CN102938841B - Method for hiding information in bearing image, image quality evaluation method and information transmission method - Google Patents

Abstract

The invention discloses a method for hiding information in a bearer image, which includes an information hiding step at the image sending end and an information extraction step at the image receiving end. The information hiding step includes: a step of preprocessing the bearer image; The step of carrying the hidden information bits of the image; the information extraction step adopts the reverse process of the information hiding step to realize the extraction of the information bits. At the same time, the invention also provides an image quality evaluation method and an information transmission method realized by the information hiding method. The method improves the anti-compression performance of the hidden information in the carrying image, further increases the amount of information that can be hidden, and realizes the objective evaluation of the image quality and the confidential transmission of information.

Description

Hiding information in bearer image, image quality evaluation and information transmission method

technical field

The invention relates to a method for hiding information in a bearer image, and also relates to an image quality evaluation method using the hiding method and an information transmission method using the image using the hiding method.

Background technique

With the development of science and technology, image resolution is getting higher and higher, and the amount of transmitted data is increasing. Generally, data compression technology is used for transmission. The status of spacecraft image compression and data transmission systems is more prominent. The image data transmission The performance evaluation of the system is more and more important.

At present, most image compression algorithms use lossy compression methods to remove visual redundant information, and sacrifice part of the information of the source after compression. The image quality compressed by image compression algorithm and compression equipment directly reflects the performance of the algorithm and equipment. Therefore, image quality evaluation has become a problem worthy of attention. At the same time, this method can also be applied to enhance the image quality, improve the error tolerance of image encoding and decoding algorithms and so on.

Image quality evaluation can be divided into subjective evaluation methods and objective evaluation methods in terms of methods. The former evaluates the image quality of objects based on the subjective perception of the perceiver. Current research focuses on the latter, which measures image quality based on quantitative indicators or parameters given by the model.

The peak signal-to-noise ratio is based on the statistical and average calculation of the gray value of the image pixel, which is a commonly used indicator to measure signal distortion. Although there may be a large deviation from the subjectively perceived quality when evaluating the quality of some images or videos, PSNR is still effective for most image quality evaluations. More than 10 quality evaluation models have been evaluated internationally, but no one model has been determined as the standard model. The main reason is that from a statistical point of view, the results achieved by these models are not significantly different from PSNR. Therefore, it is generally believed that PSNR is still a meaningful reference index for image quality evaluation at home and abroad. However, this method needs the original image for evaluation, and there is no original image in the actual image digital transmission system, so this method cannot be used for evaluation.

Traditional image compression quality evaluation methods are mainly used in the design stage and test stage, that is, when the original image can be obtained. In the absence of original images, for example, when the image acquired by the spacecraft is transmitted to the ground through the spacecraft data compression transmission system, it cannot be used, because the ground cannot obtain the original image, and can only obtain the decompressed and restored image after channel transmission. image. Therefore, when evaluating the quality of image compression on the ground, it is mainly subjective evaluation, or objective evaluation of the characteristic parameters of the received image itself, which can neither be compared with the original image nor obtain objective evaluation indicators (such as peak signal-to-noise ratio (PSNR ).

Application number: CN200810191796.8, name: A method for evaluating the compression quality of space-borne optical remote sensing images, which generates simulated images according to the characteristics of satellite sensors, and adopts a comprehensive, systematic and comprehensive method of objective evaluation and subjective evaluation. Evaluate image compression quality in a targeted manner, and construct a complete system for evaluation of image compression quality in spaceborne optical remote sensing. However, this evaluation method can only evaluate the received image itself subjectively and objectively, and cannot directly compare and evaluate the characteristics of the original image and the received image.

Application number: CN201210178979.2, this invention discloses an image transmission quality evaluation method, which mainly solves the problem that the traditional evaluation method cannot realize the comparison of the characteristic parameters of two images, and greatly improves the image transmission quality of the image data transmission system The objectivity of evaluation avoids the problems that may be caused by subjective evaluation alone. However, this method hides the characteristic parameters in the original image for uncompressed transmission, or hides the characteristic parameters in the compressed data of the original image for transmission, but the receiving end cannot obtain the peak signal-to-noise between the restored image and the original image Compare.

