JP2002135736A - Electronic watermarking device for moving picture - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an electronic watermarking device for a moving picture capable of simply and accurately determining whether watermarking data are embedded while the data are in a compressed form such as MPEG. SOLUTION: Bit information is comprised of a pair of different blocks. When embedding bit information 21 in an image frame F, for example, the bit information 21 is comprised of a pair of blocks 21a, 21b in which watermark data of different two portions are embedded, or a pair of a block 21a in which the watermark data are embedded and a block 21b in which a watermark data are not embedded. On the other hand, when detecting the bit information 21, it is detected whether the bit information 21 embedded in the frame F is 1 or 0 by comparing between the sum of a set of DCT coefficients corresponding to the block 21a and the sum of a set of DCT coefficients corresponding to the block 21b. For bit information 22, embedding and detection are also executed in a similar manner.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a moving picture digital watermarking apparatus, and more particularly to a moving picture digital watermarking apparatus capable of protecting copyright in distribution of moving picture contents.

[0002]

2. Description of the Related Art With the spread of the Internet and the like, it is expected that opportunities for communicating or distributing moving images in digital form will increase. Since a large number of image frames exist, it is necessary to embed the same watermark in all the frames or to embed the watermark in a specific frame.

As a method of embedding a watermark in a moving image, M
A method of embedding a watermark by manipulating a PEG motion vector has been proposed. According to this method, a watermark can be extracted without specifying one frame in an image frame, which is useful. As a patent related to the present invention, there is, for example, Japanese Patent Application Laid-Open No. Hei 10-178624.

As another method of embedding a watermark in a moving image, for example, as described in Japanese Patent Application Laid-Open No. 2000-175161, an image frame is divided into blocks, and the extracted blocks are subjected to DCT processing. There is a method in which the absolute value of at least one of the coefficients obtained by the DCT processing is increased to embed watermark data.

[0005]

However, the above-described method of operating the MPEG motion vector has a problem that the watermark data is easily removed by once decoding the compressed data and then re-encoding the compressed data. there were.
In addition, since moving images are generally distributed in an MPEG-compressed format, it is necessary to determine a watermark in the MPEG-compressed format. Had not been.

The latter conventional method solves the above-mentioned problem. However, the conventional method does not give special consideration to a method for simply detecting whether or not watermark data is embedded. Was.

SUMMARY OF THE INVENTION An object of the present invention has been made in view of the above-mentioned prior art, and simply and accurately determines whether or not watermark data is embedded in a compressed format such as MPEG. It is an object of the present invention to provide a moving image digital watermarking apparatus capable of performing the above. Another object of the present invention is to provide a moving image digital watermarking apparatus capable of detecting watermark data embedding at high speed.

[0008]

In order to achieve the above-mentioned object, the present invention extracts a plurality of blocks at predetermined positions from an image frame obtained by dividing a block, and extracts the plurality of extracted blocks. Means for performing DCT processing on the blocks, and embedding watermark data in at least one coefficient value of each of the plurality of blocks subjected to the DCT processing;
Watermark data embedding means for embedding 1-bit information is provided by a pair of blocks in which the watermark data is embedded at two different positions, or a pair of blocks in which the watermark data is embedded and blocks in which the watermark data is not embedded. There is a first feature in the point that it is provided.

[0009] The present invention also relates to a DCT in a DCT coefficient set of one of the blocks of the pair.
The point where the absolute value sum of the coefficient value is compared with the absolute value sum of the DCT coefficient values in the DCT coefficient set of the other block, and it is determined that the larger bit of the absolute value sum is embedded. There is a second feature.

According to the first and second features, even if the prediction is incorrect due to the original nature of the image,
The detection accuracy of the bit information embedded in the image frame can be improved.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below in detail with reference to the drawings. FIG. 1 is a block diagram showing a schematic configuration of a watermark data embedding device according to the present invention.

In the figure, a block dividing section 1 divides, for example, the image frame shown in FIG. 2A into a block bij of 8 × 8 pixels. The block extracting unit 2 extracts a macroblock (16 × 16 pixels) mxy specified from the template 9. The DCT unit 3 performs a discrete cosine transform (DCT) for each block mxiyi included in the macroblock mxy. DC
The T coefficient extraction unit 4 extracts a DCT coefficient x of one block mx2y2 (see FIG. 2B) in the DCT-processed macroblock mxy according to the instruction of the template 9.
Here, the template 9 provides information indicating a position at which watermark data to be embedded is described later. The block extracting unit 2 extracts a macroblock at a position where the watermark data is embedded, and the DCT coefficient extracting unit 4 The DCT coefficient x in which the watermark data is embedded is extracted.

