JP2008259059A - Moving image decoding apparatus, method, and program, and moving image encoding apparatus, method, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve a degree of freedom for maintaining image quality when embedding information, and to extract information even if a method of embedding information is changed. <P>SOLUTION: A moving image decoding apparatus includes: a means for dividing a moving image into a plurality of target regions, determining motion vector information indicating a predictive region and a position of the predictive region from a decoded reference image for each target region, and inputting a bit stream containing a difference signal between the predictive region and the target region and data resulting from encoding the motion vector information; a means for decoding the difference signal and the motion vector information from the bit stream and generating a regenerative moving image from a signal of the predictive region specified by the motion vector information and the difference signal from the reference image; and a means for detecting a reference motion vector for a target region within the regenerative moving image to which the motion vector information belongs, comparing the motion vector information with the reference motion vector, and extracting the embedded information on the basis of a comparison result. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a moving picture encoding apparatus, method and program for embedding information in a moving picture, and a moving picture decoding apparatus, method and program, and in particular, moving picture encoding accompanying information embedding using a motion vector. -It relates to decoding technology.

  In order to efficiently transmit and store moving image data, a compression encoding technique is used. In the case of moving images, MPEG1, 2, 4 and H.264 are used. 261-H. H.264 is widely used. In the encoding of a moving image, a prediction signal of a target image to be encoded is generated using another image that is adjacent in the time direction, and the difference between the target image and the prediction signal is encoded. Realize volume reduction.

  For example, H.C. In H.264, an image of one frame is divided into blocks of 16 × 16 pixels, and the image is encoded in units of blocks. In interframe coding, motion detection is performed on a target block of an image to be encoded using another frame that has been encoded and restored as a reference image, and a prediction signal with the least error is determined. Next, a difference value between the target block and the prediction signal is obtained, and discrete cosine transform and quantization processing are performed. A quantized discrete cosine transform coefficient and a motion vector for specifying a prediction signal are entropy-coded to generate encoded data.

  By the way, a technique for embedding watermark information and information related to the moving image (for example, telop information) in the moving image data compressed and encoded in this way is known. For example, information embedding using motion vectors is disclosed in Non-Patent Documents 1 and 2 below.

  According to Non-Patent Document 1, an encoder that embeds digital watermark information first obtains an optimal motion vector in units of one pixel. One bit is extracted from the digital watermark information, and the search range of the motion vector in units of half pixels is limited by the value (0 or 1) of the bit, and the 8 pixels around the reference pixel pointed to by the optimum motion vector obtained in units of one pixel. A pixel closest to the original pixel is obtained from two half pixels, and a vector indicating the obtained pixel is set as a new motion vector. The decoder that extracts the digital watermark information decodes the motion vector, checks whether or not the x-coordinate and y-coordinate are half pixels, and finally calculates the watermark information based on the combination of the x-coordinate and y-coordinate.

According to Non-Patent Document 2, a function F (x, y) = x + y (mod 2) is defined from x and y direction components of a motion vector. Information is embedded by moving the motion vector so that this function value is equal to 0 and 1 of the embedded information bits. This movement vector can be moved by +1 or -1 in the x and y directions, respectively. Therefore, although there are a total of four choices in the combination of the x and y directions, in order to suppress image degradation, Of the above four methods, a method of moving to a position where the error from the optimum image is minimized is used.
"A Study on Copyright Protection by Using" Digital Watermark "in MPEG2", "Symposium on Cryptography and Information Security" (January 29, 1997), Hidenori Nakazawa, Ryoyuki Komine, Jeff Morrison, Hideyoshi Tominaga Written by the Institute of Electronics, Information and Communication Engineers "A few considerations regarding digital watermarking of moving images considering tampering" (April 20, 2000), Hideyuki Tsuno, Hiroyuki Inaba, Masao Kasahara, The Institute of Image Information and Television Engineers, 593-600 page

  In the above two conventional techniques, there is a problem that the image quality that can be maintained is limited because the motion vector is limited and information is embedded according to a predetermined rule. Consider Non-Patent Document 1. For example, when the information to be embedded is “0”, if the motion vector in units of half pixels is limited to the first search range, the second search range different from the first search range has an error larger than that of the first search range. Even if there are few prediction signals, the motion vector that gives the prediction signal cannot be used. That is, there is a problem that the degree of freedom for maintaining the image quality is limited due to the rules for embedding information. Similarly, Non-Patent Document 2 has a problem that the degree of freedom for maintaining the image quality is limited because the motion vector is moved so that the function value is equal to the embedded information.

