Digital Image Watermarking Improvement Using Multiple Duplicated Signals Embedding and Adaptive Retrieval Algorithm Selection Techniques

In this paper, we propose two techniques toimprove the watermark extraction performance and accuracy.The First one is implemented in the embedding process by usinga multiple duplicated watermark signals embedding method.The second technique is archived in the retrieval process byusing adaptive retrieval algorithm selection to extract thewatermark signal. The accuracy of the extracted watermarkcan be improved by applying these two concepts together. Setsof experiment are fulfilled to testify our proposed technique.The experimental results show the improvement in terms ofNormal Correlation (NC) value, compared to the previousmethods. The obtained results also perform better in terms ofthe robustness against various common imaging processingbased attacks. Especially in cropping attacks, our proposedwatermarking methods show significantly improvement over0.2 on average in terms of NC compared to the previousproposed methods