WO2014075469A1

WO2014075469A1 - Method and device for information hiding

Info

Publication number: WO2014075469A1
Application number: PCT/CN2013/081235
Authority: WO
Inventors: 贺菲菲; 胡森; 潘志斌; 邓小曼; 王凌飞
Original assignee: 中兴通讯股份有限公司
Priority date: 2012-11-13
Filing date: 2013-08-09
Publication date: 2014-05-22
Also published as: CN103810664A; CN103810664B

Abstract

Disclosed is a method for information hiding, comprising: encrypting an original image, dividing the encrypted image into image blocks, dividing each image block into two or more parts, embedding secret information into corresponding parts of the image blocks, and issuing the image having embedded therein the secret information; receiving the image having embedded therein the secret information, decrypting the image having embedded therein the secret information, dividing the decrypted image into image blocks, dividing each image block into two or more parts, extracting the embedded secret information, and restoring the original image. Also disclosed is a device for information hiding. Utilization of the technical solution of the present invention increases the accuracy rate of secret information extraction and the accuracy rate of original image restoration.

Description

Information hiding method and device

The present invention relates to the field of information processing technologies, and in particular, to an information hiding method and apparatus. Background technique

As the most basic method to ensure information security, encryption technology plays an important role in the field of information security technology. The realization idea of the encryption technology is: using the key and the original data to perform some processing or operation on the transmitting end, so that the original data becomes unrecognizable; only the party having the key can recover the encrypted data by using the key at the receiving end. Come out and get the raw data. Although the encryption technology makes the original data unrecognizable, for multimedia data, such as video, audio and images, after encryption, the multimedia data becomes unrecognizable, that is, "invisible" or "inaudible", destroying The value and significance of the original data. In addition, the encryption technology makes the original data (plain text) into encrypted data (ciphertext). Although the ciphertext is difficult to recognize, it is the huge difference between the ciphertext and the plaintext, suggesting the importance of the original data. Increased the attention and attention of the attacker, allowing the attacker to steal the ciphertext and crack it.

In order to ensure the security of information transmission, information encryption technology and information hiding technology are usually combined. The information hiding technology is that the transmitting end encrypts the ordinary information as a carrier, and uses the embedded operation to embed the secret information into the carrier for transmission; the receiving end extracts the secret information hidden in the ordinary carrier by using the extracting operation, thereby obtaining the secret information. The carrier may be an image, a video or an audio. In this way, ordinary information is used as a transmission carrier of secret information for information exchange, which reduces attention and improves security.

In the prior art, the transmitting end encrypts the image I by using a key. Here, the encryption processing specifically refers to performing bitwise exclusive OR between the key and the image I, and then encrypting the encrypted image I as an image carrier. The specific information hiding process is as follows: The encrypted image is used as secret information. The carrier image is embedded in the information, and the image is first equally divided into a plurality of image blocks, and the pixels of each image block are Nx N; here, taking N=8, the number of bits of the secret information bits to be embedded is n=l, Then, the secret information bits to be embedded are 0 or 1; for each image block, the pixels are randomly and equally divided into 2 ⁿ parts by using a random key, namely: two parts S _G and Si;

When the secret information to be embedded is bit 0, all the pixels of the S ₀ portion of the current image block are inverted by the lower three bits, and the pixels of the Si portion remain unchanged;

When the secret information to be embedded is bit 1, all the pixels of the Si portion in the current image block are inverted by the lower three bits, and the pixels of the S _G portion remain unchanged;

Wherein, each pixel is composed of 8 bits; the above inversion of the lower three bits is to invert the lower three bits constituting each pixel to be the original embedded mode;

The above processing is performed for each image block, and the encrypted carrier image 1" after the secret information is embedded is obtained.

At the receiving end, the encrypted carrier image I" after receiving the embedded secret information is decrypted by using a key to obtain a decrypted image DI; first, the image DI is divided into a plurality of image blocks, and the pixel of each image block is Nx N, here Take N=8; use the same random key as the receiving end to randomly and equally divide each image block, because the number of bits of the secret information embedded in the transmitting end is n=l, so the embedded secret information bit is 0. Or 1, then, the number of bits of the secret information that needs to be extracted at the receiving end is still 1 bit. Here, each image block of the image DI is still divided into 2 ⁿ , that is, two parts Mo and Mi;

All the pixels in the Mo portion of the current image block are inverted by the lower three bits to obtain Mo, and all the pixels in the current image block are inverted by the lower three bits to obtain; Mo' and form a new image. Blocks B ₀ , Mo and ' constitute a new image block Bi; then, the selection is made in the new image blocks B ₀ , Bj , and B is calculated separately according to the formula ( 1 ) volatility prediction formula. And the volatility f of ^. , f _{i :}

Where x=0, 1, B _x , u, _v denotes B ₀ (B is the value of the pixel at position ( _M , v); the volatility of obtaining Β is f., the volatility of Bj is f, and judging f. and f size, when f. <fi,

B ₀ is the original image block, and the secret information bit 0 is extracted; when ί < ίο, 8 is the original image block, the secret information bit 1 is extracted;

As can be seen from the above description, the size of the comparison takes a smaller value, and finally the X of the smaller value f _x is converted from decimal to binary, which is the secret information bit that needs to be extracted.