The watermark data embedding unit 5 embeds watermark data in the DCT coefficient x. For example, the DCT coefficient x
Is changed to a coefficient x ′ obtained by the following equation to embed watermark data. x ′ = sign (x) * (| x | + a) where x is the original coefficient value, a is the watermark data, that is, the embedding value, and sign (x) is the sign of x.

That is, the watermark data embedding unit 5
Functions to change the DCT coefficient x in a direction of increasing the absolute value while maintaining the + and-signs of the DCT coefficient x. For example, assuming that the watermark data a = 5.0, if the DCT coefficient x is +15.7, x ′ = + 20.7
When the DCT coefficient x is −15.7, x ′ =
-20.7.

The inverse DCT section 6 performs an inverse DCT on the block (macroblock) in which the watermark data is embedded. The block combining unit 7 combines the block subjected to the inverse DCT with the block not extracted by the block extracting unit 2 to generate image data in which watermark data is embedded. The MPEG encoding unit 8 performs MPEG encoding (quantization and Huffman encoding) of the image data in which the watermark data is embedded in order to create moving image content to be distributed.

In one embodiment of the present invention, the block extracting unit 2 extracts macroblocks at a plurality of positions specified by the template 9 from a frame of a moving image,
The CT coefficient extraction unit 4 extracts DCT coefficients at the same or different positions designated by the template 9 from each of the macroblocks at the plurality of positions, and embeds watermark data in the DCT coefficients at the watermark data embedding unit 5. I will.

That is, the block extracting unit 2
As shown in FIG.
For example, four macro blocks M1 to M4 at a plurality of positions shown in (a) are extracted, and for another frame F, for example, four macro blocks M1 to M4 at a plurality of positions shown in FIG. Blocks M5 to M8 are extracted.

Next, the DCT coefficient extraction unit 4 applies the D macroblocks at the same position designated by the template 9 to the four macroblocks extracted by the block extraction unit 2.
Extract CT coefficients or DCT coefficients at different positions,
The watermark data embedding unit 5 embeds the watermark data.

If the macroblock extraction pattern shown in FIG. 3A is named mark A and the macroblock extraction pattern shown in FIG. 3B is named mark B, an image as shown in FIG. The frame sequence is classified into a frame with a mark A, a frame with a mark B, and a frame without a mark. In the above description, the number of macroblocks extracted by the block extracting unit 2 is four, but the number is not limited to four and may be two or more. In addition, the number of marks is two, A and B, but is not limited thereto.
Or three or more.

Next, a description will be given of a watermark data detecting apparatus for detecting watermark data embedded in a moving image as described above. FIG. 4 is a block diagram showing a schematic configuration of the detection device.

The video content 11 to be inspected is M
In the state of the PEG compressed data, the inspection target macroblock extraction unit 12 extracts the inspection target macroblock. This extraction can be performed according to an instruction from the template 13. The template 13 is preferably the same as the template 9 in FIG.
That is, the template 13 includes the information of the marks A and B and the D at which position in each macroblock of the marks A and B.
Information is output as to whether watermark data x is embedded in the CT coefficient.

According to the information from the template 13, the macroblock extracted by the macroblock extraction unit for inspection 12 is Huffman-decoded by the Huffman decoding unit 14, and then dequantized by the dequantization unit 15. The inversely quantized DCT coefficient is sent to the watermark data detection unit 16 to detect whether the watermark data is embedded.

Specifically, the watermark data detecting unit 16 is instructed from the template 13 by the information of the watermark data embedding position of the DCT coefficient.
6 extracts the DCT coefficient in which the watermark data is embedded from the macroblock at the position indicated by the mark by applying the template 13. For example, as shown in FIG. 5, DCT coefficients x1A, x2A, x
3A and x4A are extracted, and the DCT coefficient x
Extract 1B, x2B, x3B, and x4B. And the D
The sums SA and SB of the absolute values of the CT coefficients are obtained as follows. SA = | x1A | + | x2A | + | x3A | + | x4A | SB = | x1B | + | x2B | + | x3B | + | x4B |

Next, the sums SA and SB are compared, and SA
If> SB, mark A is determined, and if SA <SB, mark B is determined. Also, SA = SB or SA and SB
Are substantially equal, it is determined that there is no mark.

As a result of the watermark data detection, as shown in FIG. 6, when the mark A, the mark B, and the group without the mark are detected, the distributor of the video content of the moving image acquires the copyright. Be able to assert. In this case, even if an obstruction or an attack of thinning out a frame is applied by another person, the mark A, the mark B, and the group without the mark can always be detected, and a great resistance to the obstruction or the attack can be obtained. Will be able to show. Also, by changing the position of the macroblock in which the watermark data is embedded and / or the position of the DCT coefficient for each distribution destination, for example, when a pirated version becomes available, it is possible to assert that the user owns the copyright, and to identify the source of the pirated version. Can be detected.

