A multi-image reversible hiding method and device based on up-sampling

Abstract

This invention discloses a reversible multi-image hiding method and apparatus based on upsampling. The method selects an appropriate upsampling rate to upsample the carrier image according to the size of the multiple secret images to be hidden, thereby obtaining a target image with sufficient hiding capacity. Common embedding algorithms are then used to hide the multiple images. This invention not only achieves simultaneous hiding of multiple secret images with high capacity but also improves the quality of the disguised image while ensuring lossless recovery of the multiple secret images.

Description

Technical Field

[0001] This invention relates to the field of information security technology, and in particular to a multi-image reversible hiding method and apparatus based on upsampling. Background Technology

[0002] Multi-image reversible hiding technology refers to the reversible embedding of one or more secret images into a carrier medium. "Lossless" means that the one or more secret images can be reconstructed without any loss of quality. This characteristic of reversible image hiding technology is of great significance in many fields such as military, medical, and copyright protection.

[0003] Common reversible image hiding techniques include color-transformation-based reversible algorithms and compressed sensing-based reversible algorithms. Compressed sensing-based reversible algorithms are currently the mainstream approach. Their core idea is to compress the secret image using compressed sensing technology and then embed it into the target image using general embedding methods. Compared to color-transformation-based reversible algorithms, this algorithm significantly improves embedding capacity and the visual quality of the disguised image. However, compressed sensing is a lossy compression technique; the compressed image cannot be recovered losslessly. This could potentially lead to misinterpretation of information due to the loss of quality in reconstructing the secret image. Summary of the Invention

[0004] The purpose of this invention is to provide a multi-image reversible hiding method and apparatus based on upsampling. This method utilizes upsampling technology to enlarge the carrier image, thereby increasing the embedding capacity of the target image. This not only improves the visual quality of the camouflaged image but also ensures that the secret image can be recovered without loss.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] This invention provides a multi-image reversible hiding method based on upsampling, comprising:

[0007] The upsampling rate is determined based on the secret image to be hidden and the carrier image to be used;

[0008] The target image is obtained by upsampling the carrier image to be used using the aforementioned upsampling rate;

[0009] The secret images to be hidden are merged to obtain the merged secret image;

[0010] The merged secret image is embedded into the target image to obtain a hidden image and generate additional information;

[0011] The additional information is embedded into the hidden image to obtain the disguised image;

[0012] The original secret image was recovered based on the camouflaged image.

[0013] Furthermore, the upsampling rate is determined based on the secret image to be hidden and the carrier image to be used, including:

[0014] Calculate the minimum upsampling rate based on the secret image to be hidden and the carrier image to be used:

[0015]

[0016] Where, m i ×n i , i∈[1,k] is the size of the i-th secret image to be hidden, k is the number of secret images, M×N is the size of the carrier image, and MSR is the minimum upsampling rate;

[0017] The upsampling rate is selected based on a balance between the visual quality of the camouflaged image and the bandwidth occupied by the transmission of the camouflaged image; the upsampling rate is greater than the minimum upsampling rate.

[0018] Furthermore, the target image is obtained by upsampling the carrier image to be used using the upsampling rate, including:

[0019] The carrier image to be used can be upsampled using the upsampling rate using any of the following methods:

[0020] Nearest neighbor interpolation, bilinear interpolation, and cubic convolution interpolation.

[0021] Furthermore, the process of merging the secret images to be hidden to obtain the merged secret image includes:

[0022] The secret images to be hidden are seamlessly stitched together horizontally to form a single secret image.

[0023] Furthermore, the merged secret image is embedded into the target image to obtain a hidden image, and additional information is generated, including:

[0024] The merged secret image is embedded into the target image using any reversible color conversion embedding method to obtain a hidden image and generate additional information.

[0025] Furthermore, the additional information is embedded into the hidden image to obtain a disguised image, including:

[0026] The additional information can be embedded into the hidden image using any reversible data hiding method.

[0027] Furthermore, the reversible data hiding method includes:

[0028] Multiple residual histogram translation algorithm and least significant bit embedding algorithm.

[0029] Furthermore, recovering the original secret image based on the camouflaged image includes:

[0030] Based on the selected reversible data hiding method, the corresponding data extraction method is selected to extract the additional information embedded in the disguised image;

[0031] The selected reversible color conversion embedding method is reversed to extract additional information embedded in the camouflaged image, thereby recovering the secret image hidden in the camouflaged image.

