Feature point enhanced local watermark embedding and extracting method under perceptual constraint and guidance

Abstract

The invention discloses a feature point enhanced local watermark embedding and extracting method under perceptual constraint and guidance. The embedding method comprises the following steps: carrying out SIFT feature point detection on a watermark image to be embedded to obtain a plurality of feature points; sorting the feature points according to the intensity of the feature points from large to small, and selecting a plurality of non-overlapping areas positioned by the plurality of feature points with the highest intensity as watermark embedding areas; partitioning each watermark embedding area according to the length of a watermark information bit; a watermark is embedded in each block of each watermark embedding area by using a fusion feature point enhanced watermark embedding strategy; and replacing a corresponding region in the image to be embedded with the watermark by using the watermark embedding region embedded with the watermark to obtain a watermark-containing image. Compared with an existing self-adaptive local image watermark embedding and extracting method, the self-adaptive local image watermark embedding and extracting method has better invisibility and higher robustness.

Description

Technical Field

[0001] The present invention relates to the field of digital image copyright protection, and in particular to a feature point enhanced local watermark embedding and extraction method under perception constraints and guidance. Background Art

[0002] While the development of online media has brought convenience to people, it has also led to the phenomenon of digital works being copied and illegally disseminated, thus causing serious copyright disputes. Digital image watermarking technology embeds watermark information containing copyright identification into images covertly. When copyright infringement occurs, the watermark information can be extracted for copyright authentication and infringement tracking. Watermarking technology should be both imperceptible and robust. The former requires that the embedded watermark does not affect the visual quality of the image, and the latter requires ensuring that the watermark can be correctly extracted after the image has been subjected to various attacks. While ensuring the imperceptibility of the watermark, maximizing its anti-attack capability is the key to the effective application of watermarking technology.

[0003] Local watermarking technology embeds watermark information into multiple non-overlapping areas of an image. Local embedding reduces the visual impact on the overall image and improves the imperceptibility of the watermark. Due to the multiple embedding of watermarks, even if some areas are damaged, it will not affect the watermarks in other areas, making it more robust. This technology usually uses feature points in the image, such as SIFT (Scale-Invariant Feature Transform), to locate the embedded area. Consistent feature points must be detected in the two stages of watermark embedding and extraction, and the embedded area must be accurately located before the watermark can be successfully extracted. Since the attack will cause changes in the image content, it may cause the position of the feature points to shift or be lost, making the watermark unable to be correctly extracted. Some studies have attempted to enhance the stability of feature points by modifying the pixel values ​​around the feature points. However, since the watermark embedding process also needs to modify the pixel values ​​in the same area, these two modifications may produce contradictory effects, thereby reducing each other's effectiveness and affecting the imperceptibility and robustness of the watermark.

[0004] Watermark embedding is essentially to make subtle modifications to image pixels according to a specific scheme. Generally speaking, the larger the pixel modification amount, the harder it is to destroy the embedded watermark and the stronger its robustness; but at the same time, it will also have a greater impact on the image quality and reduce the imperceptibility of the watermark. Usually, a parameter is introduced in the watermark algorithm to control the overall pixel modification amount to balance the imperceptibility and robustness of the watermark. Some studies adaptively adjust the value of this parameter based on the statistical characteristics or content characteristics of the image, so that the modification amount varies according to the image content, aiming to make the embedded watermark less noticeable. However, they can only adjust the overall modification amount of the image or region, and cannot allocate the modification amount to a single pixel. Due to the difference in sensitivity of the human eye to different pixels, the ability of different pixels to hide watermarks is different. Simply controlling the overall modification amount cannot ensure that each pixel change is not noticed by the human eye. Summary of the invention

[0005] In order to solve the problem of poor watermark robustness caused by feature point offset after the local watermark algorithm embeds the watermark, and to overcome the defects of the existing watermark technology in optimizing the invisibility of the watermark, the present invention proposes a feature point enhanced local watermark embedding and extraction method under perceptual constraints and guidance, which maximizes the robustness of watermark extraction while ensuring the invisibility of the embedded watermark.