A method of detecting at a high speed whether or not watermark data is embedded in the above case will be described. The embedding of the watermark data according to the present invention is performed by local D
A method of disturbing the similarity of CT coefficients is employed. For this reason,
For example, in inter-frame prediction at the time of encoding according to MPEG1, the DCT coefficient at the embedding position of watermark data always becomes a non-zero value because the prediction is incorrect. Therefore, it is possible to detect the presence / absence of embedding of the watermark data by using the P1 frame and the B frame of MPEG1 using the inter-frame predictive coding.

Next, another embodiment of embedding watermark data will be described with reference to FIG. In this embodiment, a plurality of patterns of macroblock positions for embedding watermark information in a frame are determined in advance, and these are prepared as templates. For example, two patterns of a group A and a group B as shown in FIG. 7 are prepared. In the illustrated example, the group A is a pattern of “1010” in order from the upper left pixel to the lower right pixel of the frame, and the group B is a pattern of “0010”. The watermark information embedder performs no embedding when the value is 0 and embeds the value when the value is 1 according to the binary expression of the embedding information. In the figure, a triangle indicates a macroblock with embedding, and a square described later indicates a macroblock without embedding.

The simplest method of detecting the presence / absence of embedding of the watermark data is as follows. A non-zero DCT coefficient is added to the embedding position of the watermark data provided from the templates 9 and 13, that is, the position known from the secret key. Is individually examined to determine whether or not the
By making a majority decision in the coefficient set, it is determined whether watermark data has been embedded, that is, DCT.
It can be determined at a high speed whether or not the coefficient has been operated.

As described above, according to the high-speed detection method of the watermark data, the DCT coefficient at the position where the watermark data is embedded has a non-zero value because the prediction is incorrect, and the DCT coefficient at the position where the watermark data is not embedded is The presence / absence of watermark data is determined using the fact that the value becomes 0 because the prediction is not lost. May not be 0. For example, in the case of an image in which an object suddenly jumps out of the shade or an image in which the leaves are swaying by the wind,
For example, since prediction is incorrect in inter-frame prediction of MPEG1, D at a position where watermark data is not embedded is set.
Even if it is a CT coefficient, it does not become 0.

An embodiment of the present invention which addresses this problem is described below.
This will be described below. In this embodiment, as shown in FIG. 8, 1-bit embedding information to be embedded in an image is
It consists of a pair of blocks. 21a, 21b, 2
Reference numerals 2a and 22b denote blocks each having 8 × 8 pixels, and the blocks 21a, 21b, and 22a, 22
b are in a predetermined positional relationship. For example, of the pair, the blocks 21a and 22a located on the left side are:
An embedded DCT coefficient for bit “0” is shown, and blocks 21b and 22b located on the right side show embedded DCT coefficients for bit “1”.

Therefore, a bit “0” is added to a certain frame F.
When the information 21 is embedded, the block 2 for bit 0 is used.
The watermark data is embedded in the DCT coefficient of
The watermark data is not embedded in the embedded DCT coefficient in the block 21b. On the other hand, bit “1”
When embedding information 22 of bit 2, block 2 for bit 0 is used.
The watermark data is not embedded in the DCT coefficient of 2a, but is embedded in the bit 1 block 22b.

Therefore, when the pattern of the group A of FIG. 7 is formed by the embedding method of the present embodiment, the result is as shown in FIG. That is, information A1 and A3 of bit “1” are:
The watermark data is not embedded in the DCT coefficient of the left block of the pair, the watermark data is embedded in the DCT coefficient of the right block, and information A2 and A4 of bit “0” are
In the opposite manner, watermark data is embedded in the DCT coefficients of the block.

Next, when extracting the information embedded in the frame as described above, the information is first decomposed into a group A and a group B shown in FIG. It is determined whether the 1-bit embedded information composed of a pair is bit “0” or “1”.

In the majority decision section 17 (see FIG. 4),
For example, as shown in FIG. 10, (N + 1) frames of the group A are extracted, and a set of embedded DCT coefficients for the bit “0” {x0, x
, XN} and a set of embedded DCT coefficients {y0, y1,..., YN} for bit “1”. Here, xi (i = 0, 1,..., N), yi (i = 0, 1,
.., N) are the bits “0” side and “1” side, respectively.
It is a CT coefficient value. Then, when the following condition (1) is satisfied, it is determined that the bit information A1 is "0". Σ | xi |> Σ | yi | (1) When the condition (1) is reversed, the bit information A1
Is determined to be “1”.

Although the above description has been made with reference to the bit information A1, the bit information A2, A3 and A shown in FIG.
Similarly, determination is made for each pair using a set of sets.