[0032] The recovered secret image is split to obtain the original secret image.

[0033] Another aspect of the present invention provides a multi-image reversible hiding device based on upsampling, which performs multi-image reversible hiding using the aforementioned multi-image reversible hiding method based on upsampling, including:

[0034] An initialization module is used to determine the upsampling rate based on the secret image to be hidden and the carrier image to be used;

[0035] The upsampling module is used to upsample the carrier image to be used using the upsampling rate to obtain the target image;

[0036] The merging module is used to merge the secret images to be hidden to obtain the merged secret image;

[0037] The first embedding module is used to embed the merged secret image into the target image to obtain a hidden image and generate additional information;

[0038] The second embedding module is used to embed the additional information into the hidden image to obtain a disguised image;

[0039] The recovery module is used to recover the original secret image based on the disguised image.

[0040] The beneficial effects of this invention are as follows:

[0041] This invention first determines the upsampling rate by comparing the sizes of multiple secret images and the carrier image. Then, it uses an upsampling interpolation algorithm to upsample and interpolate the pixels of the original image, thereby expanding the number of pixels and size of the original image, and increasing the embedding capacity to achieve simultaneous and reversible hiding of multiple images. This method not only allows for arbitrarily increasing the embedding capacity of the carrier image to accommodate simultaneous hiding of multiple images, but also achieves reversible recovery of multiple secret images while ensuring good visual quality of the camouflaged image. Attached Figure Description

[0042] Figure 1This is a flowchart of a multi-image reversible hiding method based on upsampling provided in Embodiment 1 of the present invention;

[0043] Figure 2 It is the secret image S1 in Embodiment 3 of the present invention;

[0044] Figure 3 It is the secret image S2 in Embodiment 3 of the present invention;

[0045] Figure 4 It is the carrier image F in Embodiment 3 of the present invention;

[0046] Figure 5 It is the upsampled carrier image L in Embodiment 3 of the present invention;

[0047] Figure 6 It is the camouflaged image C in Embodiment 3 of the present invention;

[0048] Figure 7 This is a map showing the distribution of related pixels in the horizontal direction of the carrier image in Embodiment 3 of the present invention;

[0049] Figure 8 This is a map showing the distribution of relevant pixels in the vertical direction of the carrier image in Embodiment 3 of the present invention;

[0050] Figure 9 This is a map showing the distribution of related pixels in the carrier image along the diagonal direction in Embodiment 3 of the present invention;

[0051] Figure 10 This is a map showing the distribution of relevant pixels in the horizontal direction of the target image in Embodiment 3 of the present invention;

[0052] Figure 11 This is a map showing the distribution of relevant pixels in the vertical direction of the target image in Embodiment 3 of the present invention;

[0053] Figure 12 This is a map showing the distribution of relevant pixels in the diagonal direction of the target image in Embodiment 3 of the present invention;

[0054] Figure 13 This is a map showing the horizontal pixel distribution of the camouflaged image in Embodiment 3 of the present invention;

[0055] Figure 14 This is a map showing the distribution of relevant pixels in the vertical direction of the camouflage image in Embodiment 3 of the present invention;

[0056] Figure 15 This is a pixel distribution map of the camouflaged image in the diagonal direction in Embodiment 3 of the present invention. Detailed Implementation

[0057] The present invention will now be further described. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.

[0058] Example 1

[0059] This embodiment provides a multi-image reversible hiding method based on upsampling. See [link to relevant documentation]. Figure 1 ,include:

[0060] The upsampling rate is determined based on the secret image to be hidden and the carrier image to be used;

[0061] The target image is obtained by upsampling the carrier image to be used using the determined upsampling rate;

[0062] The secret images to be hidden are merged to obtain the merged secret image;

[0063] The merged secret image is embedded into the target image to obtain the hidden image and generate additional information;

[0064] The obtained additional information is embedded into the hidden image to obtain the disguised image;

[0065] The original secret image was recovered from the disguised image.

[0066] Example 2

[0067] This embodiment provides a multi-image reversible hiding method based on upsampling, including the following steps:

[0068] S1. Determine the upsampling rate:

[0069] To ensure the carrier image has a sufficiently large embedding capacity to embed multiple secret images while minimizing the storage space occupied by the target image, a minimum upsampling rate is first calculated. The target image size after sampling at the minimum upsampling rate is identical to the sum of the sizes of the multiple secret images. However, in practical applications, a suitable upsampling rate greater than the minimum upsampling rate can be chosen to achieve a balance between the visual quality of the camouflaged image and the bandwidth occupied by its transmission.

