KR100536641B1 - Method and system for embedding fingerprint into digital contents - Google Patents

Abstract

The present invention relates to a method and apparatus for embedding a fingerprint in digital content. In selecting a seed value for generating a pseudo random number, a different seed value is selected for different transaction information. The pseudo random number corresponding to the transaction information is generated according to the selected seed value. The generated pseudo random number is inserted into the digital content. The selected seed value is also inserted into the digital content as part of the fingerprint information. Accordingly, fingerprint information can be accurately detected without interference between user information even in collusion attacks such as average attack or mosaic attack.

Description

METHOD AND SYSTEM FOR EMBEDDING FINGERPRINT INTO DIGITAL CONTENTS}

The present invention relates to a method and apparatus for embedding a fingerprint in digital content, and more particularly, to a fingerprint insertion method and apparatus that are robust to collusion attacks such as an average attack or a mosaic attack.

Recently, wired and wireless digital content providers have been struggling to come up with measures to redistribute illegal content by users. Digital content, unlike computer software or hardware that can be protected from copying and redistribution using serial numbers or hard-locks, has very little protection. The application is also very difficult.

Watermarking and encryption technologies are being studied in order to protect copyright and redistribution of digital contents. Although encryption technology has already been standardized and standards have been specified, the watermarking sector has not adopted standards despite some standardization movements. Although watermarking technology is not yet as certified as encryption technology, many people are interested in research because of its high utilization.

The application fields of watermarking are diverse from broadcasting monitoring, owner information display, owner certification, authentication, copy control, and confidential information transfer. The following requirements are required for watermarking to be applied in such applications.

(1) Toughness against attack:

① Robustness against lossy compression, A / D and D / A conversion, geometric transformation, Stirmark attack

② resistance to random attacks

③ Robustness against active attacks such as attempts to delete watermarks

④ Robustness against passive attacks, such as inducing false detection of watermarks

⑤ Robustness against collusion attacks such as inducing false detection using multiple copies

⑥ Robustness against counterfeit attacks, such as adding watermarks to other information

(2) Precision: It must be made unrecognizable to human sense organs.

(3) Computation amount: For broadcast monitoring, for example, the watermark should be able to be extracted quickly, so the computation amount should not be excessively complex.

(4) False detection rate: For reliability, it is desirable to reduce the false detection rate by using Error Correcting Code (ECC) or Cyclic Redundancy Check (CRC).

The essential requirements of the watermarking mentioned above can be seen that the requirements are more diverse and require higher robustness than the initial watermarking technology.

On the other hand, another kind of copyright marking such as watermarking is a fingerprinting technique.

Fingerprinting is one application of watermarking technology. The difference between watermarking and fingerprinting is the variety of information that is inserted. In general, watermarking inserts only the author's information, that is, only one piece of information, to prove who is the original author, but in fingerprinting, not only the author's information, but also the user's information such as distributors, redistributors, and end consumers. Insert it. If a user arbitrarily delivers content to a third party without permission, it is considered illegal distribution and identifies the illegal distributor based on the detected fingerprint information. That is, by applying the fingerprinting technology, it is possible to follow up who has illegally distributed the contents and hold the responsibility.

Hereinafter, a fingerprinting process and a process of tracking an illegal user through the same will be described with reference to FIG. 1. 1 is a diagram illustrating a process of tracking an illegal user by inserting a fingerprint into digital content.

Referring to FIG. 1, the author 10 provides digital content to the content distributor 20, and the content distributor 20 inserts fingerprint information into the digital content to present the fingerprinted digital content to the user 30. to provide. The fingerprint information inserted in the digital content includes all the information of the author 10, the content distributor 20, and the user 30.

The authentication center 60 provides an authoring license to the author 10, a distribution license to the content distributor 20, and a user license to the user 30, respectively.

In the distribution process of such digital content, there may be a case where the digital content is illegally used by the illegal user 50 for the reason that the digital content is hacked by the hacker 40.

When the illegally used digital content undergoes a fingerprint detection process in the fingerprint information detector 70, it is checked 80 whether the illegally used content is detected through the detected transaction information (i.e., user information).

