A Digital Watermarking Method Based On Discrete Wavelet Transform And Its Application On Biometric Authentication

Digital watermarking technology for authentication is an important research branch of digital watermarking due to its significant effectiveness in the content verification of multimedia. As a research of a new area, Digital watermarking technologies have reached booming development in the past few decades. Digital watermarking technologies for Content authentication provide new ideas and methods in order to maintain and verify the integrity of multimedia data. After nearly ten years of research and exploration, the researchers put forward a number of different digital watermarking algorithms for media digital authentication. As an important signal processing tool, Wavelet transform is widely used in digital media information processing. It also has been extensively applied for digital watermarking technology.At first, this paper summaries the definition and application of digital watermarking and gives a general introduction of watermarking for authentication, including the definition, the main characteristics, the attacks threatening to it and the main algorithms designed at present. Wavelet theory and its application in digital image processing and digital watermarking are also introduced in detail. Based on the above, this paper proposes a robust digital watermarking scheme based on wavelet transform. In addition, we will combinative the watermarking scheme with the palmprint identification for achieving the dual certification. In the last, the digital watermark technology on the future direction and development prospects will be analyzed and summarized.This paper proposes a watermarking technique based on image features. Four subimages are obtained from the host image by using subsampling. Wavelet transform is applied to the subimages for texture analysis. An optimal Coefficient selector will be formed according to the results of the texture analysis. It will be used to select the location for embedding. Modify the wavelet coefficients according to the coefficient selector and the watermark. In this way, we will embed the user's ID and the related information as a watermark to the palm prints for identifying. In recognition stage, we extract the user's ID and the related information and analysis the extracted information. If the information is incomplete, we think the palmprint image has been attacked and another palmprint image will be needed for identifying. If the information is complete, the palmprint image will be used for the coarse level matching. We will judge whether the palmprint is in the library in accordance with the user's ID and the related information. If the user is in the library, we will extract its texture feature for the fine level matching. If the user in not in the library, we will say the user is illegal. Experimental results show that the watermark algorithm can resist a variety of digital watermarking conventional attacks, and the combination of the watermark algorithm and biological certification can provide the content integrity of the biometric data. It can improve the speed and accuracy of the identification. Because the algorithm combines the characteristics of the original image, the system can actually maintain the quality of the original image. Because of the advantages of wavelet transform, the algorithm increases the capacity of resisting to malicious attacks. An optimal coefficient selector will be formed by comparing the wavelet middle-frequent coefficients of the subimages with the watermark. Watermark extracting is simply by the optimal coefficient selector and does not need original image. Extraction will be truly blind. The combination of the watermarking scheme and the palmprint identification will improve the recognition speed and the accuracy of recognition. It is a new area of research of digital watermarking. It has high value for the practical application.