Calculating the volatility of the two parts B ₀ sum of all image blocks, all original image blocks and all secret information bits embedded in the encrypted image frame are obtained.

Whether the secret information can be correctly extracted and whether the original image can be restored correctly depends on the correct prediction of volatility. The estimation of volatility is closely related to the embedded operation and the extraction operation. The embedded operation is applied to the image block. The greater the volatility, the better the results can be obtained when estimating; the more precise the estimation method of the extraction operation, the more accurate the extracted secret information.

Although the information hiding technology in the prior art improves the security of the secret information, the original embedding mode adopted by the embedding operation is the inverse of the lower three bits of the pixel (as shown in FIG. 3(c)), and the extraction operation is adopted. The volatility estimation formula is relatively simple. Therefore, there is a defect that the correct extraction rate of secret information and the correct rate of recovery of the original image are not high.

Summary of the invention

In view of this, the main purpose of the embodiments of the present invention is to provide an information hiding method and device, which can improve the secret information extraction correct rate and the original image recovery correct rate.

To achieve the above objective, the technical solution of the embodiment of the present invention is implemented as follows:

An embodiment of the present invention provides an information hiding method, where the method includes:

Encrypting the original image, dividing the encrypted image into image blocks, and dividing each image block into two or more parts, embedding the secret information into corresponding parts of the image block, and emitting an image embedded with the secret information; Receiving an image embedded with secret information, decrypting the image embedded with the secret information, dividing the decrypted image into image blocks, and dividing each image block into two or more parts, extracting the embedded secret information, and restoring the original image.

In the above solution, the encrypting the original image includes:

Encrypting the original image with an encryption key to form an encrypted image;

Embedding the secret information into a corresponding portion of the image block, and emitting the image embedded with the secret information, including:

The secret information bits are embedded into the corresponding portions of the image block using the embedded mode one or the embedded mode two, and the image blocks in which the secret information bits are embedded are assembled into an image and transmitted.

In the above solution, the decrypting the image embedded with the secret information comprises: decrypting the image embedded with the secret information bit by using the decryption key to form the decrypted image;

The extracting the embedded secret information and restoring the original image comprises: extracting the secret information bits from the corresponding part of the image block by using the embedded mode one or the embedded mode two, and the improved volatility prediction formula until being embedded in all the images The secret information of the block is extracted and the original image is restored.

In the above solution, the image after the encryption is divided into image blocks, and each image block is divided into two or more parts, including:

Dividing the encrypted image into multiple equal and non-overlapping image blocks, each image block having pixels

^ χ ^ ; Then use the random key to divide each image block into ² "parts, each part of the pixel is ^s2 l ^r ; where S and n are positive integers.

In the above solution, the embedding mode 1 or the embedding mode 2 is used to embed the secret information bits into the corresponding part of the image block, including:

According to the embedded mode one or the embedded mode two operation rule, the partial pixels of the first part of the current image block segmentation remain unchanged, and the lower three bits or the lower four bits are XOR between the partial pixels, so that The second part of the pre-image block segmentation remains unchanged until the last part of the segmentation; the partial pixels of the second part of the current image block segmentation remain unchanged, and the lower three bits or the lower four-bit XOR between the partial pixels , keeping the pixels of the first part and the third part of the current image block segmentation to the last part of the segmentation unchanged;

And so on, until the last part of the current image block is processed.

In the above solution, the segmentation of the decrypted image into image blocks, and dividing each image block into two or more portions, includes:

Dividing the decrypted image into multiple equal and non-overlapping image blocks, each image block having pixels

7x7; Reuse the random key to divide each image block into ² " parts, each part of the pixel is 7 ² /2"; where Y and η are positive integers, and Υ is the same as S.

In the above solution, the embedded information pattern is extracted from the corresponding portion of the image block by using the embedded mode 1 or the embedded mode 2 and the improved volatility prediction formula, until the secret information embedded in all the image blocks is extracted. The original image is restored, including:

According to the embedded mode one or the embedded mode two operation rule, the partial pixels of the first part of the current image block segmentation are kept unchanged, and the lower three bits or the lower four bits are XORed between the partial pixels, so that the current image block is split second. The pixels of the partial to the last part remain unchanged, and the first part to the last part of which the current image block has been embedded and the last part of the current image block pixel are combined into a new image block;

The partial pixels of the second part of the current image block segmentation are kept unchanged, and the lower three bits or the lower four bits are XORed between the partial pixels, so that the first part, the third part and the last part of the current image block are maintained. Invariably, the second portion of the current image block that has been subjected to the embedding operation is merged with the first portion, the third portion, and the last portion of the current image block pixel to form a new image block;

And so on, until the last part of the current image block is processed;

Calculate the volatility of all new image blocks using the improved volatility prediction formula The new image block with small fluctuation is the original image block, and the corresponding secret information bits are extracted; wherein the improved volatility estimation formula is:

/

C ^C one

Where C _u , _{¥ is} the value of the image pixel at position (u, V ), and Y is the root mean square of the pixel value of the image block.

