34 results about "Color image watermarking" patented technology

The invention discloses a color image watermark embedding and extraction method based on fractional order hyper-chaos and discrete wavelet transform-singular value decomposition (DWT-SVD). The method includes: firstly, utilizing a fractional order hyper-chaos system to encrypt a color watermark image W0 to obtain an encrypted image W1; then converting the encrypted image W1 and an original color image I0 from RGB space to YIQ space to obtain images W2 and I1 respectively, and separating Y, I and Q components of the image; performing wavelet transform on components of the image I1 repeatedly, performing singular value decomposition on obtained low-frequency sub-bands, using components of the image W2 as watermark information to be embedded in corresponding singular values, and performing singular value decomposition on changed singular values again; and finally obtaining a color image with embedded watermark through wavelet inverse transformation and color space conversion. The color image watermark embedding and extraction method enhances imperceptibility, safety and abundant information of a watermark, has the advantages of being high in safety, strong in robustness, large in copyright information and rapid in arithmetic speed, and has strong application value.

The invention discloses a wavelet domain color image watermark encryption algorithm based on differential evolution. The operating method comprises the following steps: performing color space transformation on an original image to obtain a YIQ color format suitable for a human vision system and extracting a brightness component Y; adopting Arnold scrambling on a watermark image; performing three-stage discrete wavelet decomposition on the component Y of the original image to generate four sub-bands of different frequencies by utilizing discrete wavelet transform, thereby obtaining a sub-band image; representing a sub-band image matrix by using A, and performing singular value decomposition on the image subjected to three-stage wavelet decomposition; performing primary discrete wavelet decomposition on the watermark image subjected to scrambling encryption processing; performing watermark embedding; adaptively optimizing and selecting embedding strength by adopting differential evolution, and adding watermark information into the original image subjected to singular value decomposition in a multiplicative addition form. The invention has the beneficial effects that the watermark algorithm has high invisibility and is high in robustness.

The invention discloses a machine learning-based color image watermark embedding and detecting method. The method comprises the following steps of firstly, using a matrix-based image normalizing technology to map a carrier image to a geometric invariant space, and combining a invariant centroid theory to determine an important area in a normalized color image; secondly, performing non-subsample Contourlet conversion on the important area of the G-component normalized color image; finally, according to a G-component visual mask model, adaptively determining embedding intensity, so that an image watermark is embedded in a low-frequency area, and the embedded watermark is guaranteed to have very good transparency and very high robustness; during watermark detection, by using the height relativity among color image components, selecting a stable feature vector training SVR (Support Vector Regression) model, and using the SVR training model to extract digital watermark information.

The invention discloses a color image watermarking algorithm with good robustness. The color image watermarking algorithm combines the advantages of a spatial-domain watermarking method and a frequency domain watermarking method, aims to improve the instantaneity and the robustness of a digital watermarking method, and satisfies a requirement of color digital image copyright protection. The color image watermarking algorithm is characterized by blocking the blue component of a color host image; acquiring the DC coefficient and the modified value thereof of each 8*8 pixel child block in each block in the spatial domain according to a DC coefficient forming principle in the DCT domain; and then modifying the DC coefficient in the DCT domain to embed a watermark by directly modifying the size of a pixel in the spatial domain; repeating 4 times to embed the child watermarks into the image blocks on the basis of the original watermark blocks; not requiring the original watermark or the original host image in watermark extraction; determining optimum extracted child watermarks by using a majority principle; and combining the child watermarks to form a complete extracted watermark. The color image watermarking algorithm is simple and convenient, good in robustness, and suitable for copyright protection of the color image used as the host image.

The present invention belongs to image signal processing field, in particular a significative digital watermarking algorithm of the supercomplex frequency domain. The algorithm processing quick supercomplex Fourier transform to the color carrier image, selecting proper frequency band to embed in the watermark data in the supercomplex frequency domain, and repairing the value of coefficient of asymmetry, resolving the problem of the precondition of the supercomplex frequency domain embedding in the watermark, ensuring process transmittion by the red, green, blue of the color image when embedding in the watermark image, paving a road for further study of the color image watermarking algorithm of the supercomplex frequency domain. The algorithm also provides a regulation of ''intersecting redundancy embed'' of the watermark data, further improving the Robust of the digital watermark. The analyzing result shows that the proposed method diffuses the error caused by watermark to the whole imge by supercomplex Fourier transform, and diffuses to each weight of red, green, blue colors, achieving excellent combination of the non intangibity and safety of the digital watermark.

