Analysis and implementation of dual image watermarking cyber watermarking in cyber security

Abstract

Now days so many people have regular web access and use application like email, instant messaging, and social networking, online messaging, etc. To speed up this communication, there is need of digital media like Text, Image, Audio, and Video. These media transfers over an open public network, hence there is need to protect these data. From all these digital media,,Digital image is most widely transferring files over the network than any other but in most unsecure manner. So that some system must be there making them secure from threats and different kind of attacks over the network. This research work develops an image watermarking technique with hybridization of spatial and frequency domain image watermarking techniques. Here, this work is divided in two parts: watermark embedding and watermark extracting. There is hybridization of three techniques: DWT, DCT and SVD used in both parts watermark embedding and extracting. DWT (Discrete Wavelet Transform) is wavelet based transform that gives information about image in both frequency and spatial domain. DWT gives higher robustness, but the quality of the image is degrading. DCT (Discrete Cosine Transform) is orthogonal transform that transform image spatial domain to frequency domain. DCT gives better quality of image with less processing power. In this work an improved SVD-ARNOLD-LSVR algorithm; the host image of size 512 x 512 is embedded with the watermarks of size 128 x 128. To prove the reliability of the algorithm several images are analyzed and the samples alone taken into account. This algorithm enhances the visual impreceivability of the watermarked image and robustness against various removals and geometric attack newline