The embodiment of the present invention further provides an information hiding device, where the device includes: an embedded module and an extraction and recovery module;

The embedding module is configured to encrypt the original image, divide the encrypted image into image blocks, and divide each image block into two or more parts, embed the secret information into corresponding parts of the image block, and embed the embedded image An image of the secret information is sent to the extraction and recovery module;

The extracting and restoring module is configured to receive an image embedded with secret information from the embedded module, decrypt an image embedded with secret information, divide the decrypted image into image blocks, and divide each image The block is divided into two or more parts, and the embedded secret information is extracted to restore the original image.

In the above solution, the embedded module includes: an encryption submodule, a first image segmentation submodule, and an information embedding submodule;

The encryption submodule is configured to encrypt the original image by using an encryption key; the first image segmentation submodule is configured to segment the encrypted image to form an image block, and then divide each image block into two The above part;

The information embedding sub-module is configured to embed the secret information bits into respective portions of the image block using the embedding mode one or the embedding mode two, and to transmit the image embedding the secret information bits to the extraction and recovery module.

In the above solution, the extraction and recovery module includes: a decryption sub-module, a second image segmentation sub-module, an information extraction and an image restoration sub-module;

The decryption submodule is configured to receive embedded secret information ratio from the embedded module a special image, and decrypting an image embedded with a secret information bit by using a decryption key; the second image segmentation sub-module configured to segment the decrypted image into image blocks, and segment each image block into More than two parts;

The information extraction and image restoration sub-module is configured to use the embedded mode one or the embedded mode, the improved volatility prediction formula, extract the embedded secret information bits, and restore the original image.

The information hiding method and device provided by the embodiment of the present invention, when embedding the secret information, use the embedded mode-operation rule or the embedded mode two operation rule to keep some pixels in the image block unchanged, and some pixels have low-order XOR or low. Four-bit XOR, which increases the volatility between image pixels; when extracting secret information and restoring images, image block pixel recovery is performed according to the embedding mode selected when embedding secret information, and some pixels remain unchanged, and some pixels are performed. Low-order or low-four-OR operation, and the improved volatility prediction formula is used to extract the embedded secret information and restore the original image; thus, the correct rate of secret information extraction and the correct rate of original image restoration can be improved. .

DRAWINGS

FIG. 1 is a schematic flowchart diagram of an information hiding method according to an embodiment of the present invention;

2 is a schematic flowchart of a method for hiding information according to an embodiment of the present invention;

FIG. 3( a ) is a schematic diagram of an operation principle of an embedded mode according to an embodiment of the present invention;

3(b) is a schematic diagram of an operation principle of an embedded mode 2 according to an embodiment of the present invention;

Figure 3 (c) is a schematic diagram of the original embedded mode operation principle;

4 is a schematic diagram of equal division of image blocks according to an embodiment of the present invention;

FIG. 5 is a schematic structural diagram of a structure of an information hiding apparatus according to an embodiment of the present invention. detailed description

An embodiment of the present invention provides an information hiding method. As shown in FIG. 1, the method includes: Step 10: Encrypt an original image, divide the encrypted image into image blocks, and divide each image block into two. In the above section, the secret information is embedded in the corresponding portion of the image block, and the image in which the secret information is embedded is issued.

Here, the secret information may be embedded into a corresponding portion of the image block by using the embedded mode one or the embedded mode two; the corresponding portion embedded in the image block means embedded in two or more parts of each image block segmentation, and some are embedded. Secret information, some do not embed secret information, and which parts of the embedded secret information can be random.

Step 11: receiving an image embedded with secret information, decrypting the image embedded with the secret information, dividing the decrypted image into image blocks, and dividing each image block into two or more parts, and extracting the embedded secret information. , restore the original image.

Here, the embedded secret information can be extracted using the embedded mode one or the embedded mode two, and the improved volatility estimating formula, and the embedding mode employed is the same as in the step 10.

Preferably, as shown in FIG. 2, step 10 includes:

Step 100: Encrypt the original image I with an encryption key to form an encrypted image. Here, the original image is I, I ( i, j ) is used to represent the pixel value at the position (i, j ) in the original image I, and ^^... ^ is used to represent the 8-bit of each pixel, 8 of each pixel The bit value formula (2) is shown:

As shown in the formula (3), the pixel of the original image I is XORed with the pseudo-random (bit) sequence, and the bit after the XOR is subjected to the operation of the equation (4) to obtain the encrypted pixel Γ.

i, J The image consisting of pixels i, J . is an encrypted image I.

(3) 7

^J ' j =∑^', ^X2 * Q<k<l (4) The pseudo-random (bit) sequence here is an encryption key.

Step 101: Divide the encrypted image I, form an image block, and divide each image block into two or more parts.