  Further, in the above two conventional techniques, the decoding side needs to extract information according to the same rules as the encoding side. In other words, every time the information embedding method is changed, the algorithm for extracting information must be changed. Such a method is effective when it is desired to firmly protect the embedded information as in a digital watermark, but the existing method can be changed even if the embedding method is changed as in the case of embedding information such as a teletext of teletext or an Internet URL. It is not suitable when it is desired that the embedded information can be extracted by the decoding device.

  The present invention has been made to solve the above-described problems, and can improve the degree of freedom for maintaining image quality when embedding information, and at the same time can extract information even if the method of embedding information is changed. It aims to be.

  In order to solve the above-described problem, the video decoding device according to the present invention divides a video into a plurality of target areas, and shows the prediction area and the position of the prediction area from the restored reference image for the target area. Input means for determining a vector information and inputting a bitstream including a difference signal between the prediction region and the target region and data obtained by encoding the motion vector information; and the difference from the input bitstream. Image decoding means for decoding a signal and the motion vector information, and generating a reproduced moving image by restoring the moving image from the prediction region signal specified by the motion vector information and the difference signal from the reference image And the reproduced moving image to which the decoded motion vector information belongs in a predetermined method from the generated reproduced moving image and the reference image An information extraction means for detecting a reference motion vector for the target area of the image, comparing the motion vector information with the reference motion vector, and extracting information embedded in the moving image based on the comparison result. It is characterized by doing.

  The moving image decoding apparatus detects a reference motion vector for a target region in a reproduced moving image to which the decoded motion vector information belongs, from a generated reproduced moving image and a reference image, by a predetermined method, The vector information is compared with the reference motion vector, and the embedded information is extracted based on the comparison result. Therefore, it is not necessary to know the rules of the information embedding method used on the encoding side, and the embedded information can be extracted without depending on the information embedding method. That is, even if the information embedding method is slightly changed, there is an advantage that information can be extracted by the decoding device.

  In the moving picture decoding apparatus according to the present invention, the information extracting means includes an adjacent area composed of reproduced pixels adjacent to a target area in the reproduced moving picture to which the motion vector information belongs, and a signal of the reference image. Preferably, the reference motion vector is detected by comparing the motion vector information, the motion vector information is compared with the reference motion vector, and the embedded information is extracted based on the comparison result.

  The moving image encoding apparatus according to the present invention divides the target image into a plurality of target regions, an input unit that inputs a target image to be encoded among a plurality of images constituting the moving image, Based on motion vector information for detecting a plurality of motion vector information for specifying a prediction signal from a reproduced reference image for the target region, and the target region based on one motion vector information among the plurality of motion vector information A prediction signal generation unit that generates a prediction signal for the signal, an encoding unit that encodes a difference between the target region and the prediction signal, and generates an encoded differential signal, and restores and restores the encoded differential signal A reproduction unit that generates a reproduction moving image by adding the encoded difference signal and the prediction signal, and a storage unit that stores the reproduction moving image used as a reference image when generating the prediction signal And an information embedding unit that executes processing for embedding information to be embedded in the moving image in the moving image, wherein the motion detecting unit uses a first motion detection method based on a reference image reproduced for the target area. The second motion detection method detects a first motion vector and detects a motion based on the reproduced moving image and the reference image, detects a second motion vector different from the first motion vector, and embeds the information The means selects either the first motion vector or the second motion vector based on the information to be embedded, and the prediction signal generating means selects the selected first motion vector or the second motion vector. The one motion vector information is used, and a prediction signal for the target region is generated based on the one motion vector information.

  As described above, since the rules for embedding information are not determined from the viewpoint of the moving picture decoding apparatus, the moving picture encoding apparatus has a high degree of freedom for maintaining the image quality. In other words, as long as the motion vector can be discriminated in the video decoding device, the video encoding device can embed information with a motion vector with fewer errors and embed information with higher image quality than the prior art. It becomes possible.

  In the moving image encoding device according to the present invention, the motion detection means, as the second motion detection method, includes an adjacent region composed of reproduced pixels adjacent to a target region in the reproduced moving image, and the It is desirable that the second motion vector is detected based on the comparison result with a reference image signal.

