JP6172745B2 - Digital watermark detection method - Google Patents

Description

  The present invention relates to a method for reading embedded information from distributed content using a digital watermark technique in a system that enables copyright management in a content distribution service.

  Currently, CDs, DVDs, Blu-ray discs, etc. all sell the same content. Once content is illegally copied and leaked to the Internet, it cannot be prevented.

  If the copy restriction used in BS, CS, etc. and the copyright management by the B-CAS card are displayed on the display, they can be copied at a high resolution with a commercially available device. Copyright management becomes essential as the content becomes higher resolution.

  For copyright management, information is embedded in content such as still images using digital watermark technology. When information is embedded in the entire image, if the entire image is analyzed, the embedded information can be extracted using a digital watermark technique. However, user information cannot be obtained from illegal copies that are leaked on the Internet.

  In order to shorten the processing time, a technique has been proposed in which an image is divided and information is embedded using a digital watermark technique (see, for example, Patent Documents 1 and 2). In this case, user information can be obtained, but even if the entire image is analyzed, all information embedded in each area cannot be read out. For this reason, when an image is divided and information is embedded in each area using digital watermark technology, the division pattern is first acquired, and the image is divided based on the division pattern and then embedded in each area. It is essential to read the information.

JP 2006-086786 A JP 2007-251268 A

  In order to increase the effect of preventing illegal copying, it is necessary to identify the spill source in a short time. However, if the embedded information is read using the division type digital watermark technology in a state where the division pattern is unknown, there is a problem that the information cannot be read or it takes a long time even if the information can be read. .

  Therefore, an object of the present invention is to make it possible to read out information embedded using a digital watermark technique in a short time even if the division pattern is unknown.

The digital watermark detection method of the present invention is:
A division pattern acquisition procedure for reading information embedded in an image using a digital watermark technique by analyzing the entire image, and acquiring a division pattern of the image embedded in the image using a digital watermark technique;
An individual decoding procedure for dividing the image into a plurality of regions using the acquired division pattern, and reading information embedded in the region using a digital watermark technique using an analysis process different from the division pattern acquisition procedure ;
In order.

    Prior to the division pattern acquisition procedure, a combination of information embedded for each area and user information are associated with each other and stored in a storage unit. After the division pattern acquisition procedure, the information is embedded for each area. It may further include a user specifying procedure for acquiring user information that matches a combination of information stored from the storage unit.

Prior to the division pattern acquisition procedure, key frame information acquisition procedure for reading from the video key frame information indicating which of the plurality of images included in the video is embedded with information using digital watermark technology Further comprising
In the division pattern acquisition procedure, a division pattern of each image indicated by the key frame information may be acquired.

  According to the present invention, even if the division pattern is unknown, it is possible to read out the information embedded using the digital watermark technique in a short time.

An example of the division pattern of an image is shown. 10 shows a first example of an image division pattern in Embodiment 3. An example of embedding an image in each divided area in the third embodiment will be described. 10 shows a second example of an image division pattern in Embodiment 3. FIG. An example of the relationship between the area ratio of an image and the ease of detection of a digital watermark when it is performed by two types of digital watermark methods, WM1 and WM2. An example of the relationship between the area ratio of an image and the ease of detection of a digital watermark when it is performed with three types of intensity of S = 1, 2, and 3. An example of the relationship between the area ratio of an image and the ease of detecting a digital watermark when the image is changed. The 1st example of the key frame pattern in Embodiment 4 is shown. 20 shows a second example of a key frame pattern in the fourth embodiment. The 1st example of the division | segmentation pattern in Embodiment 4 is shown. The 2nd example of the division | segmentation pattern in Embodiment 4 is shown. The first example of embedding information in Embodiment 4 is shown. 10 shows a second example of embedding information in Embodiment 4.

  Embodiments of the present invention will be described with reference to the accompanying drawings. The embodiments described below are examples of the present invention, and the present invention is not limited to the following embodiments. In the present specification and drawings, the same reference numerals denote the same components.

(Embodiment 1)
The invention according to the present embodiment is premised on a service for embedding different information for each user using a digital watermark technique. In this service, even if the distributed content is illegally distributed, it is possible to identify the user who first receives the content usage permission from the information embedded using the digital watermark technology. As a result, it is possible to expect a deterrent effect of illegal activities such as publishing personal content on the web.

  The present invention is based on the premise that digital watermark technology is used in such a copyright management system, and information embedded in content using digital watermark technology can be detected at high speed even when the user's information is unknown. Realize the method.

  First, information embedding using a digital watermark technique will be described. In this embodiment, as shown in FIG. 1, an image is divided into three areas A1, A2, and A3, and information is embedded in each area using a digital watermark technique. Using information embedded by using digital watermark technology, it is possible to specify to whom the content being reproduced is a licensed content.