Here, each image block can be divided into 2" equal parts using a random key.

Specifically, the step may be: dividing the encrypted image into A equal and non-overlapping image blocks, each image block having a pixel SxS; and then equally dividing each image block into 2" portions using a random key *., 5 , &... 5*^, Thus, the pixels of each part are S ² /2".

Where n is the number of bits that need to be embedded in the secret information bits, and n is a positive integer; although the selected n is different, that is, the number of bits required to embed the secret information bits is different, so that the number of parts of the image block is different, but The process of dividing the image block into parts is similar; as long as the original image pixel and each image block pixel are >^ known, A, A can be calculated as a positive integer; the random key in the subsequent scheme and the random here The keys are the same.

Step 102: Embedding the secret information bits into the corresponding part of the image block by using the embedded mode one or the embedded mode two.

Here, in the embodiment of the present invention, the operation principles of the embedded mode 1 and the embedded mode 2 are respectively shown in FIG. 3 (a) and FIG. 3 (b), wherein an image block has 64 squares, and each square represents One pixel; where

In the image pixel model with embedded mode one and embedded mode two, a square filled with "0" indicates that the pixel remains unchanged; a square filled with "1" indicates that the pixel will be adjacent to it (left and right and/or up and down) A pixel that remains unchanged performs a lower three-bit XOR; a square filled with "2" indicates that the pixel will have a lower four-bit XOR with its neighboring pixels (left and right and/or up and down); Figure 3 (c) For the operation principle of the original embedded mode in the prior art, a square filled with "1" indicates that the pixel is inverted by the lower three bits.

In the embodiment of the present invention, the embedded mode 1 involves a lower three-bit XOR between pixel bits; Mode 2 involves both the low three-bit XOR of the pixels and the low four-bit XOR between the pixels. Although the selected embedding modes are different, the embedding process is similar.

For example, in this example, the embedded mode 1 is selected, and S=8, n=l is selected, indicating that each image block has a pixel 64 at this time, and the number of bits required to embed the secret information bit is 1 bit, that is: The embedded secret information bits are 0 or 1, then, at this time, each image block can be equally divided into 2"=2^2 parts, and the current image block is B. Then, B can be divided into two equal parts So and Si, as shown in Fig. 4, the white square in Figure 4 is S ₀ , the gray square is Si, and the pixels of Sc^pSi are

One.

When the embedded secret information bit is 0, using the embedded mode-operation rule, some pixels of the So part remain unchanged, and the lower three bits are XORed between the partial pixels, as shown by the formula (5) or the formula (6), Si All the pixels of the part remain unchanged;

Here, the operation mode of the embedded mode is: a square filled with "0" indicates that the pixel remains unchanged, and a square filled with "1" indicates that the pixel will remain adjacent thereto (left and right and/or up and down). The pixel performs a lower three-bit XOR;

. , _; , _{Λ =} 5 θ^,·- w (i,j) s ₀ k = 0,1,2 ₍₅₎ o,u, _k = B _ijt十 B _iJ+1 , _k (i, j) s ₀ k = 0,\,2 (6)

Where Bi, j, _k represents the kth bit of the position (i, j) pixel of the So part in the current image block B;

When the embedded secret information bit is 1, using the embedding mode-one operation rule, the partial pixels of the Si portion remain unchanged, and some of the pixels perform the lower three-bit XOR, as shown by the formula (7) or the formula (8), S. All the pixels of the part remain unchanged;

= B _jjk ten (i, j) es, k = 0,1,2 (7)

= B _jjk X B (i, j) e _Sl k = 0,1,2 (8) where Bi,j, _k represents the k-th bit of the position (i,j) pixel of the Si portion in the current image block B . With the above scheme, A equal and non-overlapping image blocks are embedded in secret information bits.

Here, since the number of bits n of the secret information bits to be embedded is different, the number of parts 2 ^{n of the} image block division is also different, but the process of dividing the image block into parts is similar, and the embedded mode 1 can be utilized. Or embed mode 2 to embed the secret information bits.

Here, n=l is set when the description is made; when n takes another positive integer greater than 1, the step 102 may be:

According to the embedded mode one or the embedded mode two operation rule, the partial pixels of the first part of the current image block segmentation are kept unchanged, and the lower three bits or the lower four bits are XORed between the partial pixels, so that the current image block is segmented. The second part of the segment to the last part of the segment remains unchanged;

The partial pixels of the second part of the current image block segmentation are kept unchanged, and the lower three bits or the lower four bits are XORed between the partial pixels, so that the first part and the third part of the current image block are segmented to the last part of the segmentation. The pixels remain unchanged;

And so on, until the last part of the current image block is processed.

Step 103: Combine the image blocks in which the secret information bits are embedded into the image D, and issue them. Further, as shown in FIG. 2, step 11 may include:

Step 110: Receive an image D in which a secret information bit is embedded, and decrypt the image D in which the secret information bit is embedded by using the decryption key to form a decrypted image DI.

