Abstract
With the rapid growth of Internet, digital data is seen everywhere. Several characteristics such as easy-to-modify and easy-to-duplicate make digital data so popular. There are, however, some disadvantages for digital data. For example, authors of digital data are not willing to put their stuff on the Internet since they worry about their products may be stolen and published by other people. This idea can hurt indirectly and severely the development of our Internet community.Recently, digital watermark technique has been proposed to solve the questions about how to authenticate the digital data. The key is adding some information insensible to our eyes into the data and then verifies this information. The embedded information could be the author of product or the copyright. Watermark techniques can also be used in the authentication of still images, video and digital musicPrevious works have been used to look for the position where to place watermark by DCT and wavelet coefficients. This paper uses wavelet packet to calculate the position of the significant wavelet coefficients more precisely and therefore watermark the images much better.First, we decompose the images into several subbands by wavelet packet transform. Second, We find the significant coefficients in each subband and permute watermark randomly in order to achieve the encryption. Also, we use an error correction coding scheme to protect the watermark. Based on the experimental results, the proposed scheme is robust at the attacks of cropping, noise corruption, DCT-based (JPEG) and Wavelet-based (EZW) compression.