Research On Anti-geometric Distortion Holographic Digital Watermarking Algorithm Based On Image Segmentation

With the rapid development of computer network applications,the diversity and wide application of network image transmission are also on the rise.The disputes over piracy and copyright have become more and more serious social problems.The application of digital watermarking technology has become an effective method for protecting intellectual property rights and judging the authenticity of digital information.In recent years,when people apply watermarking technology,the standards for watermark robustness and anti-attack potential are getting higher and higher.Watermarking algorithms that can achieve anti-geometric distortion have appeared steadily,but these watermarking algorithms seldom can resist distortion attacks.The anti-distortion digital watermark is still in its infancy.In this case,in order to make a balance between the robustness of the watermark and the imperceptibility of the watermark,this paper proposes a combination of image segmentation and holographic watermarking to realize the hiding algorithm of the watermark information.The carrier image is divided into different regions by image segmentation processing,and the appropriate embedding region is screened out according to the size of the watermark.Then,according to the different embedding regions selected,this paper proposes two different embedding algorithms for holographic watermarks: one method is to select the largest block in the segmentation block to embed the holographic watermark in the image as a whole after the carrier image is segmented by the algorithm;the other method is to select multiple segmentation blocks with clear blocks and relatively minimal interference as the carrier image The embedded block.The original watermark image is converted into a holographic watermark using the four-phase shift algorithm,and the holographic watermark is embedded in the selected carrier image block,so that the embedding of the watermark is completed.Aiming at the problem of geometric distortion of the watermark carrier image,the paper also proposes an image distortion correction method.The image feature point matching method is used to achieve the geometric correction of image distortion.First,the feature points of the watermark carrier image that are attacked by the distortion are extracted,and the embedded watermark is found.The coordinate transformation relationship is established by matching the feature points of the undistorted original image and the distorted image.This relationship is used to perform geometric correction of the embedded block and restore the synchronization of the watermark.The position of the feature point in the corrected embedding area is used to determine the embedding starting point and extract the holographic watermark.The original watermark image is restored by the inverse transformation of the holographic calculation,and the extraction of the watermark image is completed.Compared with the traditional watermark embedding and extraction method,the watermark embedding method based on image segmentation is divided according to the inner edge of the image,so the embedding area selection is more flexible,more concealed,and the watermark is more secure.Therefore,the difficulty of deciphering and directed destruction is increased and it is not easy to be deciphered.Compared with the feature point matching and correction algorithm for the whole image,this algorithm only performs image correction for the embedded watermark area,which can realize the non-uniform distortion correction of the image.It can also shorten the time for feature point matching and image correction,and the calculation efficiency of the algorithm is significantly improved.