KR101716319B1 - Robust reversible data hiding method and apparatus for image data of png file format - Google Patents

Abstract

An aspect of the present invention discloses a robust reversible data hiding method for image data in a Portable Network Graphics (PNG) file format. The robust reversible data hiding method for image data in a PNG file format comprises the steps of: converting image data in a PNG file format into image data in a BMP file format; separating R, G, and B channels from the image data in the BMP file format, and generating three channel-specific images; performing Discrete Wavelet Transform (DWT) on each of the three R, G and B channel-specific images; assigning high-frequency (HH) regions of the three images, on which the DWT has been performed, for insertion regions based on a Human Visual System (HVS); dividing each of the HH regions into a Group of Blocks (GoBs) having a height of 16; generating a predicted value based on average values of an 8x8 block and surrounding blocks included in the GoBs; hiding a secret bit in a residual value obtained based on the predicted value and an actual average value; changing a block value so as to correspond to a new average value into which the hidden secret bit has been incorporated; performing Inverse DWT (IDWT) on blocks having the changed block value, and generating embedded R, G, and B channel images; combining the embedded R, G, and B channel images together, and generating image data in an embedded BMP file format; and converting the image data in the embedded BMP file format into image data in the PNG file format. In this case, the secret bit having a constant value is hidden in each of R, G, and B channels for redundancy.

Description

Technical Field [0001] The present invention relates to a robust reversible information hiding method and apparatus for image data of a PNG file format,

The present invention relates to a hiding method and apparatus, and more particularly, to a robust reversible information concealment method related to PNG (Portable Network Graphics) files.

In general, digital content such as images, movies, multimedia, etc. is not easy to grasp the speed and range of movement and copying on the Internet, how much is used and distributed, and thus the rights of copyright holders of digital contents are likely to be infringed And it is true that it is difficult to make highly reliable settlement of royalties due to the uncertainty of distribution status.

Therefore, techniques for effectively inserting copyright information into digital contents are actively being developed. In most cases, the length of such copyright information is at least 50 bytes, and it is necessary to be able to accurately identify the digital content existing somewhere on the Internet and to grasp the usage status using the concealed information, Or can not be extracted and corrected.

Watermarking and fingerprinting techniques are used to automatically identify images. Watermarking can insert the relevant information either visually or invisibly into the image signal and extract it if necessary to identify the image. Fingerprinting technology extracts feature points inherently possessed by an image, stores them in a database form, extracts feature points of the image to be identified, and identifies the feature points by mutual comparison. Thus, the two techniques described above can implement a relatively robust algorithm for non-geometric transformations such as Gaussian noise, brightness changes, and geometric transformations such as enlargement and reduction, rotation, and aspect ratio change.

However, the two techniques described above are not so suitable from the viewpoint of the amount of information that can be inserted and the damage of the original. For example, considering the amount of information that can be embedded, the fingerprinting method can only identify the image and can not insert the information. The watermarking method does not insert the information in the form of image, It is not suitable for embedding a sufficient amount of information in the distribution information concealment.

Also, from the viewpoint of the original damage, the fingerprinting method does not cause any original damage, but the watermarking method inevitably causes the original damage and the stronger the attack, the more the damage becomes stronger. Distribution information needs to be updated from time to time due to copyright transactions, movement of images, etc. If image damage is continuously caused by frequent modification of watermarking, image quality is degraded and the meaning to be conveyed is distorted There is a concern.

Japanese Patent Laid-Open No. 2003-536113

According to an aspect of the present invention, there is provided an image compression method for compressing an image compressed by a PNG method, And to provide a reversible information concealment method and apparatus robust against PNG recompression and partial content modification attack that can naturally occur in distribution.

According to an aspect of the present invention, there is provided a method of concealing robust reversible information of image data in a PNG file format, the method comprising: converting image data of a PNG (Portable Network Graphics) file format into a BMP file format; Separating the R, G, and B channels from the image of the R, G, and B channels, generating three images separated by channels, converting the three separated images into DWT (Discrete Wavelet Transform) Designating an HH (high frequency) region as an insertion region based on HVS (Human Visual System), dividing the HH region into GoB (Group of Block) having a height of 16, Generating a predicted value through an average value of neighboring blocks, hiding a secret bit at a residual value of the predicted value and the actual average value, Generating embedded R, G and B channel images by performing an IDWT (Inverse DWT) transformation on the blocks having the changed block values, integrating the embedded R, G, and B channel images, And converting the image data of the embedded BMP file format into image data of a PNG file format, wherein the secret bits include R, G, and B bits for redundancy, You can conceal the same value in the channel.

