Robust Color Image Watermarking Technique based on DWT and ICA

In this paper, a robust, imperceptible and blind color image watermarking technique based on discrete wavelet transform (DWT) is proposed. The proposed embedding technique is based on computation of noise visibility function (NVF), where the strength of watermarking is controlled which results in watermarks at texture and edge areas are stronger than flat areas. For extraction, conventional techniques need embedding parameters such as strength, location and information about watermark or original image. In this work, an intelligent detection technique, namely, independent component analysis (ICA) is implemented for watermark extraction without the use of original image. The features of FastICA are quick convergence, easy to implement and suitable for watermark applications. The quality of watermarked image and extracted watermark are measured in terms of peak signal to noise ratio (PSNR) and similarity measure (SM) values respectively. Robustness of the proposed scheme is validated against various image processing attacks.