A Robust Digital Watermarking Scheme using BTC-PF in Wavelet Domain

In this paper, a non-blind watermarking technique has been proposed where a grayscale watermark is embedded. In the embedding process, first the host image is decomposed by wavelet transform and a middle frequency subband is selected where the watermark is embedded. Then, both the watermark and the selected subband are encoded by BTC-PF method. For each block, BTC-PF method returns quantization levels and an index of the selected pattern. To enhance the security level the watermark information is permuted randomly. The quantization levels of the subband are modified with the watermark and subband is reconstructed with the modified quantization levels and indices of the watermark. A watermarked image is produced by applying inverse wavelet transform. The experimental result shows that the proposed watermarking technique is a robust against different type of attacks, like blurring, motion blurred, salt and pepper noise, Gaussian filter, cropping and JPEG compression.