  In the moving image encoding apparatus according to the present invention, the motion detection means may use an error value separated from an error value by a prediction signal specified by the first motion vector by a predetermined width or more as the second motion detection method. It is desirable that the second motion vector that gives the prediction signal is detected.

  By the way, the invention relating to the moving picture decoding apparatus can be regarded as the invention relating to the moving picture decoding method and the invention relating to the moving picture decoding program as described below, and can achieve the same effects.

  A moving image decoding method according to the present invention is a moving image decoding method executed by a moving image decoding apparatus, wherein a moving image is divided into a plurality of target regions, and the target region is predicted from a restored reference image and An input step for determining a motion vector information indicating a position of the prediction region, and inputting a bit stream including a difference signal between the prediction region and the target region and data obtained by encoding the motion vector information; and In addition, the difference signal and the motion vector information are decoded from the bitstream, and the moving image is restored from the prediction region signal specified by the motion vector information and the difference signal from the reference image. An image decoding step for generating a moving image, and the decoded moving image is generated by a predetermined method from the generated reproduction moving image and the reference image. A reference motion vector for a target region in the playback moving image to which vector information belongs is detected, the motion vector information is compared with the reference motion vector, and information embedded in the moving image is determined based on the comparison result. And an information extracting step for extracting.

  At this time, in the information extraction step, the reference motion vector is compared by comparing an adjacent region composed of reproduced pixels adjacent to the target region in the reproduced moving image to which the motion vector information belongs and a signal of the reference image. It is preferable that the motion vector information is compared with the reference motion vector, and the embedded information is extracted based on the comparison result.

  A moving picture decoding program according to the present invention is a moving picture decoding program for causing a computer to function as the moving picture decoding apparatus according to claim 1 or 2.

  In addition, the invention relating to the moving image encoding apparatus can be regarded as the invention relating to the moving image encoding method and the invention relating to the moving image encoding program as described below, and can achieve the same effects.

  A moving image encoding method according to the present invention is a moving image encoding method executed by a moving image encoding device, and inputs a target image to be encoded among a plurality of images constituting the moving image. An input step that divides the target image into a plurality of target regions, and detects a plurality of motion vector information for specifying a prediction signal from a reference image reproduced for the target region, A prediction signal generation step for generating a prediction signal for the target region based on one motion vector information of the motion vector information, a difference between the target region and the prediction signal is encoded, and an encoded difference signal is An encoding step to generate, a reproduction step of restoring the encoded differential signal, and adding the restored encoded differential signal and the prediction signal to generate a reproduced moving image; A storage step for storing the reproduced moving image used as a reference image when generating a measurement signal, and an information embedding step for executing processing for embedding information to be embedded in the moving image in the moving image. In the motion detection step, a first motion vector is detected by a first motion detection method from a reference image that has been reproduced for the target area, and a motion is detected based on the reproduced moving image and the reference image. A detection method detects a second motion vector different from the first motion vector, and the information embedding step selects either the first motion vector or the second motion vector based on the information to be embedded, In the prediction signal generating step, the selected first motion vector or the second motion vector is used as the one motion vector information. Based on the single motion vector information, and generates a prediction signal for the target region.

  At this time, in the motion detection step, as the second motion detection method, an adjacent region composed of reproduced pixels adjacent to the target region in the reproduced moving image is compared with the signal of the reference image, and the comparison is performed. It is desirable to detect the second motion vector based on the result.

  Further, in the motion detection step, as the second motion detection method, the second motion vector that gives a prediction signal having an error value separated from the error value by the prediction signal specified by the first motion vector by a predetermined width or more is detected. It is desirable to do.

  The moving image encoding program according to the present invention is a moving image encoding program for causing a computer to function as the moving image encoding apparatus according to any one of claims 3 to 5.

  The video decoding device according to the present invention detects a reference motion vector for a target region in a playback video to which the decoded motion vector information belongs from a generated playback video and a reference image by a predetermined method. Then, the motion vector information is compared with the reference motion vector, and the embedded information is extracted based on the comparison result. Therefore, it is not necessary to know the rules of the information embedding method used on the encoding side, and the embedded information can be extracted without depending on the information embedding method. That is, even if the information embedding method is slightly changed, there is an advantage that information can be extracted by the decoding device.