[0070] The formula for calculating the minimum upsampling rate (MSR) is:

[0071]

[0072] In the formula, m i ×n i Let i∈[1,k] be the size of the i-th secret image, k be the number of secret images, M×N be the size of the carrier image, and MSR be the minimum upsampling rate.

[0073] The actual upsampling rate SR is generally greater than the minimum upsampling rate, expressed as: SR ≥ MSR. In practical applications, the upsampling rate is selected based on a balance between the visual quality of the camouflaged image and the bandwidth occupied by its transmission. A larger SR requires more transmission bandwidth and results in higher camouflaged image quality, while a smaller SR requires less transmission bandwidth and results in lower camouflaged image quality.

[0074] S2. Upsample the carrier image using a determined upsampling rate to obtain the target image;

[0075] In this step, based on the actual sampling rate determined in step S2, a suitable upsampling interpolation algorithm is selected to upsample the carrier image to obtain a target image (upsampled carrier image) of size M×N×SR.

[0076] Commonly used upsampling interpolation algorithms mainly include three methods: nearest neighbor interpolation, bilinear interpolation, and cubic convolution interpolation. This embodiment can use any one of these three methods to upsample the carrier image.

[0077] To facilitate the explanation of the three interpolation algorithms, we assume that the input image size is p×q and the output image size is m×n. Then, according to S1, the minimum upsampling rate can be calculated.

[0078] S2A: Upsampling of the carrier image using the nearest neighbor difference method is implemented as follows:

[0079] S2A1. Find the mapping from the position (x′, y′) of each pixel in the output image to the pixel position (x, y) in the input image. The mapping expression from (x′, y′) to (x, y) is:

[0080]

[0081] In the formula, round() is the rounding function, (x′,y′) is the position of the pixel in the output image, and (x,y) is the position of the pixel in the input image.

[0082] S2A2. Assign the pixel value corresponding to each pixel position (x, y) in the input image to the pixel value at the corresponding pixel position (x′, y′) in the output image. The pixel assignment expression is as follows:

[0083] g(x′,y′)=f(x,y);

[0084] In the formula, f(x,y) is the pixel value of pixel (x,y) in the input image, and g(x′,y′) is the pixel value of pixel (x′,y′) in the output image.

[0085] S2B uses bilinear interpolation to upsample the carrier image. Unlike nearest-neighbor interpolation, which directly selects the nearest pixel value as the interpolated pixel value, bilinear interpolation calculates the interpolated pixel by performing a linear transformation on the four nearest known pixel values. The specific implementation process is as follows:

[0086] S2B1. Find the mapping from the position (x′, y′) of each pixel in the output image to the pixel position (x, y) in the input image. The mapping expression from (x′, y′) to (x, y) is:

[0087]

[0088] In the formula, floor() is the floor function.

[0089] S2B2: Calculate the three nearest pixels around the pixel (x,y) obtained in step S2B1, which are (x+1,y), (x,y+1), and (x+1,y+1).

[0090] S2B3. Perform a bilinear transform on the four pixel values ​​f(x,y), f(x+1,y), f(x,y+1), and f(x+1,y+1) to obtain the pixel values ​​g(x′,y′) of the output image. The expression for the bilinear transform is:

[0091] g(x′,y′)=round((1-η)(1-λ)f(x,y)+η(1-λ)f(x+1,y)+(1-η)λf(x,y+1)+ηλf(x+1,y+1));

[0092] In the formula, η and λ are points The distance between the point (x, y) and the point (x, y) is...

[0093]

[0094] S2C uses cubic convolution interpolation to upsample the carrier image. Unlike nearest neighbor interpolation and bilinear interpolation, cubic convolution interpolation uses the best interpolation function S(x) on the sixteen nearest known pixel values ​​to calculate the interpolated pixels, instead of the four nearest known pixel values. This results in more accurate interpolated pixels. The specific implementation process is as follows:

[0095] The expression for S(x) is:

[0096]

[0097] In the formula, θ is a parameter, usually taken as θ = 0.5.