Contents of the invention

The problem solved by the technology of the present invention is: on the one hand, it provides a method for hiding information in the bearer image, which improves the anti-compression performance of the hidden information in the bearer image, and further increases the amount of information that can be hidden. On the other hand, an image quality evaluation method is also provided, which can hide the index parameters of the image in the image, and realize the image quality evaluation by comparing the index parameters before and after image transmission at the image receiving end. At the same time, it also provides an information transmission method using images, which can realize hidden transmission of information in images.

Technical solution of the present invention is:

One aspect of the present invention discloses a method for hiding information in a bearer image, including an information hiding step at an image sending end and an information extraction step at an image receiving end:

The information hiding steps include:

Host image preprocessing:

1) Dividing the bearer image P with a size of M×N into zigzag scanning order in rows or columns into non-overlapping image blocks Pi with a size of m×n, where M, N, m, n are Even number, i=1, 2, ..., L, L=(M×N)/(m×n);

2) Perform modulo two operation on the abscissa and ordinate of each pixel position coordinate in the image block Pi, if the results of the modulo two operation on the abscissa and ordinate are equal, then the pixel at the position coordinate is classified as set A , otherwise the pixels of the position coordinates are classified as set B;

3) For each image block Pi, use the difference between the sum of the pixel values in the set A and the sum of the pixel values in the set B as the statistic Tx of the image block Pi, and calculate For the statistics Tx of all image blocks Pi, the absolute value of the maximum value in the obtained statistics is taken as the maximum statistical absolute value Tmax;

4) Adjust the pixel value of each image block Pi:

If the statistic Tx of the image block is greater than or equal to -T1 and less than or equal to T1, the pixel value of the image block remains unchanged, wherein T1 is a positive integer value smaller than Tmax;

If the statistic Tx of the image block is greater than T1, then a positive integer K1 is added to the pixel in the set A of the image block, and a positive integer K1 is subtracted from the pixel in the set B, wherein K1=(T1+ R1)/(m×n), where R1 is a positive integer;

If the statistic Tx of the image block is less than -T1, subtract a positive integer K1 from the pixels in the set A of the image block, and add a positive integer K1 to the pixels in the set B;

Sequentially adjust each of the image blocks Pi and recalculate the statistic Ty respectively to complete the image preprocessing step;

Information bit hiding:

The 0-bit value in the binary sequence of the information to be hidden is hidden in the preprocessed image block Pi using the following process:

If the statistic Ty of the image block Pi is within the range (-T1, T1), keep the pixel value of the image block unchanged;

If the statistic Ty of the image block Pi is greater than T1, then add a positive integer K2 to the pixels in the set A of the image blocks, and subtract a positive integer K2 from the pixels in the set B of the image blocks;

If the Ty is less than -T1, then subtract a positive integer K2 from the pixels in the set A of the image blocks, and add a positive integer K2 to the pixels in the set B of the image blocks; wherein, K2=( T2+R2)/(m×n), T2=Tmax-T1 is a positive integer, R2 is a positive integer;

The 1-bit value in the binary sequence of the information to be hidden is hidden into the preprocessed image block Pi using the following process:

If the statistic Ty of the image block Pi is greater than the threshold T1 or less than -T1, then keep the pixel value of the image block unchanged;

If the statistic Ty of the image block Pi is greater than or equal to 0 and less than or equal to T1, then add a positive integer K1 to the pixels in the set A of the image block, and subtract a positive integer from the pixels in the set B K1;

If the statistic Ty of the image block Pi is greater than or equal to -T1 and less than 0, then subtract a positive integer K1 from the pixels in the set A of the image block, and add a positive integer K1 to the pixels in the set B;

Information bit extraction:

Receive the bearer image and re-divide the received bearer image into m×n non-overlapping image blocks Wi, i=1, 2, . . . , L, L=(M×N)/( m×n);

According to the step 2) obtain the set C when the operating results of the second module are equal and the set D when the operating results of the second module are not equal;

Obtain the statistic Tz of each image block Wi according to the step 3);

Extracting a 0-bit value and a 1-bit value according to the distribution interval of the statistic Tz;

Recover the binary sequence of the hidden information according to the extracted 0-bit value and 1-bit value.