  Further, as described above, since the rules for embedding information are not determined from the viewpoint of the moving picture decoding apparatus, the moving picture encoding apparatus has a high degree of freedom for maintaining the image quality. In other words, as long as the motion vector can be discriminated in the video decoding device, the video encoding device can embed information with a motion vector with fewer errors and embed information with higher image quality than the prior art. It becomes possible.

  Hereinafter, embodiments according to the present invention will be described with reference to FIGS.

(About the encoding device)
FIG. 1 shows a block diagram of a video encoding apparatus 100 that processes embedded information. As shown in FIG. 1, the moving image encoding apparatus 100 includes an input terminal 101, a block divider 102, a prediction signal generator 103, a frame memory 104, a subtractor 105, a converter 106, a quantizer 107, and an inverse quantization. , An adder 110, an entropy encoder 111, an output terminal 112, a motion detector 113, and a motion vector selector 114.

(Outline of operation in moving picture coding apparatus)
Hereinafter, the operation of each component of the moving picture coding apparatus 100 will be described. A plurality of images constituting the moving image are input to the input terminal 101. With respect to the input image, the block divider 102 divides the image to be encoded into a plurality of small regions (here, for example, blocks made up of 16 × 16 pixels). The following compression / encoding process is executed for each of the divided blocks.

  A prediction signal generation process is executed by the motion detector 113, the motion vector selector 114, and the prediction signal generator 103 for a block to be encoded (hereinafter referred to as “target block”). This prediction signal generation processing will be described in detail later.

  The prediction signal generated by the prediction signal generation process is sent to the subtractor 105 via the line L103, and the subtractor 105 generates a difference signal by subtracting the prediction signal from the target block. The generated difference signal is sent to the converter 106, which converts the difference signal into a frequency domain signal and outputs it to the quantizer 107. The quantizer 107 performs quantization processing, and outputs the quantized signal to the entropy encoder 111 and the inverse quantizer 108. The entropy encoder 111 converts the quantized signal into a variable length code and outputs it from the output terminal 112. Note that arithmetic coding may be applied instead of the variable-length code.

  Further, the inverse quantizer 108 inversely quantizes the quantized signal and outputs it to the inverse transformer 109, and the inverse transformer 109 performs inverse transformation. Thereby, the quantized signal is converted into a restored differential signal in the spatial domain. Next, the adder 110 adds the restored differential signal and the prediction signal sent via the line L103. Thereby, a reproduction moving image is generated. The reproduced moving image restored in this way is stored in the frame memory 104 because it is used as a reference image for encoding the next image.

  Next, prediction signal generation processing will be described. The motion detector 113 generates a prediction signal (also a 16 × 16 pixel block) for the target block. Specifically, the target block is input to the motion detector 113 via the line L102, and the reference image stored in the frame memory is input to the motion detector 113 via the line L104. A displacement of the target block in the image is detected, and a motion vector that gives a prediction signal with the least error to the target block is detected. For example, the motion detector 113 determines a predetermined search range in the reference image, and acquires a candidate block having the same number of pixels as the target block from the search range. Next, the motion detector 113 obtains a difference value between corresponding pixels in the target block and the candidate block, adds the absolute value of the difference value, and obtains a prediction error regarding the candidate block. In this way, the motion detector 113 calculates a prediction error for each of all candidate blocks within the search range, identifies a candidate block with the smallest prediction error, and determines the displacement of the target block that gives the candidate block, It detects as an optimal motion vector. By using the prediction signal with the smallest prediction error in this way, it is possible to efficiently encode a moving image. The optimal motion vector obtained by the block matching method as described above is referred to as a first motion vector, and in some cases, abbreviated as “MV1”.

  In addition to the first motion vector, the moving image coding apparatus 100 according to the present embodiment also obtains a second motion vector (sometimes abbreviated as “MV2”). There are roughly two methods for obtaining the second motion vector. Here, processing based on the first method will be described first.

In the first method, a motion vector that gives a prediction error value that is more than a predetermined width away from the prediction error value corresponding to the first motion vector is obtained as the second motion vector. That is, assuming that the prediction error value given by the second motion vector is SAD (MV2), the prediction error value given by the first motion vector is SAD (MV1), and the predetermined width is M, SAD ( The motion vector corresponding to the smallest prediction error value is obtained as the second motion vector from MV2).
| SAD (MV2) -SAD (MV1) |> M (1)
Here, the predetermined width M is a value that can distinguish SAD (MV2) and SAD (MV1) in the decoding apparatus.

