关键词: 人工神经网络, 水印, 猫映射, 鲁棒性

Abstract: The paper, by adopting jointly Artificial Neural Network （ANN）, general Amold mapping, and statistical methods, intends to formulate an algorithm based on image spatial domain watermark. While the ANN was used as the texture analyzer of the original image, the texture zone of the original image would be given prominently. The safety of the watermark was improved by using the amold mapping rearrangement. An optimizing strategy of minimizing the changes of pixels was adopted when the watermark is embedded. The application of statistical method guaranteed the sightlessness and the robustness of the digital watermark information. The experimental results demonstrate that the watermark is robust against cut attack, salt-pepper noise attack, Least Significant Bit (LSB) attack, winner attack, and so on.

Key words: Artificial Neural Network （ANN）, watermarking, cat mapping, robustness