Specifically, in this step, the pseudo-random (bit) sequence and the image D are used to perform a bitwise exclusive OR operation, as shown in the formula (9), and the decrypted image DI is obtained by using the formula (10);

= ® ^ 0 ≤ ≤ 7 ( 9 )

^DI u =∑ ^d u^ ^2k ( 1 ) where , , is the value of the pixel bit of the decrypted image DI position (i, j ), which is the image D, and the pixel bit of the position (i, j) takes the value; I)/ Decrypt the pixel of the image DI position (i, j); this The pseudo random bit sequence is a decryption key, and the decryption key and the encryption key may take the same value; Step 111: Segment the decrypted image DI to form an image block, and divide each image block into two or more parts.

Here, each image block can be divided into 2" equal parts using a random key.

Specifically, the step is: dividing the decrypted image DI into P equal and non-overlapping image blocks, each image block having a pixel ΓχΓ; using the same random key as in step 10, equally dividing each image block ²ⁿ parts M _Q , M _l ... M ₂ ⁿ -i, such that the pixels of each part are ² / ² ".

Here, since the number of bits of the secret information bit embedded in step 10 is n, the bit number of the secret information bit to be extracted is still n, and the division of the image block is still 2 ⁿ parts; wherein, as long as the original image pixel And each image block pixel is known, then P, P integer can be calculated;

This step corresponds to step 10. Usually, the value of P is the same as A, and the value of Y is the same as S. Step 112: using the embedded mode one or the embedded mode two, and the improved volatility prediction formula, extracting the secret information bits from the corresponding portions of the image block until the secret information bits embedded in all the image blocks are extracted, the original image Be restored

Specifically, the number of bits of the secret information bits embedded in step 10 is n=l, S=8, and the embedded mode 1 is selected; correspondingly, this step also selects the embedded mode one, and the extracted bit number n=l, Y= 8. Then, using the same random key as in step 10, the current image block C of the decrypted image DI is divided into two equal parts. And, M. The sum of the pixels is X ί72 ^η =32.

Use the algorithm of the embedded mode one, Μ. Part of the pixel remains unchanged, and some pixels perform the lower three-bit XOR as Equation (11) or Equation (12), forming M _G , and all the pixels remain unchanged;

('·,···) Ε Μ. 0,1,2

i , _K = _IJK ® C _W (j, j eM ₀ k = Q, \, 2 ₍₁₂ ) where, Ci, j, _k denotes the position of the M ₀ portion in the current image block C (i, j ) K-bit ratio Special; will be M. , part and part are grouped into a new image block Co;

By using the operation rule of the embedded mode one, part of the pixels remain unchanged, and some pixels perform the lower three-bit XOR as shown in the formula (13) or the formula (14) to form a part, M. All pixels of the section remain unchanged;

M ^ =C^ ®C _w (ί,])^Μ _λ k = 0,1,2 ₍₁₃₎

M[. _k =C _{jjt ®C} _{iJ+ k} (ι,])^Μ _λ k = 0,1,2 ₍₁₄ ) where, Ci,j, _k denotes the position (i,j) of the part in the current image block C The kth bit of the pixel; the 'part and the Mo part are grouped into a new image block.

Next, using the improved volatility estimation formula of equation (15), calculate the new image block Co and (^ the volatility fo, ft:

Where x=0, 1, Y=8; C _w denotes the value of the pixel where C ₀ (d ) is at position (u, v); compares the size of fo and , when ίο <ί, takes Co as the original image block , extracting the secret information bit 0; when ί <ίο, taking (^ is the original image block, extracting the secret information bit 1;

It can be seen that comparing the size of f _x takes a smaller value, and finally converting the X of the smaller value f _x from decimal to binary is the secret information bit that needs to be extracted, (^ is the original of the current image block C Image block.

The P equal and non-overlapping image blocks are extracted according to the above technical solution and the original image block is restored. The original image block is the original image.

The above only takes n=l as an example. Of course, n can also be other positive integers greater than 1, and the value of X varies with the value of n, x=0~2 ⁿ -l.

When n selected in step 10 is another positive integer greater than 1, the step 112 takes the same value of n as the step 10, and the step 112 may be:

According to the embedded mode one or the embedded mode two operation rule, the partial pixels of the first part of the current image block segmentation remain unchanged, and the lower three bits or the lower four bits are XOR between the partial pixels, so that The pixels from the second part to the last part of the pre-image block segmentation remain unchanged, and the first part of the current image block has been embedded and the second part of the current image block pixel remain unchanged. Image block

And so on, until the last part of the current image block is processed;

Using the improved volatility prediction formula (15), the volatility calculation is performed for all new image blocks, and the new image block with small volatility is taken as the original image block, and the corresponding secret information bits are extracted.

Based on the information hiding method, the embodiment of the present invention further provides an information hiding device. As shown in FIG. 5, the device includes:

The embedding module 50 is configured to encrypt the original image, divide the encrypted image into image blocks, and divide each image block into two or more parts, embed the secret information into corresponding parts of the image block, and embed the embedded image The image of the secret information is sent to the extraction and recovery module 51.

