Host: The Institute of Image Electronics Engineers of Japan
Co-host: The Institute of Image Information and Television Engineers
Adding noise in flat region of an image is noticeable for human eye. On the other, slightly changing pixel values in edge region is less noticeable. This characteristic is applicable to embed data in an image. In this study, we propose a method to embed the watermark data in the edge region of an image by using the discriminant analysis method. The discriminant analysis method is the method of image binarization. In the proposed method, an image is divided into blocks of fixed size at first. Secondly, pixels in each block are classified into two classes using the discriminant analysis method. At the final step, watermark information is embedded by changing the between-class variance under the pre-defined rule. A simulation experiment is performed to evaluate the quality of the embedded image and to verify effectiveness of the proposed method.