[0098] S2C1: Obtain the pixel (x,y) mapped to the input image from step S2A1;

[0099] S2C2: Calculate the 15 nearest pixels to the pixel (x, y) obtained in step S2B1. The matrix formed by these 16 pixel values ​​is as follows:

[0100]

[0101] S2C3. Perform a cubic convolution on the 16 pixel values ​​obtained in step S2C2 to obtain the output image pixel g(x′,y′). The cubic convolution expression is:

[0102] g(x′,y′) = ABC;

[0103] In the formula, A=[S(n+1)S(n+0)S(n-1)S(n-2)],

[0104] S3, Merging multiple secret images;

[0105] Specifically, k secret images S1, S2, ..., S k The horizontally seamlessly stitched together form a secret image S of size m×n, where m×n = m1×n1 + ... + m k ×n k .

[0106] It should be noted that multi-image stitching and merging is a well-known technique in the field, and will not be explained in detail here.

[0107] S4. Hide the merged secret image to obtain the hidden image;

[0108] In this step, the merged secret image S is embedded into the target image to obtain the hidden image.

[0109] It should be noted that any common reversible color transformation embedding method can be used to achieve hiding, which will not be explained in detail here. This step will generate additional information needed to recover the secret image.

[0110] S5. Embed additional information into the hidden image to obtain the disguised image;

[0111] It should be noted that common reversible data hiding methods can be used to hide the additional information generated in step S4, such as multiple residual histogram shifting and least significant bit embedding, which will not be explained in detail here.

[0112] S6. Extract additional information;

[0113] In this step, based on the reversible data hiding method selected in step S5, the corresponding data extraction method is selected to extract the additional information embedded in the disguised image.

[0114] S7. Recover multiple secret images hidden within the disguised images;

[0115] In this step, using the additional information obtained in S6, the reverse process of the image hiding method selected in S4 is implemented to recover the multiple secret images originally hidden in the disguised image.

[0116] S8. The recovered secret image is split to obtain multiple original secret images.

[0117] It should be noted that this embodiment is based on upsampling and reversible hiding of multiple images. For the same carrier image, the number and size of the secret images to be hidden are different, the upsampling rate is different, and the embedding capacity of the final target image is also different.

[0118] Example 3

[0119] This embodiment provides a multi-image reversible hiding method based on upsampling. Simulation is performed using MATLAB 2017 software. Two secret images, Van Gogh and Parrot, with sizes of 512×384 and 512×768 respectively, are selected. Figure 2 and Figure 3 As shown, the carrier image Lena, denoted as S1, S2, and with a size of 512×512, is as follows. Figure 4 As shown, it is denoted as F.

[0120] A multi-image reversible hiding method based on upsampling is used to hide the Parrot and Van Gogh images into the Lena image. The specific process is as follows:

[0121] (1) Determine the upsampling rate:

[0122] To ensure that the carrier image F has a sufficiently large embedding capacity to embed the two secret images S1 and S2, and to minimize the storage space occupied by the target image L, the minimum upsampling rate (MSR) is first calculated. The formula for calculating the minimum upsampling rate (MSR) is as follows:

[0123]

[0124] In the formula m i ×n i Let i∈[1,k] represent the sizes of the k secret images, M×N represent the size of the carrier image, and MSR represent the minimum upsampling rate.

[0125] Minimum upsampling rate in this embodiment Here we choose the minimum upsampling rate MSR = 2.25 as the actual sampling rate, i.e. SR = 2.25.

[0126] (2) Upsample the carrier image to obtain the target image:

[0127] Using the actual sampling rate SR = 2.25 determined in step (1), the nearest neighbor difference method is used to upsample the carrier image F to obtain a target image L of size M × N × SR = 512 × 512 × 2.25, as shown below. Figure 5 As shown.

[0128] Since the actual number of pixels in an image is too large and the specific interpolation process is too complex, we will now assume that 16 pixels of a 4×4 secret image are used to represent its specific nearest neighbor difference process.

[0129] Assuming the carrier image size is 4×4, its pixel value distribution is as follows: If the target image size is 8×8, then the minimum upsampling rate is...

[0130] (2.1) The position (x′, y′) of each pixel in the target image is mapped to the pixel position (x, y) in the carrier image. The mapping expression from (x′, y′) to (x, y) is:

[0131]

[0132] In the formula, round() is the rounding function, (x′,y′) is the position of the pixel in the target image, and (x,y) is the position of the pixel in the carrier image.

