KR102574832B1 - A Method of Anti-steganography for Image File - Google Patents

Abstract

As a countermeasure against steganography attacks using image files, the present invention is to provide a method for completely neutralizing meta contents hidden in image files.
An anti-steganography method of an image file according to the present invention includes a decomposition step of obtaining image data by decomposing an image file according to a specification of an image file format; A conversion step of converting the image data obtained in the decomposition step into an RGB format; an anti-steganography step of quantizing the RGB-type image data converted in the conversion step to neutralize the meta content inherent in the image file; and a restoration step of converting image data in an RGB format in which meta contents are disabled through the anti-steganography step into a format, and combining the converted image data according to the format.

Description

A Method of Anti-steganography for Image File}

The present invention relates to detoxification of image files, and in particular, to anti-steganography in preparation for steganography attacks using images.

Steganography can be viewed as secret key cryptography. The steganographic method is a method of concealing the existence of a message itself. The activation of information and communication technology has particularly developed digital steganography techniques, which include changing the LSB (Least Significant Bit) of a file or pixel RGB (RGB: Red/Green/RGB) of an image file. There are various methods such as a method of hiding a message by changing the LSB of each pixel included in Blue) information. Steganography is often used in media such as images and MP3, and messages are often hidden in document files.

The advantage of steganography is that it hides files within files. Therefore, it is difficult for third parties to see hidden files unless they know how they were hidden. In the case of encryption, even a third party can attempt decryption if they detect that it is encrypted data, so it is also an advantage that safety is further secured in the case of steganography, which is difficult to even know whether it is encrypted or not.

An encryption method (Korean Patent Laid-open Publication No. 10-2019-0005346), an information security device (Korean Patent Publication No. 10-2022-0003318) and a communication method (Korean Laid-Open Patent Publication No. 10-2022-0003318) using the advantages of steganography 10-2018-0029756) and the like have been proposed.

However, in addition to the above positive uses, it also has significant negative uses, especially as it is easy to hide computer viruses, Trojan horses, or malicious codes. September 11, 2001. At the time of the attack, despite the fact that the United States was operating a fairly sophisticated international wiretapping system, Osama bin Laden was able to use steganographic techniques on the Mona Lisa painting to send 19 terrorists a message to hijack four planes at the same time and the plane blueprints. . In addition, the malware 'Duqu', which occurred in 2011, carried out a cyber attack that collected information from the victim's system, hid it in a JPEG file, and transmitted it to its control server. Steganography techniques were also used at this time.

From the above cases, it can be seen that cyber attacks using steganography techniques mainly utilize image files. This is because many security solutions pass image files as is.

The steganography technique is very difficult to pattern, making it difficult to detect with anti-virus. In addition, the variation is very diverse, so machine learning techniques such as artificial intelligence are not effective. In particular, the fatal part is that it is quite difficult to determine whether or not the meta content of malicious intent is hidden (hereinafter referred to as 'stego image'). Simple image steganography methods, such as LSB-based, can be easily detected in detection using statistical differences, but in response, steganography methods using more sophisticated techniques are constantly being developed.

For this reason, the disarming method using CDR (Contents Disarm & Reconstruction) is attracting attention as a countermeasure against cyber attacks using steganography. Unlike the malware analysis method used in existing vaccines, CDR disassembles files and removes all components that do not conform to system policies.

However, CDR is a technology optimized for attacks using document files, and it has limitations for attacks using image files. If the malicious code is appended to the end of the image file and hidden, it can be safely reconstructed with CDR. However, if steganography is applied based on the bit values inserted into specific information in the image, the malicious code will not remain even after CDR processing. can In other words, since CDR does not remove components that conform to the system's policy, but leaves them behind, if malicious code is hidden through LBS (if it is not detected as a malicious code pattern), the file can be defined as a malicious file. there will be no

Meanwhile, an invention (registered patent No. 10-2103306) related to a steganographic discrimination device and method is disclosed. The present disclosure discloses a step of determining whether steganography is applied to an image. However, variants of steganography are expected to appear constantly. Therefore, it seems that it is only a matter of time before the steganography technique that avoids the steganography discrimination method provided by the present disclosure is used in an attack.

