Intelligent and Effective Digital Watermarking Scheme for Mobile Content Service

详细信息    查看全文

• 作者：Hang-Rae Kim ; Young Park ; Mi-Hee Yoon ; Yoon-Ho Kim
• 刊名：Lecture Notes in Computer Science
• 出版年：2005
• 出版时间：2005
• 年：2005
• 卷：3681
• 期：1
• 页码：p.492
• 全文大小：404 KB
• 刊物类别：Computer Science
• 刊物主题：Artificial Intelligence and Robotics
  Computer Communication Networks
  Software Engineering
  Data Encryption
  Database Management
  Computation by Abstract Devices
  Algorithm Analysis and Problem Complexity
  
• 出版者：Springer Berlin / Heidelberg
• ISSN：1611-3349

文摘

In this paper, digital watermarking scheme that is intelligent and effective in protecting a copyright of image data under image transformation such as scaling, rotation, and JPEG lossy compression is proposed. Especially, the proposed scheme is a robust algorithm for mobile content service. It means that the proposed watermarking is robustness to protect impulse noise attacks. The watermark and inserted watermark are very different from each other. The original watermark is recomposed using the code granted to a copyrighter or customer. The proposed algorithm is an extension of spread-spectrum watermarking algorithm that obtains the convolution value of PN sequence. Correlator and decision circuit in our algorithm are required. Because just a little watermark is embedded in the original image, this decision circuit plays an important role for detecting this watermark and then the intelligent method for the decision must be required. Degradation of image quality between original image and watermarked image in order to confirm required invisibility in the digital watermarking system and watermark robustness to protect various attacks from the outside are analyzed using the value of PSNR and recovered rate of watermark image. The experimental result shows that image quality is less degraded as the PSNR of 93.75 dB. It is also observed that excellent watermark detection is achieved under an image transformation and JPEG lossy compression, and watermark can be effectively detected in case of the image inserted impulse noise, which PSNR is 7.97 dB.