  In the present embodiment, for example, the predetermined width M is determined by a function of the quantization width used in the encoding device. Although details will be described later, the decoding apparatus obtains MV2, compares the obtained MV2 with a motion vector included in the encoded data, and extracts embedded information based on the comparison result. Therefore, MV1 may be anything as long as it is different from MV2. For example, with respect to MV2 determined in the decoding device, MV1 may be a negative value of one or both of the horizontal component and the vertical component of MV2, or may be increased or decreased by the minimum unit of the motion vector.

  The first motion vector and the second motion vector determined in this way are sent to the motion vector selector 114 via the line L113. The motion vector selector 114 is also input with binarized data of information to be embedded in the image, and the motion vector selector 114 sets MV1 or MV2 in accordance with “0” or “1” of the information to be embedded. select. The selected motion vector is sent to the prediction signal generator 103 via the line L114. The prediction signal generator 103 acquires a prediction signal from the frame memory 104 based on the selected motion vector, and outputs the acquired prediction signal to the subtractor 105 via the line L103 as a prediction signal of the target block.

(Information embedding process in moving picture encoding apparatus)
Next, an information embedding process executed by the moving picture encoding apparatus 100 will be described with reference to FIGS. In FIG. 2, the first motion vector MV1 is obtained for all the blocks of the target image by the block matching method described above (step 211). Next, in step 212, the second motion vector MV2 is obtained for all the blocks of the target image by the first method described above. MV2 is required to have at least one horizontal / vertical component different from that of MV1.

  From step 213, information is embedded in each block. First, in step 213, the counter bn is set to 0, and in the next step 214, 1 bit (0 or 1) of the binarized data of the information to be embedded is input. In step 215, a motion vector MV for generating a prediction signal of the block bn is determined based on the input 1 bit.

  One embodiment of the determination method at step 215 is shown in FIG. That is, it is determined in step 301 of FIG. If the embedded information is 1, MV1 is selected at step 302 and MV = MV1. If the embedded information is not 1, MV = MV2 is set (step 303). In step 304, the selected motion vector MV is output.

  Returning to FIG. 2, in the next step 216, the motion vector MV is used to acquire the prediction signal of the block bn, and then differential encoding between the acquired prediction signal and the target block is performed. Information is embedded in the image by executing the processing in steps 215 and 216 for all blocks.

  If it is not necessary to perform the above processing on all blocks depending on the amount of information to be embedded, information is embedded only in some of the blocks after identifying some of the blocks. Other blocks may be encoded using the first motion vector MV1. 2 may be placed after step 213, and processing may be performed so as to embed information in the image while obtaining the first motion vector and the second motion vector in units of blocks.

(Second method for obtaining the second motion vector MV2)
FIG. 4 is a schematic diagram for explaining a second method for obtaining the second motion vector. In FIG. 4, the first motion vector of the target block 402 is obtained by the conventional method described above. The second motion vector is obtained using the pixel group 403 adjacent to the target block 402. That is, a predicted pixel group that is most similar to the pixel group 403 is obtained from the search area 409 in the reference image 406.

  Specifically, an error between each of the candidate pixel groups 408 in the search area 409 and the pixel group 403 is obtained, and among all candidate pixel groups, the candidate pixel group with the smallest error from the pixel group 403 is predicted pixel. A group. Then, the displacement between the pixel group 403 and the predicted pixel group in the search area 409 is set as the second motion vector of the target block. In this embodiment, the sum of absolute values of differences between pixels is used as the error, but other methods may be used. Since the pixel group 403 has already been restored before the target block 402 is processed, the decoding device can also obtain the same second motion vector using the same pixel group 403.

  FIG. 5 shows a method for determining the motion vector MV corresponding to the method for obtaining the second motion vector in FIG. 4, and is executed in step 215 in FIG. That is, it is determined in step 501 of FIG. If the embedded information is 1, the process proceeds to step 502, and it is determined whether MV1 and MV2 are the same. Here, when the embedding information is 1, encoding is performed with a motion vector different from the second motion vector MV2. Therefore, if it is determined in step 502 that MV1 and MV2 are different, in step 503, the motion vector MV used for encoding is set to MV1. On the other hand, when MV1 and MV2 are the same, it is necessary to separately determine a motion vector different from MV1 and use the determined motion vector. Therefore, in step 504, a motion vector MV1 'different from MV1 is determined by the method shown in Equation 1 above, and the motion vector MV used for encoding is set to MV1'.