  Assume that one image is divided into a plurality of regions, and information is embedded in each region using a digital watermark technique. When trying to read the information embedded using the digital watermark technology for the entire image, the information embedded in all the areas is not read out, but only the information embedded in one of the divided areas Is read out. That one area depends on the software that embeds the information using digital watermark technology.

  Therefore, when the entire image is analyzed using software capable of reading the embedded information, the divided pattern is embedded so that the divided pattern is read first. For example, if the software first reads the area A1 shown in FIG. 1, the division pattern is embedded in A1. Then, information related to contents and user information are embedded in the area A2 or A3. In this state, the content is provided to the user.

  The area to be read first depends on the software analysis process. Therefore, in the case of software that preferentially reads a position with a small coordinate such as the left or the top, the division pattern is embedded at the position on the left side or the top side of the image. In the case of software that preferentially reads information embedded using a digital watermark technique over a wide area, a divided pattern is embedded in an area having a large area. If the software preferentially reads a digital watermark with high strength, the strength of the digital watermark in which the division pattern is embedded is increased.

  Note that the content may be a still image or a moving image including a plurality of still images. Further, the user information is not only embedded as character information, but may be a figure. For example, a division pattern may be used for user information.

  The digital watermark detection method according to the present embodiment includes a division pattern acquisition procedure, an individual decoding procedure, and a user specifying procedure in this order. The user specifying procedure may be after the division pattern acquisition procedure, and may be performed at any timing before, simultaneously with, or after the individual decoding procedure.

  In the division pattern acquisition procedure, information embedded in the image is read out using the digital watermark technique by analyzing the entire image. At this time, if analysis is performed using appropriate software, a division pattern which is information embedded in the area A1 is read.

  In the individual decoding procedure, using the acquired division pattern, the image is divided into A1, A2, and A3, and information embedded in each image is read by analyzing each divided image. When the user information is embedded in the area A2 or A3, the user information can be acquired by reading the information embedded in the A2 or A3.

  In the user specifying procedure, the user who is permitted to view the content is specified using information embedded in each image. For example, the user is specified using information read from the area A2 or A3.

  At this time, the user may be specified using a combination of information embedded in the areas A1, A2, and A3. For example, the image W1 or W2 is embedded in the area A1, the image W3 or W4 is embedded in the area A2, and the image W5 or W6 is embedded in the area A3. In this case, the combination of images embedded in each region is changed for each user, and the combination of images is associated with the user. As a result, as shown in FIG. 1, when the images embedded in the areas A1, A2 and A3 are W1, W4 and W5, it is specified that the user of this content is a registered user in advance. can do.

  In this embodiment, the example in which the division pattern of the entire image is embedded in the area A1 to be decoded first is described. However, the range of the area A2 is embedded in the area A1, and the area A3 in the area A2 to be decoded next. The range may be embedded.

  As described above, the invention according to the present embodiment embeds information on the division pattern of the entire image in the information detected in the first embedded information detection process. Therefore, if the image is divided in advance according to the division pattern, the remaining embedded information can be reliably detected.

(Embodiment 3)
Consider a case where one image is divided into sub-images Wna, Wnb, Wnc, Wnd,... As shown in FIG. As shown in FIG. 3, the image is duplicated, images A, B, C, and D are embedded in one of the sub-images Wna, Wnb, Wnc, and Wnd, and the images G, Embed H, I, and J. Then, the user is specified using a combination of the sub-image area and the embedded information.

  In the divided digital watermark, when the user is specified, the watermark information A to J is linked with the user data, so that the detection is easy. It takes time to detect watermark information when the user cannot be identified. Since it cannot be accurately detected without the division information, the division is squeezed out.

If the division method is not known, when detecting the watermark information as a not divided, as shown in FIG. 4, vertically to R _{A = A} L _{/ A} sub-images W of 2 divided by the ratio of _{R L} and W _R When divided, it is detected according to the following rules.
As indicated by WM1 in FIG. 5, depending on the type of digital watermark, the digital watermark with the larger area of the sub-image is detected. As indicated by WM2 in FIG. 5, depending on the type of digital watermark, if there is a certain amount of watermark area, the watermark information of the left sub-image is detected.
As shown by S = 3 in FIG. 6, when the watermark strength is high, the left side is detected even if the left side area is considerably small.
Note that, as shown in FIG. 7, the change in the image has little influence on the ease of detection of the digital watermark.

(Embodiment 4)
In the present embodiment, a case where the content is a moving image will be described. The content is a 4K moving image (resolution 4096 × 2160), and character information of 8 characters is embedded in an area of 256 × 256 or more of the divided sub-images. In the case of an image synthesized by two or more sub-images, the watermark of the image with the largest area is first read out.

