; it is imperceptible by humans and thus very difficult to detect. On the downside, it is one of the easiest techniques to destroy. It is, however, the most widely used technique today. It takes advantage of the weakness of HVS, human visual system, which is not able to detect the variation in colour vector’s luminance of the visual spectrum at a higher frequency. In this technique, a picture is represented by a variety of colour pixels, with the individual pixels being represented their optical attributes such as chroma and brightness, etc.; which can all be expressed digitally in forms of 0s and 1s. The binary equivalent of the secret message is distributed among the LSBs of each pixel; for instance, trying to hide the character A, into an 8-bit bitmap colour image, the equivalent binary will take the form of 00100111 1101001 11001000 0010011 11001000 11101001 11001000 00100111 The bit binary equivalent of the letter A is then serially copied from the left hand side to the LSBs equivalent of the stated binary pixels pattern. This technique has only one problem, it is vulnerable to attacks such as image formatting and compression. DCTs and FFTs This is yet another Steganography technique that is apparently very effective when it comes to hiding data in digital images; it refers to the use of DCT, Direct Cosine Transform or the use of FFT, Fast Fourier Transform to embed data in the frequency domain. One of the main JPEG compression components is the DCT algorithm. This message embedding process is useful since anyone looking at the pixel values of an image, would not in any ware be aware of something or anything being different. Block Tweaking This is another technique that makes it possible to embed information or data during the quantization phase. This is done by ensuring that all the necessary coefficients within a block are even. With this technique, a large image has the capability of storing data in a way that makes it difficult to destroy, if compared to the LSB method. It is, however, vulnerable to noise. D. Steganalysis There is need to design powerful techniques that can be able to detect the presence of steganography and avoid or help minimize illicit actions such as the distribution of child pornography pictures on the web. Steganalysis, as Pevny & Fridrich (2006) and Rocha & Goldenstein (2006) puts it, is a variety of techniques that have been or are being devised to detect secret messages contained in a digital media. In fact, it is an allusion to cryptanalysis which is the body of techniques that are made to decrypt cyphers and codes. Three different types of Steganalysis attacks exist including aural attacks, which is the stripping away of the important contents of a digital content, enable the inspection of anomalies visually; for instance, showing the LSBs of an image. The other attack is the structural attack, which is the easy detection of a pattern in a file pattern’s format as a result of inherent changes in the format of a file during the embedding of dat. Statistical attack is yet another Steganalysis attack, proper statistical analysis is employed to determine whether an image has been altered or not. This is because all digital pictures, especially of natural scenes, have very distinct statistical behaviour (Pevny & Fridrich 2008). X2 Analysis X2 analysis as presented by Westfeld & Pfitzmann (1999) is used to detect secret messages. Accordingly, they showed that 2L possible values can be represented by an L-bit colour channel, such that if these values are split into pairs of 2L-1, which is only different in terms of LSBs, then all the possible neighbouring bits patterns for LSBs are considered. Each of the said pairs is referred to as PoV, pair of value in the given sequence. In the event that all the available LSB fields in an image that can be used to embed a secret message are used, the distribution of even and odd PoV values will be the same as 1/0 distribution of the data or information bits. According to Westfeld & Pfitzmann (1999), the X2, chi squared test can be applied of the PoV values to help detect secret messages. This technique is only valuable in detecting first LSBs available pixels hidden sequential messages; this is because it only considers the value of descriptor. Thus, it is easy to defeat this approach by simply maintaining the basic statistical profiles during the embedding process. III. Conclusion This paper has presented a critical overview and analysis of Steganography and Steganalysis techniques. It has shown how both steganography and Steganalysis can be applied in real-life, and has also discussed some of the interesting data/information embedding and detection techniques that are in use today. Further, the paper has also discussed the strengths and the weaknesses if any, of each of the techniques discussed. It has shown that steganography is a powerful, secure and robust technique for data hiding; however, the recent advancements in technology have also affects steganography. Steganography being of a different purpose, which is to ensure hidden information or data is imperceptible by humans, it is also complemented in many current software tools. This paper has also managed to show that the tools for performing steganography are becoming more user friendly despite the fact that they are becoming more sophisticated and are offering, better, secure and robust means for data hiding IV. Bibliography Bhattacharyya, S., Banerjee, I. & Sanyal, G., 2011. A Survey of Steganography and Steganalysis Technique in Image, Text, Audio and Video as Cover Carrier. Journal of Global Research in Computer Science, 2(4), pp.1–16. Li, B. et al., 2011. A Survey on Image Steganography and Steganalysis. Journal of Information Hiding and Multimedia Signal Processing, 2(2), pp.142–172. Pevny, T. & Fridrich, J., 2006. Multiclass blind steganalysis for jpeg images. In Proc. of IST/SPIE Electronic Imaging: Security, Steganography, and Watermarking of Multimedia Contents VIII. pp. 1–13. Pevny, T. & Fridrich, J., 2008. Multiclass detector of current steganographic methods for jpeg format. IEEE Trans. Information Forensics and Security, 3(4), pp.635–650. Por, L.Y. & Delina, B., 2008. Information Hiding: A New Approach in Text Steganography. In 7th WSEAS International Conference on Applied Computer & Applied Computational Science. pp. 689–695. Provos, N. & Honeyman, P., 2003. Hide and seek: an introduction to steganography. IEEE Security & Privacy Magazine, pp.32–44. Rocha, A. & Goldenstein, S., 2006. Progressive Randomization for Steganalysis. In 8th IEEE Intl. Conf. on Multimedia and Signal Processing. Westfeld, A. & Pfitzmann, A., 1999. Attacks on steganographic systems. In In Pro-ceedings of the Third Intl. Workshop on Information Hiding. London, UK: Springer Verlag, pp. 61–76. V. Appendices Figure IV‑1: An example of data hiding using an image cover (Source: Li et al. (2011)) Figure IV‑2: Information Hiding hierarchy. Source: (Westfeld & Pfitzmann 1999) Read More