Further, in the above method for hiding information, the hidden information is an index parameter of the bearer image, and the index parameter includes the characteristic parameter and/or peak signal-to-noise ratio of the bearer image, and the characteristic parameter includes the bearer image The mean, variance and entropy of the image.

Another aspect of the present invention discloses an image quality evaluation method using the above hidden information method, comprising the following steps:

Send the original image with the index parameters hidden to the image receiving end;

Extracting the index parameters hidden in the original image at the image receiving end, and recalculating the index parameters of the original image at the image receiving end;

The evaluation of the original image at the image receiving end is completed by comparing the extracted index parameter with the recalculated index parameter.

Further, in the above image quality evaluation method, the original image with index parameters hidden is sent to the image receiving end after being compressed.

The present invention also provides an information transmission method using an image, and the information is hidden and extracted in the image by the above information hiding method.

Compared with the prior art, the present invention has the following advantages:

According to the distribution characteristics of the image pixel values, the method for hiding information in the bearer image according to the present invention can hide the information bits in the bearer image through the preprocessing of the bearer image and the information bit hiding steps. This method hides the information bits The carrying image will not affect the quality of the image, and it overcomes the fact that the information hiding algorithms based on DCT transform and wavelet transform in the prior art have a small hiding capacity, and the performance of the algorithm decreases rapidly with the increase of the hiding capacity. . At the same time, the embedded information will not cause irreversible damage to the bearer image, and the quality of the bearer image at the image receiving end is guaranteed.

The image quality evaluation method described in the present invention can hide the characteristic parameters of the image into the bearer image and extract them at the receiving end, thereby realizing the quality evaluation of the bearer image by using parameters such as PSNR at the receiving end, overcoming the existing technology Insufficiency of only being able to evaluate the subjective quality of the image at the receiving end. And save the additional overhead of separately transmitting image feature parameters at both ends of the image sending and receiving,

At the same time, the information transmission method using images in the present invention can also use information bits as confidential information to be transmitted, and use the method of hiding information as the encoding of information bits using the gray value of the image, which is similar to the prior art. Compared with that, the quality of the bearer image is not affected, and the bearer image after hiding the information bits has good anti-compression performance, and can effectively extract the hidden information bits.

Description of drawings

Fig. 1 is a flowchart of the present invention;

Fig. 2 is an image block pixel set division diagram;

Fig. 3 is the statistic distribution of the Lena image whose block size is 8×8;

Figure 4 is the Lena image statistics shift with a block size of 8×8;

Fig. 5 is the distribution of Lena image statistics after embedding information bits;

Fig. 6 shows the statistical distribution of the dense carrier image compressed to a certain extent.

Detailed ways

The present invention will be further introduced below in conjunction with the accompanying drawings.

As shown in Figure 1, the method for hiding information in the bearer image includes the following steps:

First, carry out image preprocessing, the preprocessing process is as follows:

1). Divide the bearing image P with a size of M×N into non-overlapping image blocks Pi with a size of m×n in a zigzag scanning order by row or column, where M, N, m, and n are even numbers , i=1, 2,..., L, L=(M×N)/(m×n);

2). The abscissa and ordinate of the position coordinates of each pixel in the image block Pi are respectively subjected to a modulo two operation, and if the results of the modulo two operations of the abscissa and ordinate are equal, then the pixels of the position coordinates are classified as a set A , otherwise the pixels at the position coordinates are classified as set B. As shown in FIG. 2 , an 8×8 image block is taken as an example. After the modulo two operation is performed on the pixels in the image block, the markings "+" and "-" at the pixel positions indicate that the pixels at the pixel positions belong to the set A and the set B respectively.

3). The difference between the sum of the pixel values in the set A of each image block Pi and the sum of the pixel values in the set B is used as the statistic Tx of the image block Pi, and the statistic Tx can be calculated by the following formula:

$\mathrm{<span\; class="notranslate">\; Tx</span>}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; \&times;</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; a</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; b</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>$

Among them, a _i , i=1, 2,..., m×n/2 corresponds to the pixel whose pixel position is marked as "+" in the image block pixel set division diagram, b _i , i=1, 2,... , m×n/2 corresponds to the pixel marked as "-" in the pixel position in the difference pair map.