The check of the illegal use content is performed by transmitting the detected transaction information to the authentication center 60 and receiving legitimate user information from the authentication center 60. When it is confirmed that the content is illegally used, the legal user 30 is notified of the illegal distribution, and illegal distribution by the hacker 40 and the illegal user 50 is blocked.

That is, such fingerprinting can track illegal distributors, which was impossible in the conventional watermarking or encryption technology. Although encryption technology is widely used to prevent illegal distribution, there is no way to arrange the distribution of decrypted content.

Fingerprinting technology allows users to track who is distributing digital content by inserting transaction information into the decrypted content, and by adding psychological measures to prevent illegal distribution, Allow culture to be created.

The above-described fingerprinting is more demanding in attack robustness and data insertion amount than in general watermarking requirements. In general, watermarking for copyright protection does not require a large amount of data insertion so that even a 1-bit watermark called a key method can be protected.

In addition, in watermarking, one or a group of information is inserted into one piece of content, thereby making it robust to collusion attacks such as an average attack or mosaic attack required for fingerprinting.

However, fingerprinting requires inserting information from copyright holders to intermediate distributors and end consumers step by step, so a large amount of data insertion is required, and content containing multiple pieces of information may show a weak point in average attack or mosaic attack. There is nothing else. Such weaknesses of the fingerprinting technique against the average attack or the mosaic attack will be described later with reference to FIG. 2.

In summary, in the fingerprinting, the following essential requirements must be added or strengthened as compared to watermarking technology.

(1) High Reliability: Must be able to reliably screen out legitimate users and minimize user duplication in one content.

(2) Low Total Cost: Fingerprinting systems should be able to perform with fewer resources to operate as part of other systems.

(3) Resilience: It must be difficult to delete without complete knowledge of fingerprinting techniques.

(4) Transparency: Transparency should be used for related software related to fingerprinted content.

(5) Partial protection: Ideally, the distributor should be able to be found by detecting only a small portion of the content.

That is, as described above, the fingerprinting technique is also an extended application field of the watermarking technique, but higher precision, robustness, and toughness are required to implement the technique.

In particular, public attacks such as an average attack that takes an average of a plurality of digital contents fingerprinted with different information to create an illegal copy or a mosaic attack that creates an illegal copy by dividing a single digital content into several areas ( Collusion Attack is a very deadly attack method for fingerprinting techniques.

Such average attack and mosaic attack are shown in Figs. 2A and 2B. FIG. 2A illustrates an average attack on digital content, and FIG. 2B illustrates a mosaic attack on digital content.

The problem in the average attack or mosaic attack as shown in Figs. 2A and 2B is that the inserted fingerprint signal is canceled due to the characteristics of such an attack. That is, unlike watermarking technology, fingerprinting inserts various user information such as author, distributor, redistributor, and final consumer information. Therefore, user information inserted when collusion attack is applied to digital content with such various information is inserted. Interference occurs between them.

Such information interference occurs due to the characteristics of collusion attacks such as average attacks and mosaic attacks.

That is, in the case of the average attack, as shown in FIG. 2A, the average of the plurality of digital contents is averaged, thereby causing interference and weakening of the inserted fingerprint data. For example, if an average attack of 10 digital contents with different user information is inserted, interference is generated between user information included in each of the 10 digital contents, and the signal is reduced to 1/10, thereby weakening the signal and extracting information. There is a difficulty.

In the case of mosaic attack, as shown in FIG. 2B, one piece of digital content is inserted into which different user information is inserted, and then pieces are fragmented, and the inserted different user information (ie, author, distributor, redistributor, etc.) is used. ) Interference occurs.

For example, if a random number "10000100000" of user A is mosaiced with a random number "1111011110" of user B, an entirely new random number of "1000011110" may be detected.

Since information interference occurs between user information inserted by collusion attack, there is a problem that "high reliability", which is an essential requirement of fingerprinting, is difficult to be guaranteed. There is a problem that it is possible to judge a completely different person as an illegal user when the information inserted when detecting the fingerprint by a collusion attack is detected even if an error occurs even if only one bit.