The extraction and recovery module 51 is configured to receive an image embedded with the secret information from the embedded module 50, decrypt the image embedded with the secret information, divide the decrypted image into image blocks, and divide each image block Split into two or more parts, extract the embedded secret information, and restore the original image.

In an actual application, the embedded module 50 and the extraction and recovery module 51 may be implemented by a central processing unit. (CPU, Central Processing Unit), or digital signal processor (DSP, Digital Signal Processor), or a programmable intermixed mutli bad ^{1 J (FPGA, Field Programmable Gate} Array) is achieved. The CPU, or DSP, or FPGA may be built into servers and/or terminals of both communication parties.

Here, the embedded secret information can be extracted using the embedded mode one or the embedded mode two, and the improved volatility estimating formula, and the embedding mode employed is the same as the embedding mode adopted by the embedding module 50.

Preferably, the embedding module 50 includes: an encryption submodule 500, a first image segmentation submodule 501, and an information embedding submodule 502;

The encryption sub-module 500 is configured to encrypt the original image by using an encryption key; the first image segmentation sub-module 501 is configured to segment the encrypted image to form an image block, and then divide each image block into More than two parts;

The information embedding sub-module 502 is configured to embed the secret information bits into the corresponding portion of the image block using the embedded mode one or the embedding mode two, and transmit the image in which the secret information bits are embedded to the extraction and recovery module 51.

Specifically, the encryption sub-module 500 encrypts the original image I by using an encryption key to form an encrypted image Γ; here, the encryption key may be a pseudo-random (bit) sequence; the original image I is at the position (i, j) The pixel value is /,,, ,. , , ₇ represents 8 bits of each pixel; the original image pixels are. The relationship between AfA is as shown in equation (2). The original image pixel is XORed with the pseudo-random (bit) sequence r _j , _k as shown in equation (3), and the result is the encrypted pixel. _ϋ , as _shown in the formula (4); wherein, the performing bitwise exclusive OR is substantially a bitwise OR between the bits bij, _k and the pseudo-random (bit) sequence constituting the original image pixel (k is 1~) 8 positive integer), the image composed of the encrypted pixel I'ij is an encrypted image Γ.

The first image segmentation sub-module 501 divides the encrypted image into two equal and non-overlapping image blocks, and equally divides each image block into 2" parts 5* using a random key. 5 ... S _r _ _x ; When each image block has a pixel SxS, each portion of the equally divided 2" portions has a pixel S ² /2"; wherein n is the number of bits of the secret information bits to be embedded, and the value is a positive integer; As long as the original image pixels and each image block pixel are known, A and A can be calculated as positive integers.

The n is preset, as long as the embedding module 50 and the extraction and recovery module 51 are consistent; A and S are preset in the embedding module 50, specifically in the first image. The dividing sub-module 501 is preset; the P and Y preset in the extracting and restoring module 51 are the same as the values of A and S here.

The embedding module 50 is preset to n=l, S=8, specifically, the first image segmentation sub-module 501 presets n=l, S=8, that is, the encrypted image I, and each equal image block that is divided. The pixel is 5 x = 64, and each image block is equally divided into 2 ⁿ = 2i = 2 parts; the current image block is B, and the random key is used to divide B into equal parts into So and Si, as shown in the figure. 4 is shown.

The embedded embedding sub-module 502 selects the embedding mode one, and the embedded secret information bits are 0, 1 as an example:

When the information embedding sub-module 502 embeds the secret information bit to 0, the operation rule of the embedding mode one determines that part of the pixels of the So part of the current image block B remain unchanged, and the lower three-bit XOR between the partial pixels, such as a formula ( 5) or (6), all the pixels of the Si portion remain unchanged; here, the operation mode of the embedded mode is: a square filled with "0" indicates that the pixel remains unchanged, and a square of "1" is filled. Indicates that this pixel will have a lower three-bit XOR for pixels that are adjacent to it (left and right and/or up and down);

When the information embedding sub-module 502 embeds the secret information bit to 1, the operation rule of the embedding mode one determines that part of the pixel of the Si portion of the current image block B remains unchanged, and the lower three-bit XOR between the partial pixels, such as a formula ( 7) or (8), all pixels of the S _Q portion remain unchanged; with the above technical solution, until the equal and non-overlapping A image blocks are embedded with secret information bits, then the secret information is embedded. All image blocks of bits are assembled into an image D, and sent to the extraction and recovery module 51. The extraction and recovery module 51 includes: a decryption sub-module 510, a second image segmentation sub-module 511, and an information extraction and image restoration sub-module 512;

The decryption sub-module 510 is configured to receive an image embedded with the secret information bit from the embedding module 50, and decrypt the image embedded with the secret information bit by using the decryption key; where the decryption sub-module 510 is specific Receiving an image embedded with the secret information bits from the information embedding sub-module 502;

The second image segmentation sub-module 511 is configured to divide the decrypted image into image blocks, and divide each image block into two or more parts;

The information extraction and image restoration sub-module 512 is configured to extract the embedded secret information bits and restore the original image by using the embedded mode one or the embedded mode two, the improved volatility prediction formula.

