A holographic image robust watermarking algorithm based on DWT-SIFT and neural network model

2019 International Conference on Advanced Electronic Materials, Computers and Materials Engineering

无线电电子学;计算机科学;材料科学

Pang, J.P.^1 ; Wang, A.L.^1 ; Zhu, X.F.^1 ; Guo, L.^1 ; Li, S.K.^1 ; Xin, K.^1 ; Fan, S.X.^1 ; Liu, F.P.^1

Beijing Institute of Graphic Communication, China^1

关键词: Digital watermarking technologies;    Fresnel diffraction;    Holographic images;    Neural network model;    Position coordinates;    Robust watermarking;    Scale invariant feature transforms;    Watermarking algorithms;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/563/5/052088/pdf
DOI  :  10.1088/1757-899X/563/5/052088

来源: IOP

【 摘 要 】

In order to improve the robustness of holographic watermarking, we propose a watermarking algorithm combining wavelet transform (DWT) and scale-invariant feature transform (SIFT). And neural network model is applied to digital watermarking technology. Firstly, the watermark image is transformed by Fresnel diffraction to generate hologram, which is divided into 16 × 16 blocks. And wavelet transform is carried out on each block to obtain the low-frequency information. Then, the feature points of the carrier image are extracted by SIFT, and the 16 × 16 blocks around the feature points are selected as the embedding areas to realize the embedding of the watermark.Finally, the random coordinates are taken as the input data of the neural network model, and the embedded position coordinates are taken as the output data of the neural network model. The relationship model between coordinates is constructed to achieve the confidentiality of the embedded position. Experimental results show that the watermarking algorithm has good robustness and visibility under the premise of ensuring image quality.