In the current situation, CDR technology, which removes all meta contents without asking whether they are hidden or malicious, seems to be the best possible. However, CDR also has the aforementioned imperfections. Therefore, there is an urgent need to invent an anti-steganography method to compensate for the imperfection of CDR.

The present invention was conceived from the recognition of the incompleteness of CDR, which is a countermeasure against steganography attacks using image files, and is intended to provide a complete neutralization method for meta contents hidden in image files.

In order to solve the above problems, the anti-steganography method of an image file according to the present invention includes a decomposition step of obtaining image data by decomposing an image file according to a specification of an image file format; A conversion step of converting the image data obtained in the decomposition step into an RGB format; an anti-steganography step of quantizing the RGB-type image data converted in the conversion step to neutralize the meta content inherent in the image file; and a restoration step of converting image data in an RGB format in which meta contents are disabled through the anti-steganography step into a format, and combining the converted image data according to the format.

In one embodiment of the present invention, the anti-steganography step converts image data in RGB format into YCbCr format; converting the image data converted into a YCR format into a frequency domain; performing quantization on image data converted to a frequency domain; performing dequantization on the quantized image data; converting the inverse quantized image data into a spatial domain; and converting the spatial domain-converted image data into an RGB format.

According to an embodiment, according to claim 2, the object of quantization performed in the anti-steganography step is at least one of Y (Y), CB (Cb), and CR (Cr) Anti-steganography method of an image file is provided It can be.

According to an embodiment, a method for anti-steganography of an image file in which the degree of quantization performed in the anti-steganography step is determined according to a predefined table may be provided.

Depending on the embodiment, the predefined table may be a table defined by assigning a weight to each frequency domain of image data converted to the frequency domain.

According to an embodiment, the anti-steganography step may further include performing chroma subsampling before converting image data converted into a YCBR form into a frequency domain.

An anti-steganography method according to an embodiment of the present invention may further include a detoxification step of removing metadata included in an image file after the decomposition step.

Depending on the embodiment, metadata removed in the disarming step may be metadata selected according to a predefined policy.

The present invention may be provided in the form of a computer-readable storage medium in which a program for performing the above-described anti-steganography method of an image file is stored.

The present invention has the following effects.

First, it perfectly compensates for the imperfections of CDR by enabling secure image files to be acquired regardless of whether the image file is a steganographic image or the degree of steganography.

Second, the quality of the existing image file can be preserved as much as possible even after detoxification.

In addition, the present invention may additionally have the following effects according to embodiments.

First, by performing quantization in the frequency domain, quantization is performed in consideration of sensitivity according to frequency of human vision, thereby minimizing damage to actual image information and effectively harmless meta-contents.

Second, by predefining a table for performing quantization, it is possible to appropriately secure image quality as needed.

Third, computational load can be reduced by reducing the amount of data to be processed through chroma subsampling.

Fourth, it is possible to more tightly block the inflow of malicious codes by removing not only meta contents hidden by the steganographic method, but also meta data included in image formats. According to the embodiment, by selecting and removing metadata to be removed, an optimized security means for each user can be provided.

1 shows each component of an anti-steganographic device for carrying out the present invention.
Figure 2 shows each step of the anti-steganography method of the present invention.
3 is a subdivided anti-steganography step (S140) shown in FIG. 2 according to an embodiment of the present invention.
4 shows Y, CB, CR, and images according to a quantization table and different tables for implementing an embodiment of the present invention.
5 shows each configuration of an anti-steganography device for implementing an embodiment of the present invention.
FIG. 6 shows each step of an anti-steganography method according to an embodiment of the present invention performed by the anti-steganography apparatus of FIG. 5 .