The present invention has been made to solve the problems described above, the object of the present invention is fingerprinting that can accurately detect the fingerprint information without interference between user information even in a collusion attack such as an average attack or mosaic attack To provide a way.

In order to achieve the above object, according to an aspect of the present invention, there is provided a method for embedding a fingerprint in digital content, the method comprising the steps of: (a) selecting a seed value for generating a pseudo random number; (b) generating a pseudo random number corresponding to transaction information according to the selected seed value; And (c) inserting the generated pseudo random number into the digital content, wherein the selection of the seed value in step (a) is selected with different seed values for different transaction information. To provide.

Here, it is preferable that the pseudo random numbers generated by the different seed values are orthogonal to each other.

The method also preferably includes inserting the selected seed value into the digital content as part of fingerprint information.

In addition, the method preferably further comprises inserting a sync bit into the digital content as part of the fingerprint information for aligning the position of the fingerprint information to be inserted.

In addition, the method preferably further comprises inserting a Cyclic Redundancy Check (CRC) or Error Correcting Code (ECC) into the digital content as part of fingerprint information.

In addition, the transaction information is preferably user information such as author, distributor, redistributor or end consumer.

According to another aspect of the invention the present invention provides a seed value selection unit for selecting a seed value for generating a pseudo random number; A pseudo random number generator for generating a pseudo random number corresponding to the transaction information according to the selected seed value; And a fingerprint insertion unit for inserting the generated pseudo random number into digital content, wherein the selection of the seed value provides an apparatus for inserting a fingerprint into digital content selected with different seed values for different transaction information. .

Hereinafter, exemplary embodiments of the present invention will be described with reference to the accompanying drawings.

First, a structure of data inserted as fingerprint information in the present invention will be described with reference to FIG. 3. 3 is a diagram illustrating data inserted into digital content as fingerprint information in the present invention.

The sync bit (Sync. Bit: 100) is for matching the positional synchronization of data, which is fingerprint information to be inserted. In other words, regardless of the fingerprint information to be inserted, it means that the peak value exists at the time of the cross correlation measurement for detecting the fingerprint information at a specific position. By inserting the sync bit, even when the information of the fingerprint-embedded digital content is shifted, it is possible to restore it when the fingerprint is detected.

Fingerprinting Seed 110 is a seed value for generating a pseudo random number corresponding to the transaction information, which is different from the seed value of fingerprint information for other transactions as described in detail with reference to FIG. 4 below. Is randomly selected.

Transaction information 120 is fingerprint information for identifying a user, that is, a distributor, a redistributor, an end consumer, and the like, and may be, for example, a name of a person or an information ID of a database.

The CRC (Cyclic Redundancy Check: 130) is designed to increase the reliability of the detected information.It is 99.99998% reliable by using the CRC, which extracts the wrong user information so that a good user can frame the illegal distributor. Prevent it.

In addition to the CRC, an ECC (Error Correcting Code) may be used as a method of improving the reliability of the above-described detection information. However, since ECC requires a larger number of bits than CRC, more data may be inserted into the content, which may cause a problem in robustness. Therefore, CRC and ECC should be used selectively in consideration of the environment.

In the present invention, the data set including the sync bit 100, the fingerprinting seed 110, the transaction information 120, and the CRC 130 shown in FIG. 3 as the fingerprint information corresponding to the respective user information. It will be embedded in digital content.

4 is a block diagram illustrating a fingerprint insertion process according to the present invention.

Referring to FIG. 4, the fingerprinting seed 200 is randomly selected among N seed values by different seed values for each user in the seed value selection unit 210. In addition, the selected seed value is included in fingerprint information for a corresponding user later to configure a data set.

The pseudo random number generator 220 generates a pseudo random number corresponding to the transaction information 230 according to the seed value selected by the seed value selector 210. In the present invention, since a different transaction, that is, a different seed value is selected according to different users and a pseudo random number is generated accordingly, even if a collusion attack such as an average attack or a mosaic attack occurs in the fingerprint-embedded digital content later, Interference between information does not occur during print detection.

That is, pseudorandom numbers generated by different seed values have orthogonal properties, and thus, fingerprint information inserted by different seed values does not interfere with each other and thus is robust to collusion attacks. .

