A New Color Image Watermarking Technique Using Hybrid Domain

Many watermarking techniques have been used for copyright protection of digital images. A color image watermarking technique which is robust and resilient to many types of attack is proposed. This technique involves three stages, viz: preprocessing, embedding and extraction. RGB image is converted to YCbCr to ensure the best imperceptibility and robustness. In spatial domain, the best quadrant of host image is selected by applying canny edge detection to ensure that the selected quadrant contains the highest number of edges. In frequency domain, the low frequency sub-band which is exploited to embed the watermark image is selected by applying DWT. The watermarked image was exposed to six types of attacks. It is encouraging to observe that, using standard quality measurements, the technique proposed performs better than those suggested by Kong & Peng [10] and Al-Asmari [1].