Three levels DWT watermarking technique for color images and watermarks

Abstract

This paper presents a three-levels Discrete Wavelet Transform (DWT) watermarking technique for color images and watermarks. The proposed technique utilizes embedding the RGB watermark information into the Y channel of the YCbCr cover image. A pre-processor is used to convert the color watermark from RGB to a binary sequence. The pre-processed output is, then, inserted several times into the Y channel of the YCbCr cover image by changing the LL3 DWT coefficients. The robustness of the proposed scheme is verified through the ability to resist several classical attacks such as scaling, cropping, JPEG compression, median filtering and low pass filtering. In addition, the proposed scheme does not need the original host image to recover the watermark; hence, the recovery method is blind. This paper also includes a qualitative comparison between the one-level and two-levels DWT.