[0006] According to one aspect of the present invention, a method for local watermark embedding enhanced by feature points under perceptual constraints and guidance is provided, and the method specifically comprises:

[0007] Perform SIFT feature point detection on the image to be embedded with watermark to obtain several feature points;

[0008] Sort the feature points according to their strength from large to small, and select multiple non-overlapping areas where multiple feature points with the highest strength are located as watermark embedding areas;

[0009] Dividing each watermark embedding area into blocks according to the length of the watermark information bit;

[0010] Embedding a bit of watermark in each of the blocks in each of the watermark embedding areas by using a watermark embedding strategy enhanced by fusion feature points;

[0011] The watermark embedding area after embedding the watermark replaces the corresponding area in the image to be embedded with the watermark to obtain the watermarked image.

[0012] The step of embedding a bit of watermark in each of the blocks in each of the watermark embedding areas by using a watermark embedding strategy enhanced by fusion feature points includes:

[0013] The embedding parameters in the watermark embedding strategy for fusion feature point enhancement are determined according to a pre-established PSNR (Peak Signal to Noise Ratio)-based quantization parameter model.

[0014] Classifying the blocks according to whether they contain feature points, the blocks containing feature points are called central image blocks, and the blocks without feature points are called non-central image blocks;

[0015] Under the guidance of JND (Just Noticeable Distortion), a quantization method is used to embed a one-bit watermark in each non-central image block. The quantization formula is shown in formula (1), the modification weight is shown in formula (2), and the embedding formula is shown in formula (3):

[0016] (1)

[0017] (2)

[0018] (3)

[0019] in, i, j Indicates the index of the image block. i,j and B′ i,j represent the non-central image blocks in the image to be embedded with watermark and the image containing watermark respectively, u i,j and u′ i,j Respectively represent B i,j and B′ i,j The pixel mean. q is the embedding parameter, also known as the quantization step size, used to control the pixel modification amount. i,j Indicates the watermark position to be embedded. round() is an integer function, and mod represents the remainder operation. i,j is the image block B i,j The JND matrix of each pixel in M i,j represents the pixel modification weight of the non-central image block. sum() is a summation function.

[0020] Under the guidance of JND, an optimization function is used to embed a one-bit watermark in the central image block. The weight matrix is ​​shown in formula (4), and the optimization function is shown in formula (5):

[0021] (4)

[0022] (5)

[0023] (6)

[0024] (7)

[0025] Among them, B o The image I to be embedded with watermark contains feature points p o The central image block, Jnd o Indicates B o The JND matrix of M o represents the pixel modification weight of the central image block. I′ is the watermarked image, B o ′ is the watermarked image with B o The corresponding image block. min means taking the minimum value, ||M o (B o -B o ′)|| represents the second norm. Constraint (C.1) ensures that p o The intensity of the o :maximun and p o :minimun respectively represents p o It is centered on itself. po The maximum and minimum intensity feature points in the neighborhood. I (p o ) represents the feature point p o The initial strength value, D I’ (p o ) represents the feature point p o The enhanced strength value, Represents the enhanced value of the feature point strength. Constraint (C.2) ensures that S po No new feature points are generated in the neighborhood. max and p min Respectively represent S po The maximum and minimum feature strengths in the neighborhood. Constraints (C.3) ensure the embedding of the watermark.

[0026] The calculation method of the PSNR-based quantization parameter model specifically includes:

[0027] According to the embedding formula (1), it is assumed that each pixel in the block is modified by Δu i,j , then the mean square error MSE (Mean Square Error) of the entire watermark embedding area and the PSNR satisfy the following relationship.

[0028] (8)

[0029] With Δu i,j 2 The expected value of q2 / 4 estimates the MSE of the embedded area and substitutes it into formula (8) to obtain the PSNR-based quantization parameter model as follows:

[0030] (9)

[0031] According to formula (9), the value of the embedding parameter is calculated using the given PSNR value.

[0032] According to another aspect of the present invention, a method for extracting a local watermark enhanced by feature points under perceptual constraints and guidance is provided. The method is used to extract a watermark embedded by the method for embedding a local watermark enhanced by feature points under perceptual constraints and guidance provided by the present invention, and specifically comprises:

[0033] Perform SIFT feature point detection on the watermarked image to obtain several feature points;

[0034] Sort the feature points according to their strength from large to small, and select multiple non-overlapping areas where multiple feature points with the highest strength are located as watermark extraction areas;

[0035] Dividing each of the watermark extraction regions into blocks according to the length of the watermark information bits;

[0036] A watermark is extracted from each block of each watermark extraction area, and finally the complete watermark information is obtained. The extraction formula is as follows:

[0037] (10)

[0038] in, i, j Represents the index of the image block. Represents the calculated image block mean. Indicates the extracted watermark position. round() is an integer function, and mod represents the remainder operation.

