Ownership Identification using Discrete Wavelet Transform (DWT)-based LOGO Watermarking

Abstract

In this paper, an efficiently DWT-based watermarking technique is proposed to embed gray-scale logos as watermark in images to attest the owner identification and discourage the unauthorized copying. The method transforms both the host image and watermark into the discrete wavelet domain where their coefficients are fused adaptively based on Human Visual System (HVS) model to hide a higher energy hidden watermark in salient image components. The method repeatedly merges the watermark coefficients at the various resolution levels of the host image which provides simultaneous spatial localization and frequency spread of the watermark to provide robustness against different attacks. From the simulation results it can be observed that proposed method is robust to wide variety of attacks, such as image compression, linear or non-linear filtering, noise addition, image resizing, cropping, and image enhancement, etc.