According to the above formula, the statistic Tx of each image block Pi is calculated sequentially, and the maximum value of the absolute value of the statistic is Tmax. Taking Figure 3 as an example, it can be seen that the statistic distribution of the Lena grayscale image with a size of 512×512 and each pixel represented by 8 bits is obtained through this step.

Since all the pixels in the image block add or subtract a fixed number, the statistics of the image block remain unchanged, even if the pixels in the image block change to varying degrees, the statistics of the image block also change very little, so the statistics The information bits are embedded as a robust parameter, and the embedded information bits can resist non-malicious attacks such as JPEG2000 lossy compression.

It can be seen from Figure 3 that most of the image statistics are concentrated around 0 because the gray values of general image pixels have strong local correlation and redundancy. This feature can be used to shift the statistics left and right to generate redundant space to embed secret information.

4). Adjust the pixel value of each image block Pi

If the value of the Pi statistic Tx of the image block is within the range of the threshold T1, ie -T1≤Tx≤T1, the pixel value of the image block remains unchanged.

If the Pi statistic Tx of the image block is greater than the threshold T1, a positive integer K1 is added to the pixels belonging to the set A in the image block, and a positive integer K1 is subtracted from the pixels belonging to the set B in the image block.

If the image block Pi statistic Tx is less than the threshold-T1, subtract a positive integer K1 from the pixels belonging to the set A in the image block, and add a positive integer K1 to the pixels belonging to the set B in the image block .

After adjusting each image block Pi in turn, recalculate the statistic Ty of each image block to complete the preprocessing of the loaded image.

Above-mentioned step 4) specific process is shown in the following formula:

The pixels belonging to the set A in the image block Pi are processed as follows:

$\mathrm{<span\; class="notranslate">\; x</span>}\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{ccc}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; K</span>}\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; if</span>}& \mathrm{<span\; class="notranslate">\; Tx</span>}<span\; class="notranslate">\; ></span>\mathrm{<span\; class="notranslate">\; T</span>}\mathrm{<span\; class="notranslate">\; 1</span>}\\ \mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; K</span>}\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; if</span>}& \mathrm{<span\; class="notranslate">\; Tx</span>}<span\; class="notranslate">\; <</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; T</span>}\mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right.$

The pixels belonging to the set B in the image block Pi are processed as follows:

$\mathrm{<span\; class="notranslate">\; x</span>}\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{ccc}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; K</span>}\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; if</span>}& \mathrm{<span\; class="notranslate">\; Tx</span>}<span\; class="notranslate">\; ></span>\mathrm{<span\; class="notranslate">\; T</span>}\mathrm{<span\; class="notranslate">\; 1</span>}\\ \mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; K</span>}\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; if</span>}& \mathrm{<span\; class="notranslate">\; Tx</span>}<span\; class="notranslate">\; <</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; T</span>}\mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right.$

Wherein, T1 is a selected threshold, T1 is a positive integer, and T1<Tmax. Since the statistic of the image block may have a larger value (as shown in FIG. 3 ), that is, Tmax may be larger. If the information bits are embedded directly by shifting the statistics of the image block left and right, the translation of the statistics will be too large due to the limitation of the maximum value Tmax of the statistics. That is, after the information bits are embedded, the pixel values of the image block change greatly, resulting in poor image quality after the information bits are embedded. By introducing the threshold T1 in the present invention, the change amount of most image block pixel values is reduced, and the performance of the algorithm is effectively improved. In the above process, K1 is a positive integer, K1=(T1+R1)/(m×n), generally 1≤K1≤10. Among them, R1 is a positive integer, and the size of R1 determines the robust performance of the algorithm. According to the above formula, taking the Lena image as an example, each image block Pi after the Lena image is divided into 8X8 pixel blocks is preprocessed by modifying the corresponding pixel value, and then the statistics of each image block are calculated to obtain the statistics Ty distribution, as shown in Figure 4.

