Digital Watermarking Algorithm Based On Transform Domain

With the rapid development of information technology and network technology, the storage, replication and transmission of digital multimedia has become very convenient. To prevent from occurring of piracy and infringement and to the wide dissemination of multimedia information,digital watermarking technology came into being. In recent years, the digital watermarking technology is the forefront of research, which involves signal and image processing, computer science, chaos, cryptography, data communications and other fields and is an interdisciplinary. In essence, the digital watermark is embedded directly into the digital carrier image and does not affect the value of the original carriers and ensure carriers undetectable. It can conforms whether the carrier be tempered by the hidden of watermark. Therefore, digital watermarking technology has played important role in copyright protection of multimedia information and the integrity of the certification. It has become a hot research field.Firstly, this paper briefly describes the background of the digital watermark, the basic concepts of theory, and the overall development of digital watermarking process and the current problems. We also discuss the characteristics of digital watermarking, classification, typical algorithms and analysis the shortcoming of the discrete cosine transform and wavelet transform and the merits of the ridge transform and the curvelet transform. The key algorithm is the effective integration of the compressed sensing theory and the wavelet transform. We put forward a adaptive watermarking algorithm.Based on the above discussion, the DCT domain based on the traditional algorithm is first selected. The grayscale watermark image is broken down into eight binary images in this algorithm and the eight binary images are imbedded in the cosine coefficients. Experimental results show that the algorithm is certain robustness and transparency. Secondly, according to the traditional wavelet theory, generally wavelet has limited direction. Therefore, in the face of mutations of the edge image, the effect of reconstruction does not play a very good result. The ridge wave and curvelet theory are developed recently. Due to the choice of the base of curvelet function, it has better result in reconstruction by the ridge wave and curvelet theory than wavelet function. Therefore, we use the curvelet transform domain to embed the coefficients to achieve a better robustness and peak signal to noise ratioFinally, we propose a algorithm combined with compressed sensing theory and discrete curvelet theory. The algorithm makes use of compressed sensing algorithm to compress the original watermark. The reconstruction coefficient can be embedded adaptively in the coefficient of the discrete curvelet transform and establish the one mapping. This algorithm is fully considered the human visual system. The original image is blocked. We choose the sub-block with high gray scale and embed the coefficients into the block. Experimental results show that the method of the original watermark is encrypted at the same time and ensure a good invisibility and robustness.