  On the other hand, if the embedded information is 0 in step 501, the process proceeds to step 505, where it is determined whether MV1 and MV2 are the same. Here, when the embedded information is 0, encoding is performed with the same motion vector as the second motion vector MV2. Therefore, if it is determined in step 505 that MV1 and MV2 are the same, the process proceeds to step 506, and the motion vector used for encoding is set to MV1. On the other hand, if MV1 and MV2 are different, the process proceeds to step 507, and MV2 is used as the motion vector used for encoding instead of MV1. The motion vector MV set in the above steps 504, 503, 506, and 507 is output in step 508.

  According to the processing in FIG. 5 described above, when the decoding side determines that the motion vector MV included in the bitstream is the same as MV2 obtained by the decoding device, the embedded information is 0, and the motion vector MV Is determined to be different from MV2 obtained by the decoding device, it can be uniquely determined that the embedded information is 1.

  As described above, rather than changing the motion vector of data in which information is embedded according to a predetermined rule, it is possible to generate encoded data with high image quality on a global basis in order to embed information in the order of small prediction error. it can.

  In the above-described method, the method of embedding 1 bit in one target block has been described. However, 2-bit data may be embedded in one block. Since the second motion vector MV2 can be obtained by the decoding apparatus, the motion vector corresponding to the smallest prediction error, the second smallest, or the third smallest motion vector is defined as MV2. Can be embedded with 2-bit data. For example, when the motion vector used for encoding is different from MV2, “00” is assigned. When the motion vector used for encoding is the same as MV2, “01” is set as the smallest prediction error due to MV2. It is only necessary to assign “10” to the second smallest and “11” to the third smallest. In the case of 3 bits or more, it may be expanded in the same way.

(Decoding device)
FIG. 6 is a block diagram of a moving picture decoding apparatus 600 for processing embedded information according to the present invention. As shown in FIG. 6, the moving picture decoding apparatus 600 includes an input terminal 601, a data analyzer 602, an inverse quantizer 603, an inverse transformer 604, an adder 605, an image output terminal 606, a frame memory 607, and a prediction signal generation. 608, a motion detector 609, an information extractor 610, and an information output terminal 611.

  In the following description of the video decoding device, the “first motion vector” corresponds to the “motion vector information” described in claim 1, and the “second motion vector” is described in claim 1. It corresponds to “reference motion vector”.

(Outline of operation in video decoding device)
Hereinafter, the operation of each component of the moving picture decoding apparatus 600 will be described. The input terminal 601 divides the moving image into a plurality of target regions, determines a prediction block and a first motion vector indicating the position of the prediction block from the restored reference image for the target block, A bit stream including data obtained by encoding the difference signal and the first motion vector is input. In the present embodiment, a bit stream obtained by processing by the video encoding device 100 in FIG. 1 is input to the video decoding device 600.

  The data analyzer 602 analyzes input data (input bit stream) and decodes the differential signal and the first motion vector. Although variable length decoding is used here, arithmetic decoding may be used. The decoded difference signal is inversely quantized by the inverse quantizer 603, and the difference signal after inverse quantization is inversely transformed by the inverse transformer 604 to be a restored differential signal in the spatial domain. On the other hand, the first motion vector decoded by the data analyzer 602 is sent to the prediction signal generator 608 via the line L602b.

  The prediction signal generator 608 obtains a prediction signal specified by the first motion vector from the reference image stored in the frame memory 607, and the prediction signal is sent to the adder 605 via the line L608. The adder 605 generates a playback moving image by adding the prediction signal and the restored difference signal. The generated playback moving image is transmitted to an external display device via the image output terminal 606 and is also sent to the frame memory 607 for storage.