Specifically, the extracting and restoring module 51 receives the image D, and specifically, the decrypting sub-module 510 receives the image D, and decrypts the image D by using a decryption key; here, the decryption key and the encryption key are available. The same value, all pseudo-random (bit) sequences, a pseudo-random (bit) sequence and image D, a bitwise XOR operation, as shown in formula (9), and then use the formula (10) to obtain the decrypted image DI;

The second image segmentation sub-module 511 divides the decrypted image DI into P equal and non-overlapping image blocks, and equally divides each image block into equal pieces by using the same random key as the first image segmentation sub-module 501. Part Μ ₀ , Μ _ι ... Μ ₂ ^η _Λ ; each image block has a pixel y, then the pixel of each part of each image block is y ² / ² ";

Here, since the number of bits of the embedded secret information bit set in the embedding module 50 is n=l, the bit number of the secret information bit to be extracted in the extraction and recovery module 51 is still n=l; The division of the image block by the second image segmentation sub-module 511 is still 2 ⁿ = 2 portions; wherein, as long as the original image pixel and each image block pixel are known, a P, P positive integer can be calculated; The value of P and P can be the same as the values of S and A in the embedded module 50, so Y=8 can be selected here; Here, the encryption sub-module 500, the first image segmentation sub-module 501, the information embedding sub-module 502, the decryption sub-module 510, the second image segmentation sub-module 511, the information extraction and image restoration sub-module 512 may be by a CPU, or a DSP, or an FPGA. The CPU, or DSP, or FPGA can be built into the server and/or terminal of both communication parties.

The embedding module 50 selects the embedding mode one. Correspondingly, the extraction and recovery module 51 is the same as the embedding mode selected by the embedding module 50, and is also an embedding mode one;

The information extraction and image restoration sub-module 512 uses the operation rule of the embedded mode one to set the M of the current image block C. Part of the pixel remains unchanged, and some pixels perform the lower three-bit XOR as shown in formula (11) or formula (12), forming M _G , and all the pixels remain unchanged, and M ₀ , part and part are assembled into a new image block C ₀ ;

The information extraction and image restoration sub-module 512 maintains a part of the pixels of the current image block C by using the operation rule of the embedded mode one, and the partial pixels perform the lower three-bit XOR as the formula (13) or the formula (14). ), forming all pixels of the ΜΛ M _Q portion remain unchanged, and combining the Mi, part and Mo parts into a new image block d;

Finally, using the improved volatility prediction formula of equation (15), calculate the new image blocks C ₀ and (of the volatility fo, fi , and compare the size of fo sum; when <, take Co as the original image block, Extract the secret information bit 0; when ί <ίο, take (^ is the original image block, extract the secret information bit 1;

It can be seen that comparing the size of f _x and taking a smaller value, and finally converting the X of the smaller value f _x from decimal to binary is the secret information bit that needs to be extracted, (^ is the original image block of the current image block C. .

The P equal and non-overlapping image blocks are subjected to extraction of secret information bits and restoration of original image blocks according to the above technical solution, and the original image block is the original image.

In the above technical solution, the embedded mode 1 is selected, the secret information bit number n=l, the pseudo-random (bit) sequence is used as the encryption key and the decryption key, and the embedded mode can also be selected. Select other sequences as the encryption key and decryption key, n can also take any other positive integer, the secret information bit X is from 0~2 ⁿ -l.

The information hiding method and device provided by the embodiment of the present invention determine whether which pixels in the image block are unchanged, which pixels perform the lower three bits or the lower four bits XOR by embedding the mode one operation rule or the embedding mode two operation rule; On the embedding side, a large volatility is implanted between the image blocks; on the extraction and recovery side, which pixels in the image block are subjected to the lower three or lower four-bit XOR and which pixels are kept according to the selected embedding mode. Invariant, combined with the improved volatility estimation formula, the secret information is extracted, and the original image is restored. With the technical solution described in the embodiment of the invention, the error rate of the secret information extraction and the error rate of the original image recovery are reduced. .

The above is only the preferred embodiment of the present invention and is not intended to limit the scope of the present invention.

Claims

claims

1. An information hiding method, the method includes:

Encrypt the original image, divide the encrypted image into image blocks, divide each image block into two or more parts, embed the secret information into the corresponding part of the image block, and send the image with the secret information embedded;

Receive an image embedded with secret information, decrypt the image embedded with secret information, divide the decrypted image into image blocks, and divide each image block into two or more parts, extract the embedded secret information, and restore the original image.

2. The information hiding method according to claim 1, wherein the encryption of the original image includes:

Use the encryption key to encrypt the original image to form an encrypted image;

Embedding secret information into the corresponding part of the image block and sending the image with the secret information embedded includes:

Embedding mode one or embedding mode two is used to embed secret information bits into corresponding parts of image blocks, and the image blocks embedded with secret information bits are assembled into images and sent out.