After the preprocessing of the bearer image is completed, information bits in the form of a binary sequence may be hidden in the preprocessed bearer image.

For the 0-bit value in the information bit, if the statistic Ty of the image block Pi is in the range (-T1, T1), then keep the pixel value of the image block unchanged; if Ty is greater than the threshold T1, then the A positive integer K2 is added to the pixels belonging to the set A in the image block, and a positive integer K2 is subtracted from the pixels belonging to the set B in the image block; if Ty is less than the threshold -T1, the pixels in the set A are A positive integer K2 is subtracted from the pixels, and a positive integer K2 is added to the pixels in the set B; wherein, K2=(T2+R2)/(m×n). T2=Tmax-T1 is a positive integer, generally 1≤K2≤10. R2 is a positive integer, and the size of R2 also determines the robust performance of the algorithm.

For the 1-bit value in the information bit, if the statistic Ty carrying the image block Pi is greater than the threshold T1 or less than the threshold-T1, then keep the pixel value of the image block unchanged; if Ty is greater than or equal to 0 and less than or equal to The threshold T1, then add a positive integer K1 to the pixels belonging to the set A in the image block, and subtract a positive integer K1 from the pixels belonging to the set B in the image block; if Ty is greater than or equal to the threshold -T1 and less than 0, then subtract a positive integer K1 from the pixels in the set A, and add a positive integer K1 to the pixels in the set B;

According to the above process, the information bit concealment is completed by modifying the pixel value of each image block Pi in turn, and the image block Wi, i=1, 2, . . . , L containing information bits is obtained. The binary sequence of the information to be hidden is in one-to-one correspondence with the preprocessed image blocks Pi, and each image block Pi hides one bit. Finally, the image block Wi is reorganized according to the corresponding image block scanning order in step 1) to obtain the bearer image W with a size of M×N hidden secret information bits. Also taking the Lena image as an example, the statistics distribution of the Lena image after embedding information bits is shown in FIG. 5 . It can be seen from Figure 5 that after embedding the index parameters, the statistics of the Lena image are divided into five areas, where the "bit 0 area" and "bit 1 area" represent the embedded index parameter bits are "0" and "1" respectively. The blank areas between bit fields are robust intervals.

After the bearer image W with information bits hidden is sent to the image receiving end, the information bits are extracted at the image receiving end.

When extracting the information bits, according to the number of 0 and 1 when embedding, and considering the trend that 0 is mainly concentrated in the bit 0 area and 1 is mainly concentrated in the bit 1 area, the statistic distribution of the bearing image W containing information bits is redistributed. partition to extract information bits.

Figure 6 shows the distribution of statistics when the bearer image W containing information bits is compressed to a certain extent.

As can be seen from Figure 6, if the statistic is in the interval [-x, x] or [-z, -y] or [y, z], then extract the bit "0"; if the statistic is in the interval [-y, -x ] or [x, y], extract bit "1";

Each image block produces a statistic, and each statistic can represent a binary bit. Therefore, the total number of embedded bits is equal to the number of image blocks. That is to say, the embedding capacity Cap=(M×N)/(m×n)bits, and the number of secret information bits is 1/(m×n×8) of the total number of bits of the carrier image. When M=N=512, m=n=8, Cap=(512*512)/(8*8)=4096 bits can be embedded, and the number of information bits is 1/512 of the total number of bits of the image.

Embodiment one

The above-mentioned information bits can be a combination of one or several parameters of the peak signal-to-noise ratio PSNR, mean value, variance, and entropy of the bearer image, and these parameters about the bearer image can be hidden in the bearer image by the above method, and sent At the image receiving end, by recovering the parameters at the image receiving end, the evaluation of the quality of the bearer image can be realized. Therefore, in the embodiment of the present invention, an image quality evaluation method using the above method is further provided. In this method, the above information bits are the index parameters of the bearing image, and the image quality can be evaluated by using the index parameters before and after transmission. evaluate.

Index parameters include one or two of peak signal-to-noise ratio and characteristic parameters.

Suppose the original image at the image sending end is X (size M*N*8), the gray value is X(i, j); the original image restored by the image receiving end is Y (size M*N*8), gray value The degree value is Y(i,j).

