DWT-Based Watermarking Using 2DLDA for Color Images

Abstract

In this paper, we propose a watermarking algorithm based on Discrete Wavelet Transform (DWT) by using Two-dimensional Linear Discriminant Analysis (2DLDA) for color images. At first, the color image is converted into the YIQ color space and then transformed into the frequency domain by DWT. During the embedding stage, two watermarks of reference and logo are embedded into the quadrature chrominance component. Then, watermark extraction is done by 2DLDA from the quadrature chrominance component based on DWT. By embedding the reference watermark and using 2DLDA, the training covariance matrix can be directly computed from 2D image samples and the robustness of watermark extraction can be ensured. Also, the computation complexity can be reduced. By considering the Human Visual System (HVS), experimental results have shown that the watermark can be correctly extracted and better robustness is provided after various image attacks.