[0039] Compared with the prior art, the above technical solution has at least the following beneficial effects:

[0040] The present invention integrates feature point enhancement into the watermark embedding process, and realizes watermark embedding and feature point enhancement simultaneously through a single modification of image pixels, thereby avoiding mutual interference between the two processes of watermark embedding and feature point enhancement, and effectively improving the stability of feature points and the robustness of watermarks. A quantization parameter model based on PSNR is used to limit the total amount of pixel modification, and a pixel-level weighting strategy guided by JND is used to achieve differentiated distribution of the total amount of modification at different pixels, thereby ensuring that the watermark has good imperceptibility at both the overall image block and the single image pixel levels. The present invention shows relatively good robustness against most common image attacks and geometric attacks, and effectively solves the problem of poor robustness caused by feature point offset in local image watermarking methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The accompanying drawings are part of the present invention and are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention, but do not constitute an improper limitation of the present invention. Obviously, the drawings described below are only some embodiments. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the accompanying drawings:

[0042] Figure 1 A schematic flow chart of a method for local watermark embedding with feature point enhancement under perceptual constraints and guidance according to an exemplary embodiment;

[0043] Figure 2 is an image to be embedded with a watermark according to an exemplary embodiment;

[0044] Figure 3 is a watermarked image according to an exemplary embodiment;

[0045] Figure 4 The figure shows the change of the intensity of the corresponding feature points in the image to be embedded with a watermark and the image containing the watermark according to an exemplary embodiment;

[0046] Figure 5 FIG. 1 is a comparison of PSNR values ​​between an embedding area in an image to be embedded with a watermark and a corresponding area in an image containing a watermark according to an exemplary embodiment;

[0047] Figure 6 A schematic flow chart of a method for extracting local watermarks enhanced by feature point enhancement under perceptual constraints and guidance according to an exemplary embodiment;

[0048] Figure 7 FIG. 4 is a histogram distribution of image block means according to an exemplary embodiment.

[0049] These drawings and textual descriptions are not intended to limit the conceptual scope of the present invention in any way, but to explain the concept of the present invention to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0050] The following is a clear and complete description of the technical problems solved by the embodiments of the present invention, the technical solutions adopted, and the technical effects achieved in conjunction with the accompanying drawings and specific embodiments. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other equivalent or obviously modified embodiments obtained by ordinary technicians in this field without creative work fall within the scope of protection of the present invention. The embodiments of the present invention can be concretized in a variety of different ways as defined and covered in the claims.

[0051] It should be noted that in the following description, many specific details are given for ease of understanding, but it is obvious that the present invention can be implemented without these specific details.

[0052] It should be noted that, in the absence of clear limitation or conflict, the various embodiments of the present invention and the technical features therein can be combined with each other to form a technical solution.

[0053] The embodiment of the present invention provides a method for embedding a local watermark enhanced by feature points under perceptual constraints and guidance. Figure 1 As shown, the method may include: step S1000 to step S1400.

[0054] Step S1000: Perform SIFT feature point detection on the image to be embedded with watermark to obtain a number of feature points.

[0055] The specific description is as follows: The Gaussian filter kernel is used to filter the image to be embedded with watermark to obtain a series of Gaussian filter images, and the differences between adjacent Gaussian filter images are extracted to establish a Gaussian difference image (DOG, Difference of Gaussian). represents a position in the DOG image at row x and column y. D(p) represents the value of position p in the DOG image. ≤ 1, represents the 26 positions in the 3×3×3 cube centered at position p in the DOG image, where and Respectively represent the row and column coordinates of pixels in the DOG image, Denotes the variance value. Let D(p) be the same as the 26-neighborhood S pIf D(p) is the maximum or minimum value, the position p is considered to be an extreme point of the image. Excluding extreme points with low contrast and at the edge, SIFT feature points are obtained.

[0056] Step S1100: Sort the feature points according to their strength from large to small, and select multiple non-overlapping areas where multiple feature points with the highest strength are located as watermark embedding areas.

