Research On Blind Digital Watermarking Algorithm In Contourlet Domain

With the development of technology such as electronics,computer and network,digital media information such as images,audio and video have become mainstream.Their security and copyright protection are seriously threatened due to their ease of reproduction.Digital watermarking technology is considered as one of the most effective methods to solve this problem.It has become a hot topic for many scholars.Based on the current status of digital watermarking technology,the paper summarizes the classification,characteristics,basic framework and evaluation criteria of digital watermarking.Based on the analysis of the principle and characteristics of Contourlet transform,two blind digital watermarking algorithms in Contourlet domain are proposed in combination with QR and SVD decomposition.All of them have achieved good results.The main work of the paper is as follows:1.A blind digital watermarking algorithm in Contourlet domain based on QR decomposition is proposed.This algorithm combines the Contourlet transform and QR decomposition.The original image is decomposed using Contourlet transform and block QR is peformed on low-frequency subband.The binary watermark image is quantized embedded in the selected three elements of the R matrix after Arnold scrambling and dimensionality preprocessing.Before the watermark is extracted,the Hough transform is used to perform the rotation correction.Afterwards,the watermark is extracted from the water-saturated image according to the majority principle and the inverse process of the watermark.2.A blind digital watermarking algorithm in Contourlet domain based on SVD decomposition is proposed.The original image is decomposed using Contourlet transform and block SVD is peformed on low-frequency subband.According to the adaptive quantization step,the binary watermark image after preprocessing and scrambling is embedded in the selected element of the S matrix.Before the watermark is extracted,the Hough transform is used to perform the rotation correction,and then the watermark is extracted from the watermark image according to the embedded inverse process.3.Two blind digital watermarking algorithms are tested for invisibility and anti-attack experiments,and compared with similar algorithms.Experiments show that both algorithms have good robustness against the attacks of JPEG compression,filtering,noise,scaling,rotation and image cropping.The blind digital watermarking algorithm in Contourlet domain based on QR decomposition has outstanding effects in againsting the attacks of pepper and salt noise,Gaussian noise and rotation;The blind watermarking algorithm based on SVD decomposition in Contourlet domain has more advantages in againsting the attacks of JPEG compression,Gaussian noise,rotation and image cropping.