As described above, in the present embodiment, the sum of the set of embedded DCT coefficients for bit “0” and the sum of the set of embedded DCT coefficients for bit “1” are compared, and Since the information A1, A2,... Are bits “0” or “1”, the image may be an image in which an object suddenly jumps out of the shade, or an image of a tree leaf swaying in the wind. Even if the bit information is "0" or "1", it can be accurately determined.

In the above embodiment, when embedding the information of bit “0” or “1”, the pair is replaced by DCT.
Although the block is constituted by the block in which the watermark data is embedded in the coefficient and the block in which the watermark data is not embedded, the pair may be constituted by a block in which the watermark data is embedded in the DCT coefficient.
However, in this case, the watermark data to be embedded in the DCT coefficient of one block is made larger than that of the other block.

[0038]

As described above, according to the present invention, bit information to be embedded in an image frame is constituted by pairs of blocks at two different positions. Even in the case of deviation, the detection accuracy of the bit information embedded in the image frame can be improved. Further, according to the present invention, the bit information embedded in the image frame can be determined by simply comparing the sum of the sets of DCT coefficients of the paired blocks, so that the determination can be performed at high speed.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a schematic configuration of an embodiment of a watermark data embedding apparatus for a moving image according to the present invention.

FIG. 2 is an explanatory diagram of an embedding position of watermark data.

FIG. 3 is an explanatory diagram of a watermark data embedding position according to the present invention.

FIG. 4 is a block diagram illustrating a schematic configuration of an embodiment of a watermark data detection device.

FIG. 5 is an explanatory diagram of a watermark data detection method.

FIG. 6 is a conceptual diagram illustrating an example of a watermark data detection result.

FIG. 7 is an explanatory diagram of another watermark data embedding position of the present invention.

FIG. 8 is an explanatory diagram of a representation format of 1-bit embedded information embedded in an image according to the present invention.

FIG. 9A is a diagram showing FIG. 7A in an expression form according to the present invention.

FIG. 10 is a diagram illustrating a method of extracting information embedded in a frame.

[Explanation of symbols]

1. Block dividing unit 2. Block extracting unit 4. DCT
Coefficient extraction unit, 5: watermark data embedding unit, 7: block synthesis unit, 8: MPEG encoding unit, 9, 13, template, 11: video content, 12: macro block extraction unit to be inspected, 14: Huffman decoding unit , 15: inverse quantizer, 16: watermark data detector, 17: majority decision determiner, 21, 22: information of bits 0 and 1, 21a, 22a
... Embedded DCT coefficients for bit "0", 21b, 22
b ... Embedded DCT coefficient for bit "1".

──────────────────────────────────────────────────の Continued on the front page (51) Int.Cl. ⁷ Identification FI FI Theme Court ゛ (Reference) H04N 7/32 (72) Inventor Yasuhiro Takishima 2-1-15 Ohara, Kamifukuoka-shi, Saitama K-die Research Inc. In-house (72) Inventor Masahiro Wada 2-1-15 Ohara, Kamifukuoka-shi, Saitama F-term in C-D Laboratory Inc. (reference) 5C059 KK43 MA00 MA05 MA23 MC11 MC38 ME02 PP06 PP07 RC35 UA02 UA05 5C063 AB03 AB07 AC01 CA11 CA12 CA23 DA07 DA13 DB09 5J064 AA01 BA09 BA16 BD02 5J104 AA14 NA15

Claims (3)

[Claims] 1. An image frame divided into blocks,
Extract a plurality of blocks at a predetermined position,
Means for performing DCT processing on the plurality of extracted blocks; embedding watermark data in at least one coefficient value of each of the plurality of blocks subjected to the DCT processing;
Watermark data embedding means for embedding 1-bit information by a pair of blocks in which watermark data at different positions are embedded, or a pair of blocks in which the watermark data is embedded and a block in which the watermark data is not embedded. A digital watermarking apparatus for a moving image. 2. The moving image digital watermarking apparatus according to claim 1, wherein one of a pair of the blocks subjected to the DCT processing at the two different positions is a bit “0” and the other block is a bit “0”. Is used for the expression of bit “1”, and the absolute value of the corresponding DCT coefficient value of the block corresponding to the bit to be expressed is defined as the absolute value of the corresponding DCT coefficient value of the other block according to the bit to be expressed. A moving picture digital watermarking apparatus characterized in that the size is made larger. 3. The moving picture digital watermarking device according to claim 1, wherein DCT of one of the blocks in the pair is used.
Comparing the absolute value sum of the DCT coefficient values in the coefficient set with the absolute value sum of the DCT coefficient values in the DCT coefficient set of the other block, and determining that the larger bit of the absolute value sum is embedded; A digital watermarking apparatus for a moving image.