Hereinafter, an embodiment of the present invention will be described in detail with reference to the accompanying drawings. In describing the present invention, if it is determined that a detailed description of a related known function or configuration may unnecessarily obscure the gist of the present invention, the detailed description will be omitted. In addition, the terms to be described later are terms defined in consideration of functions in the present invention, which may vary according to the intention or custom of a user or operator. Therefore, the definition should be made based on the contents throughout this specification.

In this specification, 'meta contents' refers to all data hidden in an image file using a steganographic technique. Depending on the purpose of the steganography attack, the meta content can be of various kinds, such as maliciously distributed information, malicious codes and viruses to capture the system receiving the image file. Since this specification is intended to explain the 'anti-steganography method', meta contents can be understood as all data or information hidden in an image file that can harm a system that has received the image file. However, it is not necessarily limited thereto, and it will suffice to be understood as all data or a set thereof that is hidden in an image file and can be harmless or removed according to the present invention.

In this specification, a 'stego file' means an image file in a state in which meta contents are hidden. All image files for which the anti-steganography method according to the present invention is performed may not be Stego files, but if the image files are Stego files, meta contents hidden in the image files are removed or harmless.

<Anti-steganography device 100>

1 is an anti-steganography apparatus 100 prepared to carry out the present invention.

Referring to FIG. 1, the anti-steganography apparatus 100 includes a storage module 110, an analysis module 120, a conversion module 130, and an anti-steganography module 140 (hereinafter referred to as 'each module'). ). In this specification, each module will be defined by the function that the corresponding module performs. That is, it should be understood that each module may not only be implemented as a single device, but may also be implemented as a part of a single device or an organically combined form of each part of several devices. Since a person skilled in the art may employ suitable known techniques to implement each module, detailed descriptions thereof are omitted.

The storage module 110 is provided to store data or information including images and operate in conjunction with each module. That is, the storage module 110 may provide data stored in each module or receive and store data from each module.

The analysis module 120 is interlocked with the storage module 110 to access data or information including image files stored in the storage module 110 . The analysis module 120 is provided to perform a disassembly operation for obtaining information on the format of an image file. The analysis module 120 is provided to acquire image data by decomposing an image file according to the format of the image file. Also, the analysis module 120 may recover an image file by recombining the decomposed image data. Information about the format of the corresponding image file can be referred to in this recovery operation. Image data to be recombined may be provided from the storage module 110 or the anti-steganography module 140 . The analysis module 120 may receive necessary data from the storage module 110 in the process of performing the above task, and may be provided to provide the data obtained as a result of performing the above task to the storage module 110. there is.

The conversion module 130 may be interlocked with the analysis module 120 to access image data acquired by the analysis module 120 and information about the format of the image file. The conversion module 130 may be interlocked with the storage module 110 to access image data stored in the storage module 110 and information about the format of the image file. The conversion module 130 is provided to convert the image data into RGB format by performing a conversion method appropriate to the format according to the format of the image data. The format of the image data that the conversion module 130 can convert to RGB format may be JPEG / JFIF, JPEG 2000, GIF, BMP, PNG, PPM, PGM, PBM, PNM and HDR raster format, but must be It is not limited here. The conversion module 130 may receive necessary data from the storage module 110 in the process of performing the above task, and may be provided to provide the data obtained as a result of performing the above task to the storage module 110. there is.

The anti-steganography module 140 may be interlocked with the conversion module 130 to access RGB-type image data acquired by the conversion module 130 . The anti-steganography module 140 may be interlocked with the storage module 110 to access RGB-type image data stored in the storage module 110 . The anti-steganography module 140 is provided to convert image data in RGB form into YCbCr (Y'CbCr) form. The anti-steganography module 140 is provided to convert image data in a YCR format into an RGB format. For this conversion, the anti-steganography module 140 is prepared to calculate the formula below, but the formula that the anti-steganography module 140 can calculate is not limited to the formula below. Additional details are described below.

[Equation 1]

[Equation 2]

, and Values can be readily adopted by those skilled in the art wishing to practice the present invention. depending on the type of conversion , and The value can be set freely. , and The types of various conversions whose values are predetermined will be described later.