[0057] The strength of the feature point is defined as |D(p)|. The feature points are sorted from large to small according to their strength, and a non-overlapping area centered on the first N feature points is selected as the watermark embedding area. The ratio of the size of the watermark embedding area to the size of the image to be embedded is K.

[0058] Step S1200: Divide each watermark embedding area into blocks according to the length L of the watermark information bit. Divide the watermark embedding area into a×b image blocks according to the watermark length L, ensuring that a and b are as equal as possible, where a and b are two positive integers and satisfy a×b≥L. The length of the image block is h and the width is w.

[0059] Step S1300: embedding a bit of watermark in each of the blocks in each of the watermark embedding areas by using a watermark embedding strategy that incorporates feature point enhancement.

[0060] Specifically, this step may include: step S1310 to step S1340.

[0061] Step S1310: Determine embedding parameters in the watermark embedding strategy for fusion feature point enhancement according to a pre-established PSNR-based quantization parameter model.

[0062] The reason for performing this step is: the embedded parameter limiting pixel modification amount Δu is calculated using the PSNR-based quantization parameter model i,j , the PSNR of the watermarked image can be kept consistent with the preset value, ensuring that the overall invisibility of the watermark meets expectations.

[0063] Step S1320: Classify the blocks according to whether they contain feature points. The blocks containing feature points are called central image blocks, and the blocks without feature points are called non-central image blocks.

[0064] Step S1330: Under the guidance of JND, a quantization method is used to embed a one-bit watermark in each non-central image block. The quantization formula is shown in formula (1), the modified weight is shown in formula (2), and the embedding formula is shown in formula (3):

[0065] (1)

[0066] (2)

[0067] (3)

[0068] in, i, j Indicates the index of the image block. i,j and B′ i,j represent the non-central image blocks in the image to be embedded with watermark and the image containing watermark respectively, u i,j and u′ i,j Respectively represent B i,j and B′ i,j The pixel mean. q is an embedding parameter, also known as the quantization step size, which is used to control the amount of pixel modification. i,j Indicates the watermark position to be embedded. round() is an integer function, and mod represents the remainder operation. i,j is the image block B i,j The JND matrix of each pixel in M i,j represents the pixel modification weight of the non-central image block. sum() is a summation function. if represents if.

[0069] The reason for performing the steps is that the sensitivity of the human eye to different pixels in the airspace is different, which makes the ability of different pixels to hide watermarks different. Simply controlling the overall modification amount cannot ensure that each pixel change is not perceived by the human eye. It is necessary to further allocate the modification amount of different pixels according to the hiding ability of different pixels. Based on this, in order to ensure that the image modification is consistent with the visual perception of the human eye and optimize the imperceptibility of the watermark, when modifying pixels in the airspace, the present invention, under the guidance of JND, weights the modification amount of the quantization value (i.e., the overall pixel modification amount) to each pixel, and allocates more modification amounts to pixels with lower sensitivity to the human eye (stronger hiding ability).

[0070] Step S1340: Under the guidance of JND, a one-bit watermark is embedded in the central image block using an optimization function. The weight matrix is ​​shown in formula (4), and the optimization function is shown in formula (5):

[0071] (4)

[0072] (5)

[0073] (6)

[0074] (7)

[0075] Among them, B o The image I to be embedded with watermark contains feature points p oThe central image block, Jnd o Indicates B o The JND matrix of M o represents the pixel modification weight of the central image block. I′ is the watermarked image, B o ′ is the watermarked image with B o The corresponding image block. min means taking the minimum value, ||M o (B o -B o ′)|| represents the second norm. Constraint (C.1) ensures that p o The intensity of the o :maximun and p o :minimun respectively represents p o It is centered on itself. po The maximum and minimum intensity feature points in the neighborhood. I (p o ) represents the feature point p o The initial strength value, D I’ (p o ) represents the feature point p o The enhanced strength value, Represents the enhanced value of the feature point strength. Constraint (C.2) ensures that S po No new feature points are generated in the neighborhood. max and p min Respectively represent the neighborhood S po The maximum and minimum feature strengths within . Constraints (C.3) ensure the embedding of watermarks.

