DWT-based Blind Digital Watermarking Methods

Digital watermarking appends some information (e.g. watermark) into the digital information (for example image) to be protected directly to realize the copyright protection and information security, which was invented in the initial stages of the last decade of 20th century with the rapid development of Internet and multimedia and soon brought a research upsurge. Digital watermarking can be widely used in many ways such as copyright protection, fingerprint adding, caption and remark, tamper prevention, access control, communication security and so on.Wavelet analysis is a new kind of time-frequency technology, it developed on the basis of Fourier Transform. It's well space-frequency localization characteristic makes it been widely used in image processing. In first ages digital watermarking in transform domain mainly based on Discrete Fourier Transform and Discrete Cosine Transform, now Discrete Wavelet Transform becomes primary too.Blind digital watermarking is method needs not the original image to detect the watermark. Non-blind methods are usually more secure and robust than blind ones, but it is not insure that the original image can be offered, and which is more important that how to prove the truth of the "original image". So the idea that original image should not be used directly in the watermark detecting is bring forwarded. The blind digital watermarking is more practical in many applications such as ownership validation and so on, it is one of the important evolutional ways of digital watermark.On the basis of many existent methods, in this thesis the techniques of blind digital watermarking in discrete wavelet transform domain is studied, and two schemes of watermarking technique are presented. Thecontribution of this thesis is following:Using correlativity in the statistical theory, in this thesis a scheme of DWT blind digital watermarking dealing with one-dimensional watermark is presented. Then the performance under kinds of image manipulation is test, performance under similar DCT scheme is contrasted. The test proves that this scheme is better than the similar one in DCT domain, robust in many aspects and very practical. To raise the security of this scheme, the watermark is modulated with pseud random sequence.Using the characteristic of human visual system, combined with the LSB ( least significant bits) scheme, in this thesis a scheme of semi-fragile DWT blind digital watermarking based on quantization of the wavelet coefficients dealing with two-dimensional image watermark is presented. Restriction of the size of watermark can be hold is discussed. This scheme is adapt in image authentication. The watermark is compressed before embedded to raise the size of watermark that can be hold; Pseud random sequence is used to modulate the watermark to raise the security of this scheme; The codes are repeated to resist external disturbance and raise the scheme's robustness.