The mean square error of X and Y:, $\mathrm{MSE}=\frac{1}{\mathrm{MN}}\underset{i=1}{\overset{m}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{N}{\mathrm{\&Sigma;}}}{(x(i,j)-Y(i,j))}^2$

Peak Signal to Noise Ratio: $\mathrm{PSNR}=10.\log \frac{{255}^2}{\mathrm{MSE}}\left(\mathrm{dB}\right),$

PSNR can be set according to system requirements, generally above 30dB.

image mean $\mathrm{EX}=\frac{1}{\mathrm{MN}}\underset{i=1}{\overset{m}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{N}{\mathrm{\&Sigma;}}}\left(x(i,j)\right)$

image variance ${\mathrm{\&sigma;}}^2=\frac{1}{\mathrm{MN}}\underset{i=1}{\overset{m}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{N}{\mathrm{\&Sigma;}}}{(x(i,j)-\mathrm{EX})}^2$

The calculation of the entropy of the image is known to those skilled in the art.

After the indicator parameters are hidden in the original image using the method of hiding information above, it is transmitted to the image receiving end.

The image receiving end further extracts the index parameters before image transmission from the received image according to the above information bit extraction method. And further calculate the index parameter of the received image at the image receiving end. And compare the index parameters before and after the transmission one by one to complete the evaluation of image quality.

Before transmitting the image with hidden index parameters, the image can also be compressed, and the index parameters can be extracted after the image is decompressed at the image receiving end. After the information bits are hidden in the carrying image by adopting the method described in the embodiment of the present invention, no matter whether all pixels in the image block are added or subtracted by a fixed number, the statistics of the image block remain unchanged, even if the pixels in the image block are different As the degree changes, the statistics of the image block also change very little. The invention utilizes the robustness of the statistics of the image blocks to embed the information bits, and after the image blocks are subjected to non-malicious attacks such as JPEG2000 lossy compression, the changes in the statistics of the image blocks are small, and the hidden information bits can still be correctly extracted. Therefore, the present invention has good compression resistance.

Embodiment two

The process of hiding the information bits into the bearer image by using the present invention can be further regarded as the encoding process of the bearer image at the image sending end, and the encoding of each information bit is realized by adjusting the gray value of the bearer image, and At the receiving end, the corresponding decoding method is used to realize the extraction of information bits, which can realize the confidential transmission of information bits by using the bearer image.

Therefore, in this embodiment, an information hiding step is used at the image sending end to hide the information bits to be encrypted in the bearer image, and an information extraction step is used at the image receiving end to extract the information bits, so that the information described in this embodiment can be realized. Secure transmission of bits.

It is used to change the pixel value of each image block in the carrying image when the information hiding step is used to hide the information bits, which has little influence on the quality of the carrying image. The invention divides the pixels in the image block into sets, and completes the embedding of the information bits by adding and subtracting a small positive integer to the pixels in the two sets respectively. Therefore, after the information bits are embedded in the present invention, the amount of change to the pixels of the image block is very small, and the change amount is evenly distributed, and the influence on the image quality is small. In addition, the information bit hiding capacity of the present invention Cap=(M×N)/(m×n)bits, when the image block size m and n decrease, the hiding amount increases, and the hiding capacity can be adjusted according to performance requirements in practical applications, Algorithms are very practical.

Parts not described in detail in the present invention belong to the common knowledge of those skilled in the art.

Claims (5)

1. the method hidden Info in load image, comprises the Information hiding step of image transmitting terminal and the information extracting step of image-receptive end, it is characterized in that:

Described Information hiding step comprises:

Load image preliminary treatment:

1) by size be the load image P of M × N by row or column image block Pi for the non-overlapping copies of m × n sized by Z-shaped scanning sequency is divided, wherein, M, N, m, n are even number, i=1,2 ..., L, L=(M × N)/(m × n);

2) abscissa of location of pixels coordinate each in described image block Pi and ordinate are carried out mould two computing respectively, if mould two operation result of abscissa and ordinate is equal, then the pixel of this position coordinates is classified as set A, otherwise the pixel of this position coordinates is classified as set B;