[0076] The reason for performing the steps is that if the region is modified twice independently, feature point enhancement and watermark embedding are achieved respectively. Although the stability of feature points can be improved, there are two main disadvantages: first, the feature point enhancement process brings additional modifications, which will reduce the image quality; second, watermark embedding may weaken the strength of feature points, resulting in failure of feature point enhancement. Therefore, the present invention realizes a watermark embedding algorithm that integrates feature point enhancement through one modification, which can effectively enhance the stability of feature points.

[0077] Specifically, the calculation method of the PSNR-based quantization parameter model in step S1310 may include: steps S1311 to S1313.

[0078] Step S1311: According to the embedding formula (1), assuming that the modification amount of each pixel in the block is , then the mean square error MSE and PSNR of the entire watermark embedding area satisfy the following relationship:

[0079] (8)

[0080] The calculation formula of PSNR is: , extract the MSE through conversion, and then combine it with the calculation formula of MSE We can get formula (8).

[0081] Step S1312: Use The expected value of q 2 / 4 estimates the MSE of the embedded area and substitutes it into formula (8) to obtain the PSNR-based quantization parameter model as follows:

[0082] (9)

[0083] in, The calculation of the expected value of is based on the following considerations:

[0084] When quantizing, the modification amount of each image block is Distributed in [-l q, l q ]between, Distributed in [0 , l q ], and its value follows the mean value l q / 2 uniform distribution, so we can get The expected value of q 2 / 4.

[0085] The above conclusion is supported by a preferred embodiment: 24 images are randomly selected, 5 embedding regions are determined for each image, and each embedding region is divided into 64 image blocks, for a total of 7680 image blocks. q =14, according to formula (1), we get 7680 The average value of is 7.02. Their histogram distribution ( Figure 7 ) can be verified The mean is l q / 2 uniform distribution.

[0086] Step S1313: According to formula (9), the value of the embedding parameter is calculated using the given PSNR value.

[0087] Step S1400: replacing the corresponding area in the image to be embedded with the watermarked area after the watermark is embedded, to obtain the watermarked image.

[0088] The embodiment of the present invention provides a method for extracting local watermarks with feature point enhancement under perceptual constraints and guidance. Figure 6As shown, the method may include: step T1000 to step T1300.

[0089] Step T1000: using the same method as step S1000 of the feature point enhanced local watermark embedding method under perception constraint and guidance in the first embodiment, SIFT feature point detection is performed on the watermarked image to obtain a plurality of feature points;

[0090] Step T1100: using the same method as step S1100 of the feature point enhanced local watermark embedding method under perception constraint and guidance in the first embodiment, the feature points are sorted from large to small according to their strength, and multiple non-overlapping areas where multiple feature points with the highest strength are located are selected as watermark extraction areas;

[0091] Step T1200: using the same method as step S1200 of the feature point enhanced local watermark embedding method under perception constraint and guidance in the embodiment, dividing each watermark extraction area into blocks according to the length L of the watermark information bit;

[0092] Step T1300: extract one bit of watermark from each of the blocks in each of the watermark extraction areas, and finally obtain complete L-bit watermark information. The extraction formula is as follows:

[0093] (10)

[0094] Among them, i, j represents the index of the image block. Represents the calculated image block mean. Indicates the extracted watermark position. round() is an integer function, and mod represents the remainder operation.

[0095] like Figure 2 and 3 As shown, Figure 2 An image before watermarking is exemplarily shown, i.e., the image to be watermarked; Figure 3 An image after embedding a watermark, ie, the watermarked image, is shown as an example.

[0096] A method for local watermark embedding enhanced by feature points under perceptual constraints and guidance is described in detail below with a preferred embodiment.

[0097] This embodiment randomly selects one image as the image to be embedded with watermark, such as Figure 2 As shown, the image size is 512 512.

[0098] This embodiment uses a 64-bit binary sequence as the original watermark, and the length L of the watermark information is 64 bits, which can basically achieve a relatively comprehensive description of the copyright information.

[0099] Perform SIFT feature point detection on the image to be embedded with watermark and obtain 100 feature points.

[0100] The feature points are sorted from large to small according to their strength, and five non-overlapping areas located by the five feature points with the highest strength are selected as watermark embedding areas. The ratio K of the watermark embedding area size to the size of the image to be embedded with the watermark is 1 / 6, which is 85. 85.

[0101] According to the length L = 64 of the watermark information bit, each watermark embedding area is divided into 8×8 image blocks.