The invention relates to a low-false-alarm-rate color QR code digital watermarking method, which is a quick and efficient double color image watermarking algorithm based on color QR codes and DWT-SVD technology. Tone-level-rich color QR codes are normalized and embedded to a carrier image obtained after DWT conversion and SVD decomposition through different channels, so that the watermarking method not only has good invisibility, but also can be resistant to conventional attacks of cutting, rotation, JPEG compression, Gaussian noise and salt-and-pepper noise and the like, and has good robustness and higher QR code decoding rate. Insertion of a lot of information can be realized through insertion and extraction of the tone-level-rich color QR codes, thereby meeting requirements of high-end companies for the watermarking algorithm; meanwhile, the method also has faster operation speed, so that efficiency is improved greatly; and through insertion of a check mark, under the condition that the watermarked carrier image obtained after attacks can be extracted normally, the problem of high false alarm rate of an SVD decomposition algorithm is effectively solved. The watermarking method has a certain application value in digital right management.

The invention discloses a color image watermarking method based on a quaternion index matrix. The color image watermarking method can effectively resist the desynchronizing attack. The method comprises the following steps of calculating the quaternion index matrix of an original color image, reconstructing the original color image according to the quaternion index matrix, selecting a stable low-order quaternion index matrix, adopting a quantization modulation strategy to embed watermarking information into a quaternion index matrix amplitude, reconstructing the color image into which the watermarking information is embedded according to the selected quaternion index matrix and carrying out the margin calculation on the reconstructed image to superpose the obtained differential value into the original image.

The invention discloses a dual-color image blind watermarking method based on QR decomposition and compensation, which aims to change the current situation that the current digital watermarking is mostly a binary or grayscale image and meet the requirement that a color digital image serves as digital watermarking. According to the technical key point, color image watermarking is embedded into an orthogonal matrix Q obtained by QR decomposition; visual distortion caused by watermarking embedding is compensated by utilizing a triangular matrix R; and assistance of the original watermarking or the original host image is not needed during watermarking extraction, so that blind detection is realized. The method is simple and quick, has the advantages of high watermarking invisibility, strong robustness and high capacity, and is applicable to copyright protection of the color digital image serving as the digital watermarking.

The invention discloses a color image blind watermarking method. A color digital image is used as a digital watermark to be inserted in a color host image by utilizing Hessenberg matrix disintegration, and the inserted color image watermark is good in invisibility, so that the requirement for protecting a copyright based on a color image marker can be met. A 4*4 pixel block is disintegrated in a Hessenberg way, the encrypted color image digital watermark information is inserted into a maximum coefficient of the Hessenberg matrix by utilizing a coefficient quantization technology, the original host image or the original watermark image is not needed when the watermark is extracted, and a blind detection object can be realized. The method is good in watermark algorithm performance and suitable for the copyright protection adopting the color digital image as the digital watermark.

One example embodiment is a method to embed a watermark image into a host image such that a watermarked image is generated. The method divides a host image into an Red (R) component, a Green (G) component, and a Blue (B) component, and partitions each component of the R, G, and B components into non-overlapping blocks from which embedding blocks are selected to embed watermark information such that a watermarked image is generated.

The invention provides an imperceptible watermark attack method based on residual learning, a storage medium and an electronic device. The method comprises the steps: constructing a watermark attack model based on a convolutional neural network, carrying out the end-to-end nonlinear learning between a watermark-containing image and a watermark-free image, mapping the watermark-containing image to the watermark-free image, and carrying out watermark attack; selecting a proper number of feature extraction blocks according to the embedded area of the watermark information to extract a feature map containing the watermark information; introducing a residual learning mechanism to improve the convergence speed and learning ability of a watermark attack model, and improving the imperceptibility of an attacked image by reducing the difference between a residual image and the watermarking-free image; and constructing a data set for training a watermark attack model according to a DIV2K2017 super-resolution data set and a robust color image watermark algorithm based on a quaternion exponential moment. The watermark attack model provided by the invention can attack a robust watermark algorithm at a high bit error rate on the premise of not destroying the visual quality of a watermark-containing image.

The invention discloses a color image watermark embedding and detecting method based on quaternion Legendre moment correction. First, the Legendre invariant theory for traditional gray images is extended to a quaternion level, and a quaternion Legendre moment invariant suitable for describing color image content is constructed; and then, a digital watermark is embedded into a low-frequency region of a discrete quaternion cosine transform domain by a quantization index modulation (QIM) method. During watermark detection, part of low-order quaternion Legendre moment invariants are selected as feature vectors to obtain a least squares support vector machine (LS-SVM) training model so as to improve the performance of a color image classifier. Simulation experiments show that the method has good imperceptibility, and has good robustness to conventional signal processing (especially for the color attack mode) and de-synchronization attack.