  On the other hand, the motion detector 609 receives the signal of the target block and the signal of the reference image in the generated playback moving image from the frame memory 607, and the decoded first motion vector is obtained by a block matching method. A second motion vector for the target block in the reproduced moving image to which it belongs is detected. The first motion vector is sent to the information extractor 610 via the line L602b, and the second motion vector is sent to the information extractor 610 via the line L610. The information extractor 610 compares the first motion vector and the second motion vector, and extracts (determines) information embedded in the moving image (embedded information) based on the comparison result. Specifically, if the first motion vector and the second motion vector are the same, the information extractor 610 determines that the embedding information is “0”, and if the first motion vector and the second motion vector are different, It is determined that the information is “1”. This results in a result corresponding to FIG. 3, but “0” and “1” may be assigned in reverse.

  In addition to the first method described above, the motion detector 609 receives, from the frame memory 607, an adjacent block signal composed of reconstructed pixels adjacent to the target block and a reference image signal. The second motion vector may be detected by the method described in. That is, the motion detector 609 obtains an error between the pixel group 403 and the candidate pixel group 408 in the search area 409 in the reference image 406 for the pixel group 403 adjacent to the target block 402 in FIG. Among all candidate pixel groups 408 in the region 409, a candidate pixel group having the smallest error from the pixel group 403 is set as a predicted pixel group. Then, the motion detector 609 sets the displacement between the pixel group 403 and the predicted pixel group as the second motion vector of the target block. The information extractor 610 compares the second motion vector thus obtained with the first motion vector, and extracts (determines) information (embedded information) embedded in the moving image based on the comparison result. . For example, when the first motion vector and the second motion vector are the same, the information extractor 610 determines that the embedded information is “0”. When the first motion vector and the second motion vector are different, the embedded information is “ 1 ”.

  FIG. 7 shows a flowchart of a moving picture decoding process for processing embedded information, which is executed by the moving picture decoding apparatus 600. In step 701 in FIG. 7, compression-encoded data is input. Next, the data analyzer 602 performs entropy decoding on the input data, and extracts information including the quantized transform coefficient, the quantized information, and the first motion vector (step 702). The prediction signal generator 608 generates a prediction signal based on the extracted first motion vector (step 703). The inverse quantizer 603 inversely quantizes the extracted quantized transform coefficient using the quantization information (step 704). The inverse transformer 604 generates a reproduction difference signal by inversely transforming the result (step 705). Then, the adder 605 generates a reproduction signal by adding the prediction signal and the reproduction difference signal (step 706). The generated reproduction moving image is output to the frame memory 607 and temporarily stored (step 707). The processes in steps 701 to 707 are executed for all the data, thereby restoring the image.

  FIG. 8 shows a flowchart of information extraction processing executed by the motion detector 609 and the information extractor 610 of the video decoding device 600. In step 801 of FIG. 8, the first motion vector included in the bit stream and decoded is input to the information extractor 610, and the block signal and reference image signal of the playback moving image corresponding to the first motion vector are stored in the frame memory. 607 to the motion detector 609. Next, in step 802, the motion detector 609 detects a second motion vector from the block signal of the reproduced moving image and the reference image signal. As described above, the second motion vector detection process is performed using the reproduced target block or a pixel group adjacent to the target block.

  In step 803, the information extractor 610 compares the input first motion vector with the second motion vector detected by the motion detector 609. Here, if the first motion vector and the second motion vector are the same, the information extractor 610 determines that the embedded information is “0” (step 804), and the first motion vector and the second motion vector are different. The embedded information is determined to be “1” (step 805). Thereafter, the processing in steps 801 to 805 is executed for all target blocks, and when the execution is completed for all target blocks, the information extractor 610 outputs the extracted embedded information to the outside from the information output terminal 611 (step 807). .

  When information is embedded only in a part of the blocks, the embedded information may be extracted only for the part of the corresponding block.

  In the above-described embodiment, the case where 1 bit is embedded in one target block has been described, but the same processing is performed even when 2 bits of data are embedded per block. Depending on whether the second motion vector detected by the motion detector 609 corresponds to the smallest prediction error, the second smallest, or the third smallest, Distinguish embedded information. For example, when the first motion vector is different from the second motion vector, “00”, and when the first motion vector is the same as the second motion vector, “01” is set to the smallest prediction error due to the second motion vector. The information is extracted as “10” for the second smallest and “11” for the third smallest. In the case of 3 bits or more, it may be expanded in the same way. Thus, in order to embed and extract information in ascending order of prediction error, instead of changing the motion vector according to a predetermined rule, the data to embed information is generated with high quality encoded data on a global basis. Can do. Further, since the second motion vector is obtained from the reproduced moving image and the embedded information is extracted based on the obtained second motion vector, it is not necessary to know the rules of the information embedding method used on the encoding side, and the information embedding is performed. Information can be extracted without depending on the method. That is, there is an advantage that information can be extracted by the same decoding device even if the information embedding method is slightly changed.