The anti-steganography module 140 may be provided to perform chroma subsampling.

The anti-steganography module 140 performs at least one of discrete cosine transform (DCT), discrete Fourier transform (DFT), and discrete wavelet transform (DWT) on data in the spatial domain. It is set up to be able to do one. The anti-steganography module 140 performs inverse discrete cosine transformation (Inverse-DCT), inverse discrete Fourier transformation (Inverse-DFT) and inverse discrete wavelet transformation ( It is provided to perform at least one of Inverse Discrete Wavelet Transformation (Inverse-DWT).

The anti-steganography module 140 is provided to perform quantization in the frequency domain. The anti-steganography module 140 is provided to perform dequantization on quantized image data.

5 shows each component of an anti-steganography device for carrying out an embodiment of the present invention.

Referring to FIG. 5 , according to an embodiment of the method for anti-steganography of an image file according to the present invention, the anti-steganography apparatus 100 may further include a disarming module 150 . The disarming module 150 is provided to remove metadata from the image data of the image file dismantled by the analysis module 120 . The harmless module 150 may be provided to receive image data from the storage module 110 and/or the analysis module 120 . The harmless module 150 may be provided to provide image data after removing metadata to the storage module 110 and/or the analysis module 120 .

The anti-steganography device 100 may be a computer that runs a program that performs an anti-steganography method of an image file, which will be described later, and a part or all of the computer further includes each module (disarming module 150). ) can be. In addition, the anti-steganography apparatus 100 may be a security server that must be passed through in order to access an in-house intranet from the outside.

<Anti-steganography method of image file>

Figure 2 shows each step of the anti-steganography method of the present invention.

Referring to FIGS. 1 and 2, the anti-steganography method of an image file according to the present invention includes a decomposition step (S120), a conversion step (S130), an anti-steganography step (S140), and a recovery step (S120'). do.

The decomposition step (S120) may be performed by the analysis module 120. In the decomposition step (S120), the analysis module 120 may obtain image data by decomposing the image file according to the specification of the format of the image file. The image file decomposed by the analysis module 120 may be provided by the storage module 110 . The analysis module 120 may acquire format information of the image file by analyzing the image file before performing the decomposition step (S120). The analysis module 120 may decompose the image file based on the acquired format information. The image data acquired by the analysis module 120 may exist in the form of an image data set for each cell. The analysis module 120 may provide the acquired image data to the storage module 110, the conversion module 130, and/or the harmless module 150.

The conversion step (S130) may be performed by the conversion module 130. In the conversion step (S130), the conversion module 130 may convert the image data acquired in the decomposition step (S120) into an RGB format. The conversion module 130 may receive image data from the analysis module 120 , the storage module 110 and/or the harmless module 150 . The conversion of image data performed by the conversion module 130 into an RGB format may be performed based on format information of an image file acquired by the analysis module 120 . The conversion module 130 may convert the corresponding image data into an RGB file by adopting an appropriate conversion method according to the format of the image file. The format of the image data that the conversion module 130 can convert to RGB format may be JPEG / JFIF, JPEG 2000, GIF, BMP, PNG, PPM, PGM, PBM, PNM and HDR raster format, but must be It is not limited here. The conversion module 130 may provide the converted RGB image data to the storage module 110 and/or the anti-steganography module 140 .

The anti-steganography step (S140) may be performed by the anti-steganography module 140. In the anti-steganography step (S140), the anti-steganography module 140 quantizes the RGB image data converted in the conversion step (S130) to neutralize the meta content inherent in the image file. The anti-steganography module 140 may receive RGB format image data from the storage module 110 and/or the conversion module 130 . In RGB type image data that has undergone quantization, values of some pixels may be approximated to values of adjacent pixels.

In this process, data stored as values of some pixels by steganography are lost. Therefore, even if the meta content is stored as a corresponding pixel value in the image file, some or all of it may be destroyed in the quantization process. After all, even if image data in RGB form that has undergone quantization is recovered, the meta content hidden in the image file is lost. That is, in this process, the meta contents hidden in the image file in the anti-steganography step (S140) are harmless. As a result, regardless of whether the image file is a Stego file, it may have an effect of fundamentally blocking an attack by a steganography technique.