The insertion region is designated as a top-end high-frequency region by utilizing the characteristics of HVS, and the insertion size can be calculated as H * W * 5/16 (where H is the height of the image and W is the size of the image) .

According to the robust reversible information concealment method and apparatus of the image data of the PNG file format according to an aspect of the present invention, it is possible to conceal a sufficiently large amount of information without damaging the original compressed image in PNG format, There is a strong effect on content modification attack.

1 is a conceptual diagram showing a structure of a PNG file,
FIG. 2 is a flowchart schematically illustrating a robust reversible information concealment method of image data of a PNG file format according to an embodiment of the present invention.
FIGS. 3A and 3B are detailed flowcharts illustrating a file conversion process of a robust reversible information concealment method of image data of a PNG file format according to an exemplary embodiment of the present invention;
4 is a conceptual diagram for explaining a process of generating three images separated into R, G, and B channels,
FIG. 5 is a conceptual diagram for explaining a DWT transformation, a high-frequency domain segmentation, and a GoB segmentation process of the robust reversible information concealment method of the image data of the PNG file format according to an embodiment of the present invention.
6 is a conceptual diagram for explaining a two-dimensional DWT transformation,
7 is a conceptual diagram for explaining a process of generating an embedded PNG file.

While the invention is susceptible to various modifications and alternative forms, specific embodiments thereof are shown by way of example in the drawings and will herein be described in detail.

It is to be understood, however, that the invention is not to be limited to the specific embodiments, but includes all modifications, equivalents, and alternatives falling within the spirit and scope of the invention.

The terms first, second, etc. may be used to describe various components, but the components should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, the first component may be referred to as a second component, and similarly, the second component may also be referred to as a first component. And / or < / RTI > includes any combination of a plurality of related listed items or any of a plurality of related listed items.

It is to be understood that when an element is referred to as being "connected" or "connected" to another element, it may be directly connected or connected to the other element, . On the other hand, when an element is referred to as being "directly connected" or "directly connected" to another element, it should be understood that there are no other elements in between.

The terminology used in this application is used only to describe a specific embodiment and is not intended to limit the invention. The singular expressions include plural expressions unless the context clearly dictates otherwise. In the present application, the terms "comprises" or "having" and the like are used to specify that there is a feature, a number, a step, an operation, an element, a component or a combination thereof described in the specification, But do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, or combinations thereof.

Unless defined otherwise, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning in the context of the relevant art and are to be interpreted in an ideal or overly formal sense unless explicitly defined in the present application Do not.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. In order to facilitate the understanding of the present invention, the same reference numerals are used for the same constituent elements in the drawings and redundant explanations for the same constituent elements are omitted.

PNG  File structure

1 is a conceptual diagram showing a structure of a PNG file.

Referring to FIG. 1, a PNG data stream (PNG datastream) consists of a PNG signature followed by consecutive chunks. The PNG data stream begins with an 8-byte signal representing the PNG file, which is called the PNG signature.

The header (PNG signature) is followed by a junk containing information about the image. A chunk can have a chunk type that is specific to each function.

IHDR is the image header and is the first chunk of the PNG data stream. It displays the width, height, number of bits and color type of the image in order. The PLTE provides a palette table (indicating the color space) associated with the index image. IDAT is an image data chunk that displays an image that can be split into multiple IDAT chunks. Although the file size is slightly larger, it makes the PNG streamable. The IDAT may also have a real image file through the output stream of the compression algorithm. IEND is the image trailer as the last chunk of the PNG data stream.

In this embodiment, the PNG data stream includes information related to transparency (tRNS), color space information (cHRM, gAMA, iCCP, sBIT, sRGB), character information (iTXt, tEXt, zTXt), time information And other information (bKGD, hIST, pHYs, sPLT).

Also, with respect to the chunk layout, each chunk can have three or four fields. This may include length information, chunk type information, chunk data, and CRC. At this time, the chunk data field may be empty in some cases.

PNG  How to hide reversible information on file format video

2 is a flowchart schematically illustrating a robust reversible information concealment method of image data of a PNG file format according to an embodiment of the present invention.