3) for each described image block Pi, using the statistic T x of the difference of the pixel value sum in described set A and the pixel value sum in described set B as image block Pi described in this, and calculate the statistic T x of all image block Pi respectively, using the absolute value of maximum in obtained statistic as maximum statistics absolute value Tmax;

4) pixel value of each described image block Pi is adjusted:

If the statistic T x of described image block is more than or equal to-T1 and is less than or equal to T1, then the pixel value of described image block remains unchanged, and wherein, T1 is the positive integer value being less than Tmax;

If the statistic T x of described image block is greater than T1, then the pixel in the set A of image block described in this is added a positive integer K1, the pixel in set B deducts a positive integer K1, wherein, K1=(T1+R1)/(m × n), wherein, R1 is positive integer;

If the statistic T x of described image block is less than-T1, then the pixel in the set A of image block described in this is deducted a positive integer K1, the pixel in set B adds a positive integer K1;

Adjust each described image block Pi successively and recalculate respectively and obtain statistic T y, complete described Image semantic classification step;

Information bit is hidden:

By 0 bit value in the binary sequence of information to be concealed, following process is adopted to be hidden in described pretreated image block Pi:

If the statistic T y of described image block Pi is in scope (-T1, T1), then keep the pixel value of this image block constant;

If the statistic T y of described image block Pi is greater than T1, then the pixel in the set A of described image block is added a positive integer K2, and the pixel in the set B of described image block is deducted a positive integer K2;

If described Ty is less than-T1, then the pixel in the set A of described image block is deducted a positive integer K2, and the pixel in the set B of described image block is added a positive integer K2; Wherein, K2=(T2+R2)/(m × n), T2=Tmax-T1 are positive integer, and R2 is positive integer;

By 1 bit value in the binary sequence of information to be concealed, following process is adopted to be hidden in described pretreated image block Pi:

If the statistic T y of described image block Pi is greater than threshold value T1 or is less than-T1, then keep the pixel value of image block constant;

If the statistic T y of described image block Pi is more than or equal to 0 and is less than or equal to T1, then the pixel in the set A of described image block is added a positive integer K1, and the pixel in described set B is deducted a positive integer K1;

If the statistic T y of described image block Pi is more than or equal to-T1 and is less than 0, then the pixel in the set A of described image block is deducted a positive integer K1, and the pixel in set B is added a positive integer K1;

Complete hiding of information bit by the pixel value revising each image block Pi successively by above process, obtain the image block Wi containing information bit, i=1,2 ..., L; The binary sequence of information to be concealed and pretreated image block Pi one_to_one corresponding, each image block Pi hides a bit;

Information bit extracts:

Receive load image and by step 1) load image received is reclassified as the image block Wi of the non-overlapping copies of m × n, i=1,2 ..., L, L=(M × N)/(m × n);

According to described step 2) obtain described mould two operation result equal time set C and described mould two operation result anisochrouous set D;

According to described step 3) obtain the statistic T z of each described image block Wi;

0 bit value and 1 bit value is extracted according to the distributed area of described statistic T z;

The binary sequence hidden Info described in recovering according to described 0 bit value that extracts and 1 bit value.

2. the method hidden Info in load image as claimed in claim 1, it is characterized in that, describedly hide Info as the index parameter of described load image, described index parameter comprises characteristic parameter and/or the Y-PSNR of described load image, and described characteristic parameter comprises the average of load image, variance and entropy.

3. adopt the method hidden Info in load image according to claim 2, it is characterized in that comprising the following steps:

The original image being concealed with described index parameter is sent to image-receptive end;

The described index parameter hidden in described original image is extracted at image-receptive end, and the index parameter of computed image receiving terminal original image again;

Utilize the index parameter of described extraction and the described index parameter recalculated to contrast, complete in the evaluation of image-receptive end to original image.

4. the method hidden Info in load image as claimed in claim 3, is characterized in that: described in be concealed with index parameter original image after overcompression, be sent to image-receptive end.

5. utilize an information transferring method for image, it is characterized in that, described information adopts method described in claim 1 to carry out in the picture hiding and extracting.