[0133] In this example, the mapping expression from (x′,y′) to (x,y) is:

[0134]

[0135] The matrix (x, y) obtained in this way is:

[0136]

[0137] (2.2) Assign the pixel value corresponding to each pixel position (x, y) in the carrier image to the pixel value at the corresponding pixel position (x′, y′) in the target image. The pixel assignment expression is as follows:

[0138] g(x′,y′)=f(x,y);

[0139] In the formula, f(x,y) is the pixel value of pixel (x,y) in the carrier image, and g(x′,y′) is the pixel value of pixel (x′,y′) in the target image.

[0140] In this embodiment, the distribution of target image pixel values ​​obtained by upsampling the carrier image using the nearest neighbor difference method is as follows:

[0141] (3) Merging multiple secret images:

[0142] In this embodiment, two secret images S1 and S2 are merged into a single merged secret image S.

[0143] (4) Hide the merged secret image:

[0144] The merged secret image S is embedded into the target image L to obtain the hidden image.

[0145] In this embodiment, a reversible color conversion method is used to achieve concealment, which mainly consists of two stages: block matching and color conversion. In the block matching stage, the merged secret image and the target image are first divided into 4×4 blocks. Then, the standard deviation of each block is calculated, and each block is sorted in ascending order according to the standard deviation. Finally, the sorted merged secret image blocks and target image blocks are matched sequentially to obtain the block matching index.

[0146] The color conversion stage includes two steps: color conversion and pixel value overflow handling. The color conversion formula is as follows:

[0147] Δu=round(u L -u S ),

[0148] p″ t =p t +Δu,

[0149] Where u L and u S p represents the average pixel value of the target image patch and the merged secret image patch to be matched, respectively. t p″ represents the pixel value of each pixel in the merged secret image block. t These are the pixel values ​​after color conversion.

[0150] The formula for handling pixel value overflow is as follows:

[0151]

[0152] OV max UN represents the maximum value of the overflow pixels. min This represents the minimum value of the overflow pixel.

[0153] This step generates the additional information needed to recover the secret image.

[0154] (5) Embed additional information into the hidden image to obtain the disguised image C, such as Figure 6 As shown,

[0155] In this embodiment, the additional information is embedded using the multiple residual histogram translation method. This method is well-known and will not be explained in detail here.

[0156] (6) Extract additional information:

[0157] Based on the reversible data hiding method selected in step (5), the corresponding data extraction method is selected to extract the additional information embedded in the camouflaged image C.

[0158] (7) Recover multiple secret images hidden within the disguised images;

[0159] Using the additional information obtained in step (6), the reverse process of the image hiding method selected in step (4) is implemented to recover the secret image S originally hidden in the disguised image.

[0160] (8) Split the secret image S to obtain two secret images S1 and S2.

[0161] The performance of the upsampling-based multi-image reversible hiding method in this embodiment is analyzed below.

[0162] I. Analysis of the quality of camouflaged images

[0163] Currently, the most widely used image quality evaluation metrics are Peak Signal-to-Noise Ratio (PSNR) and Mean Structural Similarity (MSSIM). A higher PSNR value and an MSSIM closer to 1 indicate better visual quality. The calculation formulas are as follows:

[0164]

[0165]

[0166]

[0167]

[0168] With Van Gogh (512×384) and Parrot (512×768) selected as the secret images and Lena (512×512) as the carrier image, and an upsampling rate of 4, the quality of the camouflage images is shown in Table 1 below.

[0169] Table 1. Camouflage Image Quality

[0170]

[0171] As can be seen from Table 1, the upsampled carrier image in this embodiment has an extremely high embedding capacity, and can be hidden even if the secret image is larger than the carrier image, and the disguised image has good quality.

[0172] II. Correlation Analysis of Adjacent Pixels

[0173] The correlation between adjacent pixels is used to analyze the relationship between adjacent pixel values. For a good upsampling method, the sampled carrier image (target image) should have high pixel correlation. Here, the correlation coefficient is used to measure the strength of the correlation between pixels; the closer the correlation coefficient is to 1, the stronger the correlation between pixels. The calculation formula is:

[0174] The correlation coefficients of the carrier image, target image, and camouflage image in the horizontal, vertical, and diagonal directions are shown in Table 2 below. The pixel distributions of the carrier image, target image, and camouflage image are shown below. Figure 7-9 , Figure 10-12 , Figure 13-15 As shown.