A procedure for embedding a digital watermark in this embodiment will be described.
Step S101: A digital watermark key frame KWM, which is a key frame for embedding a digital watermark, is determined for each user. For example, as shown in FIG. 8, a frame of a moving image and a time code (TS1, TS2,... TSn) are associated with each other, and time codes TS1, TS4,... TSm,. When TSp is the key frame k, TS4,... TSm,... TSn are determined as the digital watermark key frame KWM. Here, the key frame is, for example, an I frame that is a reference for motion prediction. In FIG. 8, B and P frames are denoted by G as other frames.

  Here, the digital watermark key frame in which the digital watermark is embedded may be different for each user, or may be different for each content, but may be common to all users in the case of the same content. For example, even if the content is the same as the content shown in FIG. 8, if the users are different, as shown in FIG. 9, instead of TS4 and TSn, TS1 and TSp may be an electronic watermark key frame KWM.

  In addition, as a method for determining an electronic watermark key frame, any one of predetermined electronic watermark key frame patterns may be selected. For example, the electronic watermark key frame pattern k1 shown in FIG. 8 is prepared. In the key frame pattern k1, the time code TS4, TSm, TSn is used as a digital watermark key frame. The key frame pattern k2 is a pattern in which the time codes TS1, TSm, and TSp are electronic watermark key frames.

Step S102: Extract the digital watermark key frame KWM from the moving image based on the time code included in the header information defined by the moving image format.
Step S103: An image division pattern is selected. 10 and 11 show examples of the division pattern according to the present embodiment. FIG. 10 shows the division pattern p1, and FIG. 11 shows the division pattern p2.
Step S104: Divide the image using the selected division pattern, and create a divided sub-image. For example, the image is divided into sub-images S1, S2, S3, and S4 using the division pattern p1.

  Step S105: An electronic watermark is embedded in each sub-image. At this time, based on the pattern information p1 and the watermark information 1234, the electronic watermark is embedded in the sub-image s1 using p1-1234 as the embedded information. Based on the pattern information p1 and the watermark information 5678, an electronic watermark is embedded in the sub-image s1 using p1-5678 as embedded information. An electronic watermark is embedded in the sub-image s2 based on the pattern information p1 and the watermark information 2341, with p1-2341 as the embedded information. An electronic watermark is embedded in the sub-image s2 based on the pattern information p1 and the watermark information 6785, with p1-6785 as embedded information. Similarly, the digital watermark shown in FIG. 12 is embedded in each sub-image after the sub-image s3.

  Here, it is preferable to embed not only the pattern information but also the pattern information of the key frame containing the digital watermark in the first digital watermark key frame KWM. The first digital watermark key frame KWM is a TS4 key frame in FIG. 8, and a TS1 key frame in FIG. In the case of the key frame pattern k1 shown in FIG. 8, as shown in FIG. 13, k1-p1-1234 and k1-p1-5678 are embedded in the sub-image s1 of the digital watermark key frame KWM of the time code TS4. In addition, the sub image in which the key frame pattern information is embedded is not limited to the sub image s1 that is read first, but may be embedded in all the sub images as shown in FIG.

  The key frame in which the pattern information is embedded is not limited to the first one of the digital watermark key frames KWM, but may be all the digital watermark key frames KWM. As a result, the key frame pattern information can be read even if the copy is partially illegal.

  In the present embodiment, a key frame information acquisition procedure is further included before the division pattern acquisition procedure. In the key frame information acquisition procedure, key frame information indicating which of a plurality of images included in a moving image is embedded with information using a digital watermark technique is read from the moving image. For example, the key frame pattern embedded in the key frame is read. For example, in the case of the moving picture shown in FIG. 8, k1-p1-1234 is read from the sub-image s1 of the time code TS4. Thereby, it can be acquired that the key frame pattern k1 shown in FIG.

  The present invention can be applied to the information communication industry such as a content distribution service.

A1, A2, A3, Wna, Wnb, Wnc, Wnd: Image areas W1, W4, W5, A, B, C, D, G, H, I, J: Images TS1, TS2, TS3, TS4, TSn, : Time code

Claims (3)

A division pattern acquisition procedure for reading information embedded in an image using a digital watermark technique by analyzing the entire image, and acquiring a division pattern of the image embedded in the image using a digital watermark technique;
An individual decoding procedure for dividing the image into a plurality of regions using the acquired division pattern, and reading information embedded in the region using a digital watermark technique using an analysis process different from the division pattern acquisition procedure ;
An electronic watermark detection method comprising: Prior to the division pattern acquisition procedure, a combination of information embedded for each area and user information are associated with each other and stored in a storage unit,
The digital watermark detection method according to claim 1, further comprising a user specifying procedure for acquiring, from the storage unit, user information that matches a combination of information embedded for each region after the division pattern acquisition procedure. . Prior to the division pattern acquisition procedure, key frame information acquisition procedure for reading from the video key frame information indicating which of the plurality of images included in the video is embedded with information using digital watermark technology Further comprising
The digital watermark detection method according to claim 1, wherein in the division pattern acquisition procedure, a division pattern of each image indicated by the key frame information is acquired.

Images