Referring to FIG. 2, the reversible information concealment apparatus (not shown) of the PNG file first converts the PNG file into a BMP file (S210). When hiding data during compression, the PNG file is converted into the BMP format in order to minimize the influence of data distortion, and the data is concealed.

After the conversion into the BMP file, the image of the BMP file is divided into the images of the three channels of R, G, and B (S220). Then, DWT (Discrete Wavelet Transform) is performed on the images of the three separated channels, respectively (S230). At this time, a two-level DWT transformation is preferably used.

Then, the uppermost high frequency area (HH area) is designated as the secret bit insertion area using the characteristics of the HVS (Human Visual System) (S240). After designating the insertion area, the apparatus divides the insertion area into GoB (Group of Block) having a height of 16 (S250).

)을 생성하고(S260), 예측값(

)과 실제 평균값(M)의 차이값(d)에 비밀 비트(b)를 은닉한다(S270). Then, the average value of the 8x8 block pixels is obtained, and the average value (M) of the neighboring blocks is used to calculate the predicted value

(S260), and generates a predicted value (

The secret bit b is hidden in the difference value d between the actual average value M and the actual average value M in step S270.

Then, a new average value M 'considering the hidden secret bit b is obtained, and then the block value is changed so that the average of the 8x8 block becomes M' (S280). When information is concealed, it is desirable to conceal the same information on each of the R, G, and B channels in order to secure redundancy.

Then, an embedded channel is generated through an IDWT (Inverse DWT) conversion (S290), and the same process is performed on the R, G, and B channels. Then, an embedded BMP file is generated by integrating the embedded R, G, and B channels (S300). Then, in step S 210, the embedded BMP is converted into an embedded PNG file, and a PNG file in which robust reversible information is inserted is generated (S 310). Hereinafter, each step will be described in more detail with reference to FIGS. 3A to 7.

PNG  File format and BMP file Between formats  conversion

3A and 3B are detailed flowcharts illustrating a file conversion process of the robust reversible information concealment method of the image data of the PNG file format according to an embodiment of the present invention.

Referring to FIG. 3A, a PNG data stream may be transformed into a BMP image through a reconstruction process. On the contrary, the BMP image extracts the reduced image through the extraction process, performs scanline serialization, performs filtering, and compresses and chunking to generate a PNG data stream. At this time, the compression step 310 is performed as shown in FIG. 3B.

Referring to FIG. 3B, the compression 310 may be performed by a DEFLATE algorithm, which may be performed through the LZ77 algorithm 320 and the Hoffman coding 330. That is, the DEFLATE algorithm includes raw data, i.e., input data, and is encoded into LLD data by the LZ77 algorithm 320, and the encoded data has a format of length, distance, and literal. With this frequency, it is represented by the Huffman algorithm 330 as the non-repeated bits of 1 and 0 for the frequency and completes the compression. The DEFLATE algorithm is a lossless extrusion algorithm and has the feature that the compression / decompression speed is faster than the compression rate.

Thus, the BMP image can be converted into the PNG data stream through the compression algorithm of FIGS. 3A and 3B, and conversely, the process of converting the PNG data stream into the BMP image is performed in the reverse process of the PNG data stream generation process.

3 channel video produce

4 is a conceptual diagram for explaining a process of generating three images separated into R, G, and B channels.

4, the apparatus receives a PNG file format (lena512.png) as an input, converts it into a BMP file (lena512.bmp), and converts the data of the BMP file format into R, G, B Channels, and three images are generated. The reason for creating a file in BMP format is that, if data is hidden during compression, the LZ77 algorithm included in the Deflate algorithm may be affected and the data may be distorted.

DWT  Conversion and GoB  Division

FIG. 5 is a conceptual view for explaining a DWT transformation, a high-frequency domain segmentation, and a GoB segmentation process of the robust reversible information concealment method of the image data of the PNG file format according to an embodiment of the present invention.

Referring to FIG. 5, DWT conversion is performed on three image data of R, G, and B channels, respectively. In the present embodiment, DWT conversion for an R-channel image will be described as an example. It should be apparent to those skilled in the art that while the embodiment describes a transformation for one R channel image, the same approach must be performed for all other channel images. At this time, since the image is two-dimensional, it is preferable to use a two-dimensional DWT transformation.

6 is a conceptual diagram for explaining a two-dimensional DWT transformation.

