الفهرس | Only 14 pages are availabe for public view |
Abstract Chaos complex behaviour, based on iterating simple equations, is the occurrence of randomness and unpredictability in completely deterministic manner. Its most important sensitivity to initial conditions can be used to generate an infinite number of chaotic random sequences that can find application in the newly evolving fields of security and identity protection of digital data, digital images, and digital videos through Cryptography, Steganography and Digital Watermarking. As the chaotic random sequence is used for embedding the watermark, embedding non- linearity combined with chaos extreme initial conditions sensitivity ensure that a watermark generated even by a key in the vicinity of the correct one will not be positively detected in a watermarked image or extracting hidden data. This thesis presents several techniques to secure multimedia that depend on using spread spectrum broadband techniques based on tailored matched chaotic maps. Three new image authentication and information hiding schemes based on chaotic encryption are proposed. All schemes are based on the invariant property of the chaotic sequence matching histogram as well as chaotic map synthesis using a confidential starting point. In the first method we present a chaos-based spread spectrum watermarking algorithm for still images. The chaotic map together with the initial condition, used to produce the pseudo-random signal, behave as embedding and detecting key. In the second method the chaotic sequence having the same histogram of the chosen location and a cover image are used to generate the chaotic scrambled cover sequence. The DWT of this sequence is then used to modulate a part, having a confidential location, of the discrete wavelet transform of a host image. Either the sequence or the cover image can be the hidden information. In the third method, only the chaotic sequence histogram features are hidden in the host image 2DWT coefficients. This chaotic sequence is considered as a public key for all system users. Each user can use this public key and his confidential initial condition to generate his own chaotic sequence. Each user can use his own chaotic sequence to authenticate his data and to implement secure communication [6]. The difference between using a chaotic map that matches the host image in generating the embedding sequence and using an unmatched one is also discussed. The thesis also investigates the possibility of generating a random chaotic sequence, having a uniform histogram, using a chaotic sequence matching a given histogram and a chaotic sequence matching the complement of this histogram. To study the relation between these two sequences three methods are used to determine the complement of the histogram . In the third method, to obtain a sequence having uniform histogram, it is found necessary to combine, through simple concatenation, two relatively computed different lengths of first and second chaotic sequences generated using original and complement histograms of the second or first methods. This section presents briefly the contents and organization of the remaining chapters of the thesis Chapter 2 provides an introduction to chaos theory, its properties, history of random generators and its relation to chaos. Chapter 3 briefly discusses watermarking, its requirements, types and communication model. Chapter 4 introduces the map tailoring method. Chapter 5 introduces the random sequence generation technique we developed. Chapter 6 presents the proposed methods for watermarking and their results. Chapter 7 offers a technique to use the last proposed method in chapter 6 for multi=user watermarking. Last Chapter 8 includes our conclusions and future work. |