[0175] In Table 2, the correlation coefficients of the target image and the camouflage image in all three directions are larger and closer to 1 than the correlation coefficient of the carrier image. Figure 7-15 The compact distribution of pixel values ​​indicates that the target image and the camouflage image have a high pixel correlation, further proving that the scheme in this embodiment has a good concealment effect.

[0176] Table 2 Correlation coefficients

[0177]

[0178] Example 4

[0179] This embodiment provides a multi-image reversible hiding device based on upsampling, which performs multi-image reversible hiding using the upsampling-based multi-image reversible hiding method of Embodiment 1 or Embodiment 2, including:

[0180] An initialization module is used to determine the upsampling rate based on the secret image to be hidden and the carrier image to be used;

[0181] The upsampling module is used to upsample the carrier image to be used using the upsampling rate to obtain the target image;

[0182] The merging module is used to merge the secret images to be hidden to obtain the merged secret image;

[0183] The first embedding module is used to embed the merged secret image into the target image to obtain a hidden image and generate additional information;

[0184] The second embedding module is used to embed the additional information into the hidden image to obtain a disguised image;

[0185] The recovery module is used to recover the original secret image based on the disguised image.

[0186] It is worth noting that this device embodiment corresponds to the above method embodiment. The implementation methods of the above method embodiments are all applicable to this device embodiment and can achieve the same or similar technical effects, so they will not be described in detail here.

[0187] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0188] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0189] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0190] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0191] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A multi-image reversible hiding method based on upsampling, characterized in that, include: The upsampling rate is determined based on the secret image to be hidden and the carrier image to be used, including: Calculate the minimum upsampling rate based on the secret image to be hidden and the carrier image to be used: ； in, , For the first The size of the secret image to be hidden. The number of secret images. The size of the carrier image is MSR, where MSR is the minimum upsampling rate. The upsampling rate is selected based on a balance between the visual quality of the camouflaged image and the bandwidth occupied by the transmission of the camouflaged image; the upsampling rate is greater than the minimum upsampling rate. The target image is obtained by upsampling the carrier image to be used using the aforementioned upsampling rate; The secret images to be hidden are merged to obtain the merged secret image; The merged secret image is embedded into the target image to obtain a hidden image and generate additional information, including: using any reversible color conversion embedding method to embed the merged secret image into the target image to obtain a hidden image and generate additional information; The additional information is embedded into the hidden image to obtain the disguised image; The original secret image was recovered based on the camouflaged image.

2. The multi-image reversible hiding method based on upsampling according to claim 1, characterized in that, The upsampling rate is used to upsample the carrier image to be used to obtain the target image, including: The carrier image to be used can be upsampled using the upsampling rate using any of the following methods: Nearest neighbor interpolation, bilinear interpolation, and cubic convolution interpolation.

3. The multi-image reversible hiding method based on upsampling according to claim 1, characterized in that, The step of merging the secret images to be hidden to obtain the merged secret image includes: The secret images to be hidden are seamlessly stitched together horizontally to form a single secret image.

4. The multi-image reversible hiding method based on upsampling according to claim 1, characterized in that, The additional information is embedded into the hidden image to obtain a disguised image, including: The additional information can be embedded into the hidden image using any reversible data hiding method.

5. The multi-image reversible hiding method based on upsampling according to claim 4, characterized in that, The reversible data hiding method includes: Multiple residual histogram translation algorithm and least significant bit embedding algorithm.

6. The multi-image reversible hiding method based on upsampling according to claim 4, characterized in that, Recovering the original secret image based on the camouflaged image includes: Based on the selected reversible data hiding method, the corresponding data extraction method is selected to extract the additional information embedded in the disguised image; The selected reversible color conversion embedding method is reversed to extract additional information embedded in the camouflaged image, thereby recovering the secret image hidden in the camouflaged image. The recovered secret image is split to obtain the original secret image.

7. A multi-image reversible hiding device based on upsampling, characterized in that, Multi-image reversible hiding is performed using the upsampling-based multi-image reversible hiding method according to any one of claims 1 to 6, including: An initialization module is used to determine the upsampling rate based on the secret image to be hidden and the carrier image to be used; The upsampling module is used to upsample the carrier image to be used using the upsampling rate to obtain the target image; The merging module is used to merge the secret images to be hidden to obtain the merged secret image; The first embedding module is used to embed the merged secret image into the target image to obtain a hidden image and generate additional information; The second embedding module is used to embed the additional information into the hidden image to obtain a disguised image; The recovery module is used to recover the original secret image based on the disguised image.

Images