A robust approach of watermarking in contourlet domain based on probabilistic neural network

作者：Liu, Jia-Xing^*; Wen, Xian-Bin; Yuan, Li-Ming; Xu, Hai-Xia

来源：Multimedia Tools and Applications, 2017, 76(22): 24009-24026.

DOI：10.1007/s11042-016-4178-4

• Probabilistic neural network (PNN)
• Contourlet transform (CT)
• Correlation
• Digital watermarking

A novel algorithm of image watermarking in contourlet transform(CT) has been proposed based on probabilistic neural network(PNN) in this paper. In the proposed process, host image is divided into several blocks, and then each block will be decomposed by CT firstly. Then coefficients of each block will be divided into many small coefficient blocks, and the Pseudo random Noise(PN) sequence is generated for the watermarking and embedded into the coefficient blocks that we selected with certain intensity. Thirdly, the correlation between each of embedded coefficient block and same-sized PN sequence is calculated, and put as the input of the probabilistic neural network system for training. Fastly, after training, the trained system will be robust for watermarking extraction. Results show that this scheme is blind, strongly robust and perceptual invisible, which are two major characteristics for digital watermarking, and does not need the watermark in the extraction process.

• 出版日期2017-11
• 单位天津理工大学