[0102] The image blocks are classified according to whether they contain feature points. The image blocks containing feature points are called central image blocks, and the image blocks without feature points are called non-central image blocks. Then, for each watermark embedding area, 1 central image block and 63 non-central image blocks can be obtained.

[0103] Set the target PSNR to 40dB and use the PSNR-based quantization parameter model formula (9) to determine the quantization step size l q The value of is 15.

[0104] The quantized watermark embedding shown in formula (1), formula (2) and formula (3) is performed on the 63 non-central image blocks respectively.

[0105] For the central image block, the watermark embedding based on the optimization function shown in formula (4), formula (5), formula (6) and formula (7) is performed. The optimization function is used to find a pixel modification scheme that can simultaneously embed the watermark and enhance the strength of the feature points. The enhanced value of the embedding strength is Set to 3.

[0106] All the central image blocks and the non-central image blocks that meet the conditions are found, and the corresponding image blocks in the image to be embedded with watermark are replaced to achieve watermark embedding with enhanced fusion feature points, and obtain a watermarked image.

[0107] It should be noted that:

[0108] If the number of feature points that meet the conditions is less than 5, only the non-overlapping embedding area is determined by the feature points that actually meet the conditions. In this preferred embodiment, after the watermark is embedded according to the above steps, the peak signal-to-noise ratio of the watermarked image and the image to be embedded with the watermark is 40.23dB, which basically meets the preset value of 40dB. Generally, the larger the peak signal-to-noise ratio, the better the visual quality of the image. When the peak signal-to-noise ratio reaches above 40dB, the human eye cannot see the difference between the image before and after the watermark is embedded. Figure 2 and Figure 3shown.

[0109] A method for extracting local watermarks enhanced by feature points under perceptual constraints and guidance is described in detail below using a preferred embodiment.

[0110] This embodiment randomly selects one image as the watermarked image. Figure 3 As shown, the image size is 512 512.

[0111] Perform SIFT feature point detection on the watermarked image and obtain 100 feature points.

[0112] The feature points are sorted from large to small according to their strength, and 10 non-overlapping areas located by the 10 feature points with the highest strength are selected as watermark extraction areas. The ratio of the watermark extraction area size to the watermarked image size is 1 / 6, which is 85. 85.

[0113] Each watermark extraction region is divided into blocks according to the number of bits of the watermark information 64, and the watermark extraction region is divided into 8×8 image blocks.

[0114] According to formula (1), the mean value of a certain block in the watermark extraction area is obtained in the spatial domain. , the watermark extraction formula (10) is used to extract the watermark in each image block in turn, and finally the complete 64-bit watermark information is obtained.

[0115] The reason for selecting 10 feature points to locate the watermark extraction area during extraction is that in practical applications, after the image has been processed in various ways, the original feature points may change in intensity, making it difficult to determine all the original feature points and their corresponding areas through the original number. When extracting watermarks, increasing the number of selected feature points can cover a wider range, reduce the risk of missing the original feature points, and increase the probability of successfully extracting watermarks.

[0116] In this embodiment, 110 images of different styles, sizes and scenes are randomly selected as images to be embedded with watermarks. 110 images containing watermarks are obtained by embedding watermarks using the method proposed in the present invention. The stability of the feature points is evaluated by the change in the intensity of the feature points before and after the watermark is embedded in the image. The deviation between the PSNR values ​​of the original image area and the watermarked area and the preset PSNR value is compared to evaluate the effectiveness of the PSNR-based quantization parameter model of the present invention. The robustness of the watermark is evaluated by the average extraction accuracy under different types of image attacks with different parameters, that is, the percentage of images that can correctly extract the complete watermark to the total number of images containing watermarks.

[0117] For the 110 watermarked images obtained by the watermark embedding method embodiment, the intensity of the feature points in the watermarked images is detected and compared with the intensity of the corresponding feature points in the original images, and the intensity change results of the feature points before and after the watermark is embedded are obtained as shown in FIG. Figure 4 As shown in the results, it can be seen that the invention enhances the fusion of feature points into the watermark embedding process, avoids the weakening of feature point enhancement by watermark embedding, and can effectively enhance the strength of feature points and improve their stability.