An RGB image file in which meta content is disabled may be provided to the storage module 110 and/or the analysis module 120 .

The recovery step (S120') may be performed by the analysis module 120. In the recovery step (S120'), the analysis module 120 converts the RGB-type image data in which the meta contents are disabled through the anti-steganography step (S140) into the format of the corresponding image file, and converts the converted image data into the corresponding image data. It can be combined according to the format. The image file in RGB format in which the meta content is disabled may be provided to the analysis module 120 by the storage module 110 and/or the anti-steganography module 140 . In performing the recovery step (S120'), the analysis module 120 may utilize information about the format of the corresponding image file acquired for the decomposition step (S120).

3 is a subdivided anti-steganography step (S140) shown in FIG. 2 according to an embodiment of the present invention.

Referring to FIGS. 1 and 3, in one embodiment of the present invention, the anti-steganography step (S140) is the step of converting the image data in RGB format to YCR format (S141), to YCR format Converting the converted image data into the frequency domain (S142), performing quantization on the image data converted into the frequency domain (S143), and performing inverse quantization on the quantized image data (S144) , converting the inverse quantized image data into a spatial domain (S145), and converting the spatial domain converted image data into an RGB format (S146). Also in this embodiment, the anti-steganography step (S140) may be performed by the anti-steganography module 140.

In step S141, the anti-steganography module may convert image data in an RGB format into a YCR format. The anti-steganography module is ready to transform image data in the spatial domain into frequency domain by acquiring image data in the form of YCR. The YCBCR converted by the anti-steganography module contains Y'CbCr as well as YCbCr. Step S141 is performed according to [Equation 1] and [Equation 2] described above.

[Equation 1]

[Equation 2]

, and The value can be freely determined under the condition of [Equation 2], but several known conversion methods determine each value. ITU-R BT.601 conversion ( =0.299/ =0.587/ =0.114), ITU-R BT.709 conversion ( =0.2126/ =0.7152/ =0.0722), ITU-R BT.2020 conversion ( =0.2627/ =0.678/ =0.0593) and SMPTE 240M conversion ( =0.212/ =0.701/ = 0.087) is already known. However, the conversion method for performing step S141 is within the range satisfying [Equation 2]. , and It should be understood that any value may be employed. In step S141, the anti-steganography module satisfies [Equation 2] , and Y (Y), CB (Cb), and CR (Cr) values are found by calculating [Equation 1] with the values determined. For reference, the Y value is a value for luminance, and the CB and CIR are values for color difference.

Depending on the embodiment, the anti-steganography module 140 may further perform subsampling on the acquired YCR-type image data in consideration of characteristics of the human eye. Considering the characteristics of the human eye may mean trying to select each region, such as color, luminance, brightness, and saturation, in which the human eye has low sensitivity, as a sampling target section. By performing subsampling, the anti-steganography module 140 can reduce the computational load.

According to the embodiment, the anti-steganography module 140 further includes a step of performing chroma (chrominance) subsampling before converting the image data converted into YCBR form into a frequency domain (step S141') can do. Since the human eye is relatively less sensitive to color difference, even if chroma subsampling is performed, the effect on the cover image will be insignificant. By going through step S141', the anti-steganography module 140 can minimize damage to the image and reduce the computational load.

In step S142, the anti-steganography module may convert the image data converted into a YCR format into a frequency domain. The anti-steganography module may transform the image data converted into a YCBR form in the spatial domain into a frequency domain by performing at least one of discrete cosine transform, discrete Fourier transform, and discrete wavelet transform. Image data in the frequency domain obtained by the anti-steganography module 140 performing step S143 is Y (luminance), CB (color difference-blue), and CIR (color difference-red) frequency domains, respectively. It can mean a value expressed as .

In step S143, the anti-steganography module 140 may perform quantization on the image data converted to the frequency domain.

