Digital watermarking using decision tree on color images

Digital watermarking is an art of hiding the ownership information of the digital content. Traditional watermarks that are applied to visible media like images or video, whereas in digital watermarking, the media may be audio, pictures, video, texts or 3D models. In this approach a decision tree based ID3 algorithm is developed for performing watermarking on color digital images. The approach makes use of Discrete Cosine Transform coefficients of the image for identifying the most prominent blocks within the image in all image planes. Selecting only those blocks whose PSNR values are high are used for embedding. Using Arnold Transform and keys in the approach makes the overall system highly secure. The design is tested against various intentional and unintentional attacks and is found to be highly robust against the various type of attacks, also the approach is scalable, since it allows to change the value of DC coefficient values to get the required no of blocks for embedding. The approach satisfies all the challenges of ideal digital watermarking system including perceptibility, security, robustness against various attacks, multiple watermark embedding, file format conversions and statistics of the image.