The invention provides a hamming-code-based quantum image watermarking method. The method comprises the following steps that the original watermark image is subjected to correcting coding through the hamming code; pseudo random numbers are utilized to generate a watermark embedding position on the original quantum carrier image; an auxiliary quantum bit sequence is additionally arranged to the original quantum carrier image during embedding of the watermark; the coded watermark image is decoded through the hamming code. With the adoption of the method, the completeness and accuracy of embedding and extracting of the watermark image can be achieved; the efficiency is high; the operation complexity is small; the robustness is high; the watermark can be embedded into part of the position of the quantum image; the original watermark image can be obtained without the original quantum carrier image; the method can be applied to encrypting of the gray image watermark and encrypting of the colored image watermark.

The invention provides a color image watermark embedding method, a detection method and system. The color image watermark embedding method comprises the following steps: converting a source image froman RGB color space to a YCbCr color space; taking a Y-channel image as a carrier image, performing block discrete cosine transform, and extracting a coefficient vector of the watermark to be embeddedin each block; calculating a JND value based on the direction diversity for each block of the Y channel, selecting a Cb color channel to calculate a color complexity influence factor, and constructing a JND perception model based on the direction diversity and the color contrast; calculating a quantization step size corresponding to each coefficient vector, and embedding watermark sequence information into each block by utilizing a jitter modulation algorithm of a logarithm domain; and performing reconstruction to obtain a Y-channel image with the watermark embedded, and converting the Y-channel image, a Cb-channel image of the source image and a Cr-channel image into an RGB color space to obtain a final color image with the watermark embedded.

The invention discloses a color image watermark embedding and detection method based on quaternion PHT synchronous correction, comprising: converting color RGB images into a YUV space, and respectively selecting appropriate embedding intensities to embed watermarks; defining the quaternion PHT of the color images; training an LS-SVR model, constructing a training sample set, performing circumcircle quaternion PHT decomposition on each training image, extracting a specific moment value as the feature vector of the training image, and using a relevant information coefficient to describe each training sample; utilizing circumcircle quaternion PHT to decompose an images to be detected, and decomposing an obtained moment value as a sample feature vector; utilizing a training model to train the feature vector of the image to be detected, solving a geometric transformation parameter, and utilizing the geometric transformation parameter to perform synchronous correction on the image be detected to obtain a corrected image; and extracting the watermarks in a corresponding mode of the embedded watermarks.

The invention discloses a color image watermarking device, which comprises a color image processing system and a watermark processing system, and is characterized in that the color image processing system comprises a color image processing module, a color image extraction module and an image analysis module, and the watermark processing system comprises a watermark generation module, a watermark extraction module and a watermark analysis module; the output ends of the picture analysis module and the watermark analysis module are connected to the input end of the preprocessing module, the inputend of the preprocessing module is further connected to the output end of the position marking module, the output end of the preprocessing module is connected to the input end of the processing module, and the output end of the processing module is connected to the input end of the judgment module. According to the invention, the simulation operation of watermark adding by the subsequent preprocessing module is facilitated, and the situation that the watermark cannot be added to the color picture or the watermark is not obvious after being added is avoided; simulation addition can be carriedout before final watermark addition, and meanwhile, the effect after watermark addition can be simulated for the first time.

The invention discloses a color image watermarking method based on local histogram characteristics. The method comprises the following steps that a color invariant model of a color image is established by utilizing the chromatic characteristics and the geometric construction characteristics of the color image; an image characteristic point detector based on color invariants and probability density is structured, and stable and even characteristic points are extracted from an original color image; an oval local characteristic area is structured by utilizing characteristic values and characteristic vectors of a probability density second order autocorrelation matrix, the extracted characteristic points and the screened-out oval local characteristic area are mapped back to the original color image, and a color affine invariant local characteristic area is extracted; a high-order flat image histogram of the color affine invariant local characteristic area is extracted, and watermarking information is embedded by carrying out histogram adjustment on RGB components of an extracted high-order flat image respectively.

The invention discloses a self-adaptive color digital image blind watermarking method by utilizing the advantage of high robustness of a transform domain digital watermarking algorithm. According to the method, different embedding coefficients are adaptively selected from different image blocks, and part of intermediate frequency coefficients after two-dimensional discrete cosine transform are modified in a transform domain, so that embedding and blind extraction of the color digital watermark are completed; in the watermark preprocessing stage, affine transformation and Arnold transformationare sequentially used for carrying out dual encryption on the watermark, and the safety of a watermark algorithm is greatly improved. According to the method, the color image watermark is embedded into the color host image, so that the method not only has better watermark invisibility and stronger robustness, but also has very high safety, and is suitable for occasions of digital image copyright protection.