3. The information hiding method according to claim 1 or 2, wherein the decrypting the image embedded with secret information includes:

Use the decryption key to decrypt the image embedded with secret information bits to form a decrypted image;

The method of extracting embedded secret information and restoring the original image includes:

Using embedding mode one or embedding mode two and the improved volatility prediction formula, the secret information bits are extracted from the corresponding parts of the image blocks until the secret information embedded in all image blocks is extracted and the original image is restored.

4. The information hiding method according to claim 1 or 2, wherein the encrypted image Divide the image into image blocks, and divide each image block into two or more parts, including: Divide the encrypted image into multiple equal and non-overlapping image blocks, each image block has pixels

S x S then uses a random key to divide each image block equally into ² " parts, and the pixels of each part are ^s2 l ^T ; where S and n are both positive integers.

5. The information hiding method according to claim 4, wherein the use of embedding mode one or embedding mode two to embed secret information bits into corresponding parts of the image block includes;

According to the operation rules of embedding mode one or embedding mode two, some pixels in the first part of the current image block segmentation are kept unchanged, and the lower three bits or lower four bits are XORed between the partial pixels, so that the third part of the current image block segmentation is The pixels from the two parts to the last part of the segmentation remain unchanged;

Keep part of the pixels in the second part of the current image block unchanged, and perform XOR on the lower three or lower four bits between the partial pixels, so that the first part and the third part of the current image block are divided into the last part of the division. pixels remain unchanged;

And so on until the last part of the current image block is processed.

6. The information hiding method according to claim 4, wherein said dividing the decrypted image into image blocks and dividing each image block into two or more parts includes:

Divide the decrypted image into multiple equal and non-overlapping image blocks, each image block has pixels 7x7; then use the random key to equally divide each image block into ² " parts, each part has 7 pixels ² /2"; Among them, Y and eta are both positive integers, and Y and S have the same value.

7. The information hiding method according to claim 6, wherein the secret information bits are extracted from the corresponding parts of the image block using embedding mode one or embedding mode two and an improved volatility prediction formula until embedding Secret information to all image patches is extracted and the original image is restored, including:

According to the operation rules of embedding mode one or embedding mode two, some pixels in the first part of the current image block segmentation are kept unchanged, and the lower three bits or lower four bits are XORed between the partial pixels, so that the third part of the current image block segmentation is The pixels from the second part to the last part remain unchanged, and the current image block The first part that has been embedded and the second part to the last part whose pixels remain unchanged in the current image block are combined into a new image block;

Keep some of the pixels in the second part of the current image block segmented unchanged, and perform XOR on the lower three or lower four bits between the partial pixels, so that the first, third and last part of the current image block segmented pixels remain unchanged. unchanged, the second part of the current image block that has been embedded and the first part, the third part to the last part of the current image block whose pixels remain unchanged are combined into a new image block;

And so on, until the last part of the current image block is processed;

Use the improved volatility prediction formula to calculate the volatility of all new image blocks, take the new image blocks with small volatility as the original image blocks, and extract the corresponding secret information bits; where, the improved volatility The prediction formula is: = C _u , _vx2 ® C _u ^ _vx2 _ ₁ where C _u , _¥ is the value of the image pixel at the position (u, v), and Y is the root mean square of the pixel value of the image block.

8. An information hiding device, the device includes: an embedding module and an extraction and recovery module; wherein,

The embedding module is configured to encrypt the original image, divide the encrypted image into image blocks, divide each image block into two or more parts, embed the secret information into the corresponding part of the image block, and embed the The image of the secret information is sent to the extraction and recovery module;

The extraction and recovery module is configured to receive an image embedded with secret information from the embedding module, decrypt the image embedded with secret information, divide the decrypted image into image blocks, and divide each image into image blocks. The block is split into two or more parts, the embedded secret information is extracted, and the original image is restored.

9. The information hiding device according to claim 8, wherein the embedding module includes: an encryption sub-module, a first image segmentation sub-module, and an information embedding sub-module; wherein, The encryption sub-module is configured to encrypt the original image using an encryption key; the first image segmentation sub-module is configured to segment the encrypted image to form image blocks, and then divide each image block into two The above parts;

The information embedding sub-module is configured to embed secret information bits into corresponding parts of the image block using embedding mode one or embedding mode two, and send the image with the secret information bits embedded to the extraction and recovery module.

10. The information hiding device according to claim 8 or 9, wherein the extraction and recovery module includes: a decryption sub-module, a second image segmentation sub-module, an information extraction and image recovery sub-module; wherein,

The decryption sub-module is configured to receive an image embedded with secret information bits from the embedding module, and use a decryption key to decrypt the image embedded with secret information bits;

The second image segmentation sub-module is configured to segment the decrypted image into image blocks, and segment each image block into two or more parts;

The information extraction and image restoration sub-module is configured to use the improved volatility prediction formula of embedding mode one or embedding mode two to extract the embedded secret information bits and restore the original image.