SUMMARYIn recent years, high-quality digital media have spread rapidly. Accompanying the diffusion of digital media which can be simply copied and edited without deterioration of quality, research on digital watermarking techniques is being pursued as a means of preventing illegal copying and protecting copyright. In this paper, we will propose a technique for digital watermarking dealing with color images. The embedding of watermark information by the proposed technique is performed by using interblock correlation of the DCT coefficients of the brightness of a color image, and the strength of the watermark information to be embedded will vary with the visual features of the image. It is strongly embedded at places where there are intense changes of the brightness value and is weakly embedded at places where there are only mild changes of the brightness value. Since no distinction between the original image and the image in which the watermark information is embedded can be made visually, third-party estimation of the position where the embedding is performed is very difficult. Moreover, detection of the embedded watermark information can be performed without referring to the original image and the parameters used in embedding. It has been confirmed by experiments that the watermark information embedded by the proposed technique is robust to JPEG compression and Gaussian noise.
SUMMARYAs digital media which can be copied and edited readily without loss of quality has spread, several electronic watermarking methods used to protect copyrights and prevent illegal copying have been proposed. Electronic watermarking technology which can embed author information secretly in data consists primarily of embedding the watermark information, extracting watermark information, and verifying watermark information. Various methods have been proposed for embedding and extracting watermark information, depending on the scope of use and purpose. On the other hand, methods to verify the detected watermark information are limited, with BER (Bit Error Rate) and correlation between the extracted watermark information and the original watermark information widely used for evaluation. In contrast, in this paper, the authors propose a method to verify electronic watermarks based on probability. In the proposed method, the reliability that the original watermark information is embedded in the detected image is calculated as a probability, and then using this value, whether or not the watermark information is there is determined. The calculated value gives an evaluation value directly, one that includes the number of bits in the original watermark information. As a result, this verification method can be considered valid for use with video, in which watermarks are repeated. The authors use experiments to confirm the validity of their method by applying it to electronic watermarks for still images and video.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.