[0118] For the 110 watermarked images obtained by the above watermark embedding method embodiment, the relationship between the actual PSNR and the preset PSNR of the watermark embedding area is statistically analyzed. Figure 5 It can be seen that the PSNR of the embedded area is basically consistent with the preset value, which can prove the effectiveness of the quantization parameter model based on PSNR in the present invention.

[0119] For the 110 watermarked images obtained by the above watermark embedding method embodiment, various image attacks were respectively implemented, including: speckle noise with energy of 0.01 and 0.02, salt and pepper noise with energy of 0.01 and 0.02, 3×3 and 5×5 median filtering, 3×3 and 5×5 mean filtering, scaling attack with parameters of 0.80 and 0.75, cropping with ratios of 0.1 and 0.2, and rotation attack with parameter 2, etc., and then a feature point enhanced local watermark extraction method embodiment under perception constraint and guidance was used to extract the watermark from the watermarked image after the attack, and compared with the original watermark, the average extraction accuracy of the 110 extracted watermarks is shown in Table 1:

[0120] Table 1

[0121]

[0122] From Table 1, Figure 2 , Figure 3 , Figure 5 It can be seen from the results that, in the face of common image attacks, the embodiments of the present invention can basically extract 100% of the watermark, which proves that the present invention can not only effectively resist various image attacks and has strong robustness, but also ensure the invisibility of the watermark in each image.

[0123] Those skilled in the art should understand that the above assumptions are merely examples and should not be regarded as improper limitations on the present invention.

[0124] Although the various steps in the above embodiment are described in the above order, those skilled in the art can understand that in order to achieve the effect of this embodiment, different steps do not have to be executed in such an order, and they can be executed simultaneously (in parallel) or in a reverse order. These simple changes are within the scope of protection of the present invention.

[0125] The technical solutions provided by the embodiments of the present invention are described in detail above. Although this article uses specific examples to illustrate the principles and implementation methods of the present invention, the description of the above embodiments is only applicable to help understand the principles of the embodiments of the present invention; at the same time, for those skilled in the art, changes will be made in the specific implementation methods and application scopes according to the embodiments of the present invention.

[0126] It should be noted that the flowcharts or block diagrams involved in this article are not limited to the forms shown in this article, and they can also be divided and / or combined.

[0127] It should be noted that the marks and texts in the accompanying drawings are only for more clear explanation of the present invention and shall not be regarded as improper limitation of the protection scope of the present invention.

[0128] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein.

[0129] The term "comprise" or any other similar term is intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus / device that includes a list of elements includes not only those elements but also other elements not expressly listed, or also includes elements inherent to such process, method, article, or apparatus / device.

[0130] The various steps of the present invention can be implemented by a general-purpose computing device, for example, they can be concentrated on a single computing device, such as a personal computer, a server computer, a handheld device or a portable device, a tablet device or a multi-processor device, or distributed on a network composed of multiple computing devices, which can perform the steps shown or described in a different order from that herein, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Therefore, the present invention is not limited to any specific hardware and software or their combination.

[0131] The method provided by the present invention can be implemented using a programmable logic device, or can be implemented as a computer program software or a program module (which includes routines, programs, objects, components or data structures, etc. that perform specific tasks or implement specific abstract data types). For example, according to an embodiment of the present invention, it can be a computer program product, and running the computer program product enables the computer to execute the method for demonstration. The computer program product includes a computer-readable storage medium, which contains a computer program logic or code portion for implementing the method. The computer-readable storage medium can be a built-in medium installed in a computer or a removable medium that can be removed from the computer body (for example, a storage device using hot-swap technology). The built-in medium includes, but is not limited to, a rewritable non-volatile memory, such as RAM, ROM, flash memory and hard disk. The removable medium includes, but is not limited to, optical storage media (for example, CD-ROM and DVD), magneto-optical storage media (for example, MO), magnetic storage media (for example, magnetic tape or mobile hard disk), a medium with a built-in rewritable non-volatile memory (for example, a memory card) and a medium with a built-in ROM (for example, a ROM box).

[0132] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A feature point enhanced local watermark embedding method under perceptual constraints and guidance, specifically comprising: Step S1: Perform SIFT feature point detection on the image to be embedded with watermark to obtain several feature points; Step S2: sorting the feature points from large to small according to their strength, and selecting multiple non-overlapping areas where multiple feature points with the highest strength are located as watermark embedding areas; Step S3: dividing each watermark embedding area into blocks according to the length of the watermark information bit; Step S4: embedding a bit of watermark in each of the blocks in each of the watermark embedding areas by using a watermark embedding strategy enhanced by fusion feature points; Step S5: replacing the corresponding area in the image to be embedded with the watermark with the watermark-embedded area after the watermark is embedded, to obtain the watermarked image.

