Watermark embedding method based on Fourier descriptor, watermark extraction method based on Fourier descriptor, watermark embedding device based on Fourier descriptor, watermark extraction device based on Fourier descriptor and system

Abstract

The invention discloses a watermark embedding method based on a Fourier descriptor, a watermark extraction method based on the Fourier descriptor, a watermark embedding device based on the Fourier descriptor, a watermark extraction device based on the Fourier descriptor and a system. The watermark embedding method comprises the following steps of: extracting boundary coordinate points of an image watermark; constructing a complex sequence on the basis of the boundary coordinate points; performing discrete Fourier transform on the complex sequences to obtain the Fourier descriptors of the boundary coordinate points, and obtaining first M Fourier descriptor coefficients of the Fourier descriptor; performing block discrete Fourier transform on a host image to obtain an initial discrete cosine transform coefficient; respectively embedding the M Fourier descriptor coefficients into low-frequency coefficients and / or medium-frequency coefficients of an initial discrete cosine transform domain to obtain the current discrete cosine transform coefficient with the embedded Fourier descriptor coefficient; and performing inverse discrete cosine transform on the current discrete cosine transform coefficient to obtain the host image with embedded watermark information. When the method, the device and the system are used, the watermark embedding capacity achieves the breakthrough improvement on the premise of not influencing the quality of a carrier image and the invisibility of the watermark.

Description

technical field [0001] The invention belongs to the field of multimedia information security, and in particular relates to a method, device and system for embedding and extracting watermarks based on Fourier descriptors. Background technique [0002] Digital watermarking technology is a kind of copyright protection technology. It uses the characteristics of human hearing and visual systems to add certain identification information (digital watermarking) to images, audio, video and other carriers, making it difficult for people to distinguish the watermarked content. The difference between the original digital works and the original digital works, and then the added identification information can be extracted through special detection methods, which can be used for digital product copyright protection, authenticity identification of original data, data detection and copyright tracking, etc. [0003] Most of the traditional watermarking technology is to embed bits into the car...

AI Technical Summary

Problems solved by technology

This method improves the security of the watermark to a certain extent through watermark preprocessing, but has not made a qualitative breakthrough in the embedding capacity of the watermark. Although the grayscale image obtained through preprocessing is smaller than the original watermark image, it is known that the The watermark embedding method based on discrete wavelet transform cannot achieve a watermark embedding capacity of 1:1. Once the watermark is embedded in a 1:1 manner, it will cause a large area of ​​modification to the discrete wavelet coefficients, which will seriously affect the quality of the carrier image. Quality and watermark invisibility

[0007] In the patent scheme titled "Watermark Embedding and Extraction Method, Device and System" and Chinese Patent Application No. CN201310535907.3, the watermark embedding method is: parse the binary watermark image into several Regular discrete point sets, and then embed these discrete point sets into the host image, using this scheme, can introduce less redundant embedding, reduce the amount of modification to the original multimedia carrier to a certain extent, and improve the accuracy of the watermark information. embedding capacity, but this method is only suitable for those occasions where the watermark content is a relatively simple geometric shape. When the watermark content to be embedded is complex and diverse, and there are no fixed geometric rules to follow, the watermark content cannot be well processed at this time. Therefore, this shape-based watermark embedding and extraction method has certain limitations.

Images