Meanwhile, FIG. 4 shows Y, CB, CR, and images according to the quantization table and different tables for implementing this embodiment.

According to some embodiments, the target of quantization performed in step S143 may be at least one of Y, CB, and CIR. As referred to in [Equation 1] and FIG. 4 introduced above, since Y, CB, and CIR are mutually independent values, quantization can also be performed independently. Meanwhile, human vision has a smaller sensitivity to chrominance information than luminance information. For this reason, chrominance information expressed in CB and CIR is easy to conceal meta contents through steganography. A person who intends to practice the present invention will be able to focus on this point and perform quantization only for CB and CIR without performing quantization for Y. Through this, it will be possible to effectively harmless the meta contents while minimizing damage to the cover image. Of course, since there are various types of images, in some cases, the degree of quantization is different for each of Y, CB, and CIR, or quantization is performed only for the rest without performing quantization for CB and/or CR, which is not Y. Anything will be possible.

As referenced in FIG. 4 , the anti-steganography module 140 may perform quantization such that the degree of quantization is determined according to a predefined table in step S143 according to an embodiment. The higher the degree of quantization, the greater the amount of data loss. (a), (b) and (c) of FIG. 4 show each quantization table and cover image according to the required quality of the cover image. (a) is a quantization table that means that quantization is not substantially performed at all. (b) is a quantization table for performing intermediate quantization. (c) is a quantization table for performing high-level quantization. Of course, it should be understood that the tables shown in (a), (b) and (c) mean only feature points in the spectrum of successively prepared quantization degrees, and that the quantization degrees can be freely selected within the spectrum. That is, a person who intends to implement the present invention can variously select a desirable quantization table as shown in FIG. 4 . Therefore, by determining the degree of quantization through this embodiment, it is possible to select an appropriate point (balance point) according to needs and circumstances between metadata detoxification and cover image preservation.

4 (b) and (c), in this embodiment, going further than the above-described embodiment, a predefined table for determining the degree of quantization in step S143 is image data converted to the frequency domain. It may be a table defined by assigning weights to each frequency domain of . Referring to the table and arrows in FIG. 4 , it can be seen that the higher frequency region is located in the lower right corner and the lower frequency region is located in the upper left corner. Referring to the tables of FIGS. 4 (b) and (c), it can also be seen that the quantization value is larger as it is located in the lower right corner. That is, it is also possible to prepare a quantization table to which different weights are assigned to each frequency domain. Through this embodiment, it is possible to select an optimal balance between metadata detoxification and cover image preservation by considering the need and situation more specifically. In addition, while further minimizing the damage of the cover image, it will be possible to more effectively harmless the meta contents.

In step S144, the anti-steganography module 140 may perform inverse quantization on the quantized image data. For the inverse quantization performed in step S144, the anti-steganography module 140 may implement all inverse quantization techniques that can be employed by a person skilled in the art at the time of filing of the present application. As the image data undergoes inverse quantization in step S144, a more natural and similar image file to the original cover image can be obtained when restoring the image file.

In step S145, the anti-steganography module 140 may convert the inverse quantized image data into a spatial domain. The anti-steganography module 140 acquires spatial domain image data by performing at least one of inverse discrete cosine transform, inverse discrete Fourier transform, and inverse discrete wavelet transform on frequency domain data in order to perform step S145. .

In step S146, the anti-steganography module 140 may convert the spatial domain-converted image data into an RGB format. To perform step S146, the anti-steganography module 140 may use [Equation 3]. , and The value used in step S141 is used.

[Equation 3]

For reference, [Equation 3] is a summary of [Equation 1] for R, G, and B. In step S146, the anti-steganography module 140 obtains RGB-type image data by calculating R, G, and B values using [Equation 3].

When step S146 is finished, the analysis module 120 performs a recovery step (S120') on the image data in RGB format.

FIG. 6 shows each step of an anti-steganography method according to an embodiment of the present invention performed by the anti-steganography apparatus of FIG. 5 .