2. According to the method of claim 1, the feature point enhanced local watermark embedding method under perceptual constraints and guidance is characterized in that: The step S4 comprises: Step S41: Determine the embedding parameters in the watermark embedding strategy for fusion feature point enhancement according to the pre-established PSNR-based quantization parameter model; Step S42: classifying the blocks according to whether they contain feature points, the blocks containing feature points are called central image blocks, and the blocks without feature points are called non-central image blocks; Step S43: Under the guidance of JND, a quantization method is used to embed a one-bit watermark in each non-central image block; the quantization formula is shown in formula (1), the modification weight is shown in formula (2), and the embedding formula is shown in formula (3): （1） （2） （3） in, i, j Indicates the index of the image block; B i,j and B′ i,j represent the non-central image blocks in the image to be embedded with watermark and the image containing watermark respectively, u i,j and u′ i,j Respectively represent B i,j and B′ i,j The pixel mean; l q is the embedding parameter, also called quantization step size, used to control the pixel modification amount; i,j Indicates the watermark position to be embedded; round() is a rounding function, mod represents the remainder operation; Jnd i,j is the image block B i,j The JND matrix of each pixel in M i,j represents the pixel modification weight of the non-central image block; sum() is a summation function; Step S44: Under the guidance of JND, a one-bit watermark is embedded in the central image block using an optimization function; the weight matrix is ​​shown in formula (4), and the optimization function is shown in formula (5): （4） （5） （6） （7） Among them, B o The image I to be embedded with watermark contains feature points p o The central image block, Jnd o Indicates B o The JND matrix of M o represents the pixel modification weight of the central image block; I′ is the watermarked image, B o ′ is the watermarked image with B o The corresponding image block; min means taking the minimum value, ||M o (B o -B o ′)|| represents the bi-norm; the constraint (C.1) ensures that p o The intensity of the o :maximun and p o :minimun respectively represents p o It is centered on itself. po The maximum and minimum intensity feature points in the neighborhood; D I (p o ) represents the feature point p o The initial strength value, D I’ (p o ) represents the feature point p o The enhanced strength value, Represents the enhanced value of the feature point strength; the constraint (C.2) ensures that S po No new feature points are generated in the neighborhood; p max and p min Respectively represent S po The maximum and minimum feature strengths in the neighborhood. Constraints (C.3) ensure the embedding of the watermark.

3. According to the method of embedding local watermarks with feature point enhancement under perceptual constraints and guidance according to claim 2, it is characterized in that: The step S41 comprises: Step S411: According to the embedding formula (1), assuming that the modification amount of each pixel in the block is Δu i,j , then the mean square error MSE and PSNR of the entire watermark embedding area satisfy the following relationship: （8） Step S412: Use Δu i,j 2 The expected value of q 2 / 4 estimates the MSE of the embedded area and substitutes it into formula (8) to obtain the quantization parameter model based on PSNR as follows: （9） Step S413: According to formula (9), the value of the embedding parameter is calculated using the given PSNR value.

4. A method for extracting local watermarks enhanced by feature points under perceptual constraints and guidance, the method being used to extract a watermark embedded by the method for embedding local watermarks enhanced by feature points under perceptual constraints and guidance as claimed in any one of claims 1 to 3, comprising the following steps: Step T1: Using the same method as step S1, SIFT feature point detection is performed on the watermarked image to obtain a number of feature points; Step T2: using the same method as step S2, the feature points are sorted from large to small according to their strength, and multiple non-overlapping areas where multiple feature points with the highest strength are located are selected as watermark extraction areas; Step T3: using the same method as step S3, dividing each watermark extraction area into blocks according to the length of the watermark information bit; Step T4: extracting a watermark from each of the blocks in each of the watermark extraction areas, and finally obtaining complete watermark information. The extraction formula is as follows: （10） in, i, j Represents the index of the image block; Represents the calculated image block mean; Indicates the extracted watermark position; round() is a rounding function, and mod represents the remainder operation.

Images