Semi-fragilewatermarking algorithm based on texture segmentation

Based on texture visual features, region segmentation operator and quantization step equations were put forward. This guarantees the transparency of carrier image and the robustness of watermarking image extracted. After segmentation, the texture contour was clear, and the segmented results of smooth and texture region were satisfying. And it brought up the pixel value adjustment operator of IDWT. The basic idea of the algorithm is that after discrete wavelet transform, divided the low frequency coefficients LL into 2×2 blocks, then defined block coefficient suni Σ. If the value of Σ was greater than the threshold, the block was segmented into texture area, or segmented into smooth area. When quantifying the step, the intensity and texture coefficient were considered, which made the transparency and robustness optimal. Experimental results showed that the result of texture segmentation was obviously better than the present algorithms. The pixel value could be adjusted by coefficient adjustment operator exactly. The carrier image had not only good transparency, but also better anti-attack capability.