Referring to FIGS. 5 and 6 , the method for anti-steganography of an image file according to another embodiment of the present invention further includes a disarming step (S150) of removing metadata included in the image file after the decomposition step (S120). can do. For this embodiment, the anti-steganography apparatus 100 may further include a harmless module 150, and the harmless step (S150) may be performed by the harmless module 150. The disarming module 150 may delete metadata from the image file decomposed by the analysis module 120 . By deleting the metadata, the path through which malicious code can enter can be more thoroughly blocked. However, depending on the embodiment, the metadata removed in the disarming step (S150) may be metadata selected according to a predefined policy. For example, a policy may be defined in advance so that information that is not determined to be harmful, such as the creator of an image file, the location and time of image file generation, etc., is not selected as a target for removal. In this way, in some cases, necessary data is not disabled, so that the user's convenience of the image file can be promoted. The detoxification step (S150) is preferably performed after the decomposition step (S120) of decomposing the image file and before the conversion step (S130), but it is also possible to be performed simultaneously with or after the conversion step (S130).

Another embodiment of the present invention may be a computer-readable storage medium in which a program for performing the above-described anti-steganography method of an image file is stored. Codes and code segments implementing the above program can be easily inferred by a computer programmer in the art. A computer-readable recording medium may include all types of recording devices storing data that can be read by a computer system. Examples of computer-readable recording media may include ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical disk, and the like. In addition, the computer-readable recording medium may be distributed among computer systems connected through a network, and may be written and executed as computer-readable codes in a distributed manner.

So far, the present invention has been looked at mainly with its preferred embodiments. Those skilled in the art to which the present invention pertains will be able to understand that the present invention can be implemented in a modified form without departing from the essential characteristics of the present invention. Therefore, the scope of the present invention should be construed to include various embodiments within the scope equivalent to those described in the claims without being limited to the above-described embodiments.

100: anti-steganography device
110: storage module
120: analysis module
130: conversion module
140: anti-steganography module
150: detoxification module
S120: disassembly step S120': recovery step
S130: conversion step
S140: Anti-steganography step

Claims (9)

In the anti-steganography method performed by the anti-steganography device
A decomposition step of obtaining image data by decomposing an image file according to a specification of a format of the image file;
a conversion step of converting the image data obtained in the decomposition step into an RGB format;
An anti-stereostat that neutralizes the meta content inherent in the image file by quantizing the RGB-type image data converted in the conversion step and destroying all or part of the meta-contents stored as pixel values in the RGB-type image data. ganography step; and
A recovery step of converting RGB-type image data in which the meta content is disabled through the anti-steganography step into the format, and combining the converted image data according to the format;
The anti-steganography step,
converting the image data in RGB format into YCbCr format;
converting the image data converted into a YCR format into a frequency domain;
performing quantization on the image data converted into a frequency domain;
performing inverse quantization on the quantized image data;
converting the inverse quantized image data into a spatial domain; and
Converting the image data converted to spatial domain into an RGB format; comprising
Anti-steganographic methods of image files. delete According to claim 1
The target of quantization performed in the anti-steganography step is at least one of Y, Cb, and Cr
Anti-steganographic methods of image files. According to claim 1
In the quantization performed in the anti-steganography step, the degree of quantization is determined according to a predefined table.
Anti-steganographic methods of image files. According to claim 4
The predefined table is a table defined by assigning a weight to each frequency domain of the image data converted to the frequency domain.
Anti-steganographic methods of image files. According to claim 1
The anti-steganography step
Performing chroma subsampling before converting the image data converted into YCR form into a frequency domain; further comprising
Anti-steganographic methods of image files. According to claim 1
Further comprising a detoxification step of removing metadata included in the image file after the decomposition step.
Anti-steganographic methods of image files. According to claim 7
The metadata removed in the detoxification step is metadata selected according to a predefined policy.
Anti-steganographic methods of image files. A computer-readable storage medium in which a program for performing the method of any one of claims 1 or 3 to 8 is stored.