For example, the first transaction information generated and inserted into the pseudo random number generated by the seed 1 is mixed with the average transaction or mosaic attack with the second transaction information generated and inserted into the pseudo random number generated by the seed 2 However, this is only treated as noise in the spread spectrum technique and is not affected by the detection of information at all. This is based on the fundamental feature that spread spectrum techniques are robust to noise.

The pseudo random number generated by the pseudo random number generation unit 220 is inserted into the digital content 240 by the fingerprint inserting unit 250, thereby generating the digital content 260 with the fingerprint inserted therein. The fingerprint embedded digital content 260 has the data structure of FIG. 3 described above.

In this way, each time a different transaction occurs, the fingerprint information is inserted by the same process with different seed values.

Hereinafter, a fingerprint detection process of the digital content with a fingerprint inserted in the above-described manner will be described.

First, the seed value inserted from the digital content is extracted. This is possible because the seed value itself is included in the fingerprint information during the fingerprint insertion process.

A pseudo random number is generated according to the extracted seed value, and the embedded fingerprint information (ie, transaction information) is detected by measuring a cross correlation between the digital content inserted into the fingerprint and the generated pseudo random number. In this process, the original digital content is not necessary.

While the invention has been particularly shown and described with reference to the above embodiments, it has been used merely for the purpose of illustration and those of ordinary skill in the art to which the invention pertains may have as defined in the appended claims. It should be understood that various modifications may be made without departing from the spirit and scope.

As described above, according to the present invention, there is an effect of providing a fingerprinting method capable of accurately detecting fingerprint information without interference between user information even in collusion attacks such as an average attack or a mosaic attack.

1 is a diagram illustrating a process of tracking an illegal user by inserting a fingerprint into digital content.

FIG. 2A illustrates an average attack on digital content, and FIG. 2B illustrates a mosaic attack on digital content.

3 is a diagram illustrating data inserted into digital content as fingerprint information in the present invention.

4 is a block diagram illustrating a fingerprint insertion process according to the present invention.

Explanation of symbols on main parts of drawing

100: sync bit 110: fingerprinting seed

120: Transaction Information 130: CRC

200: fingerprinting seed 210: seed value selection unit

220: pseudo random number generator 230: transaction information

240: digital content 250: fingerprint insertion unit

260: fingerprint embedded digital content

Claims (12)

In the method of embedding a fingerprint in digital content, (a) selecting seed values for generating pseudo random numbers; (b) generating a pseudo random number corresponding to transaction information according to the selected seed value; And (c) inserting the generated pseudo random number into the digital content, Fingerprint insertion method characterized in that the selection of the seed value in the step (a) is selected with different seed value for different transaction information. The method of claim 1, And the pseudo random numbers generated by the different seed values are orthogonal to each other. The method of claim 1, And inserting the selected seed value into the digital content as part of fingerprint information. The method of claim 1, And inserting a synchronization bit into the digital content as part of the fingerprint information for aligning the positional synchronization of the inserted fingerprint information. The method of claim 1, And inserting a cyclic redundancy check (CRC) or an error correcting code (ECC) into the digital content as part of the fingerprint information. The method of claim 1, And the transaction information is user information such as author, distributor, redistributor or end consumer. A seed value selection unit for selecting a seed value for generating a pseudo random number; A pseudo random number generator for generating a pseudo random number corresponding to the transaction information according to the selected seed value; And A fingerprint insertion unit for inserting the generated pseudo random number into digital content, And the seed value is selected as a different seed value for different transaction information. The method of claim 7, wherein And a pseudo random number generated by the different seed values is orthogonal to each other. The method of claim 7, wherein And the fingerprint inserter further inserts the selected seed value into the digital content as part of fingerprint information. The method of claim 7, wherein The fingerprint inserting unit may further include inserting a synchronization bit to the digital content as part of the fingerprint information to match the positional synchronization of the fingerprint information to be inserted into the digital content. Device. The method of claim 7, wherein And the fingerprint inserting unit further inserts a cyclic redundancy check (CRC) into the digital content as part of fingerprint information. The method of claim 7, wherein And the transaction information is user information such as author, distributor, redistributor or end consumer.