Referring to FIG. 6, filtering is performed for the row direction and the column direction, respectively, to generate the final four band-specific conversion data. At this time, the decomposition filter first performs filtering on the low level, and the synthesis filter performs the filtering on the column level first.

The band division of the wavelet can be divided into LL region, LH region, HL region and HH region when divided into four regions. From the LL region to the HH region can be sequentially corresponded from the lowest frequency region to the highest frequency region.

Referring back to FIG. 5, in designating an insert row based on the DWT-transformed image of the R channel, the uppermost high frequency region (HH region = HH ₁ , HH ₂ ) can be designated as an insertion region by utilizing the characteristics of HVS . That is, when the height of the image is H and the width is W, the size of the insertion area may be H * W * 5/16 per channel. For example, assuming an image of 512x512, the size of the insert region is 81920.

Then, for the preprocessing of the DE algorithm, the segment is divided into GoBs with an insertion area height of 16. Then, the DE algorithm is applied to the upper left part of the divided GoB in units of 16 × 16 blocks. The application of the DE algorithm is described in more detail with reference to FIG.

Embedded PNG  File format Image data completion

7 is a conceptual diagram for explaining a process of generating an embedded PNG file.

Referring to FIG. 7, a 16x16 block is considered to be divided into four 8x8 block units in the application of the DE algorithm to a 16x16 block. At this time, M is defined as an average value of 8x8 block pixels.

)을 산출한다. Then, an average value is obtained for each 8x8 block, and the predicted value (

).

)과 실제 평균값(M)의 차이값(d)에 비밀비트(b)를 은닉한다. 이를 수학식으로 표현하면, 다음과 같다.Then,

) And the difference value d between the actual average value M and the secret bit b. This can be expressed by the following equation.

[Equation 1]

Then, a new secret bit (b) is calculated for the block that has been hidden, and a new average value (M ') is calculated. At this time, in order to secure redundancy, it is preferable to conceal the same information on each of the R, G and B channels.

After calculating the new average value M ', the values of the blocks are converted so that the average of the 8x8 block becomes M', and an image of the embedded channel is generated through the IDWT transformation. The process of applying IDWT for image generation can be expressed as the following equation.

&Quot; (2) "

Here, r _i may be a block value of 8 × 8 blocks before the secret bit concealment, and r _i 'may be a changed block value of 8 × 8 block for matching a new average value M'. That is, in the case of an image of H * W, it can have a payload of H * W * (15/16) * 1 / (16 * 16) (b).

The above process is performed in the same manner on the R, G, and B channels.

After the embedded R, G, and B channels are generated by performing the IWDT based on the changed block values, the embedded R, G, and B channels are combined to generate the combined embedded BMP format image data .

Then, the image data of the embedded BMP format is converted to generate image data of the embedded PNG format.

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit or scope of the inventions as defined by the following claims It will be understood that various modifications and changes may be made thereto without departing from the spirit and scope of the invention.

Claims (2)

Converting image data of a PNG (Portable Network Graphics) file format into a BMP file format;
Separating the R, G, and B channels from the image of the BMP file format to generate three images separated by channels;
Performing DWT (Discrete Wavelet Transform) conversion on the three images separated by the R, G, and B channels;
Designating an HH (high frequency) region as an insertion region based on HVS (Human Visual System);
Dividing the HH region into GoB (Group of Block) having a height of 16;
Generating predicted values through an average value of 8x8 blocks and neighboring blocks in the GoB;
Hiding a secret bit in the residual value of the predicted value and the actual average value;
Modifying the block value according to a new average value considering the hidden secret bits,

And

Changed using

D is the residual value according to the secret bit insertion, M 'is the new average value, r _i is the secret bit, A block value of 8x8 block before concealment, and r _i 'means a changed block value of 8x8 block for matching a new average value M');
Performing an IDWT (Inverse DWT) transformation on the blocks having the changed block values to generate embedded R, G, and B channel images;
Generating image data of an embedded BMP file format by integrating the embedded R, G, and B channel images; And
Converting the image data of the embedded BMP file format into image data of a PNG file format,
The secret bit hides the same value for the R, G, and B channels for redundancy,
The insertion region is designated as a top-end high-frequency region by utilizing the characteristics of HVS,
The insertion size is calculated by H * W * 5/16 (where H is the height of the image and W is the size of the image).
delete

Images