Abstract
The curvelet is more suitable for image processing than the wavelet, able to represent smooth and edge parts of image with sparsity. In addition, the representation contains more directional information. The primary experimental results show its potential in image processing. A watermark algorithm based on curvelet transform and Arnold transform is represented. Firstly the digital watermarking is scrambled by the Arnold transformation. Then the chaotic watermarking is embedded to the coarse coefficients in the curvelet transform domain. In our method, the watermark is embedded to the coarse coefficients larger than some threshold values. We have developed an optimization technique using the genetic algorithm to search for optimal threshold values and strength of the watermark to improve the quality of watermarked image and robustness of the watermark. Finally, certain algorithm can detect the watermarking. Because this algorithm choose the appropriate position to insert the watermarking, the experiments indicated that this algorithm enabled the watermarking to have the very good invisibility and made the watermarking have very strong robustness to the general image processing like noise, filter, rotation, compression and so on.