(Video coding program / video decoding program and operating environment)
By the way, the invention relating to the moving picture coding apparatus can be regarded as an invention relating to a moving picture coding program for causing a computer to function as a moving picture coding apparatus. Further, the invention relating to the moving picture decoding apparatus can be regarded as an invention relating to a moving picture decoding program for causing a computer to function as a moving picture decoding apparatus. The moving image encoding program and the moving image decoding program are provided by being stored in a recording medium, for example. Examples of the recording medium include a recording medium such as a flexible disk, a CD-ROM, and a DVD, a semiconductor memory, and the like.

  FIG. 9 is a diagram illustrating a hardware configuration of the computer 30 for executing the program recorded on the recording medium 10. FIG. 10 is a perspective view of the computer 30 for executing the program recorded on the recording medium 10. FIG. Here, the computer 30 includes a DVD player, a set top box, a mobile phone, and the like that have a CPU and perform processing and control by software.

  As shown in FIG. 9, a computer 30 includes a reading device 12 such as a floppy disk drive device, a CD-ROM drive device, a DVD drive device, a working memory (RAM) 14 in which an operating system is resident, and a recording medium 10. A memory 16 for storing the program stored therein, a display 18, a mouse 20 and a keyboard 22 as input devices, a communication device 24 for transmitting and receiving data and the like, and a CPU 26 for controlling execution of the program. ing. When the recording medium 10 is inserted into the reading device 12, the computer 30 can access the moving image encoding program and the moving image decoding program stored in the recording medium 10 from the reading device 12, and the moving image encoding program And the moving picture decoding program enables the computer 30 to operate as a moving picture encoding apparatus and a moving picture decoding apparatus according to the present invention.

  As shown in FIG. 10, the moving picture encoding program and the moving picture decoding program may be provided as a computer data signal 40 superimposed on a carrier wave via a wired network / wireless network. In this case, the computer 30 can store the moving image encoding program and the moving image decoding program received by the communication device 24 in the memory 16 and execute the moving image encoding program and the moving image decoding program.

  According to the present embodiment described above, there is an advantage that information can be extracted with an existing decoding device even if the information embedding method is slightly changed without depending on the rules of the information embedding method used on the encoding side. As long as the motion vector can be discriminated by the decoding device, the information can be embedded with a motion vector with a smaller overall error in the encoding device, and the information can be embedded with higher image quality than the prior art.

It is a block diagram of the moving image encoder which processes embedding information. It is a flowchart which shows the information embedding process in a moving image encoder. 3 is a flowchart showing a first method relating to the processing of step 215; FIG. 10 is a schematic diagram for explaining a second method relating to the processing in step 212; 12 is a flowchart showing a second method relating to the processing in step 215. It is a block diagram which shows the moving image decoding apparatus which processes embedding information. It is a flowchart which shows the moving image decoding process which processes embedding information. It is a flowchart which shows an information extraction process. It is a figure which shows the hardware constitutions of the computer for performing the program recorded on the recording medium. It is a perspective view of a computer for executing a program stored in a recording medium.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 10 ... Recording medium, 12 ... Reading apparatus, 14 ... Working memory, 16 ... Memory, 18 ... Display, 20 ... Mouse, 22 ... Keyboard, 24 ... Communication apparatus, 30 ... Computer, 40 ... Computer data signal, 100 ... Movie Image coding apparatus 101 ... Input terminal 102 ... Block divider 103 ... Predicted signal generator 104 ... Frame memory 105 ... Subtractor 106 ... Transformer 107 ... Quantizer 108 ... Inverse quantizer , 109: inverse converter, 110: adder, 111: entropy encoder, 112: output terminal, 113: motion detector, 114: motion vector selector, 402: target block, 403: adjacent pixel group, 406 Reference image, 408 ... Candidate pixel group, 409 ... Search area, 600 ... Moving picture decoding device, 601 ... Input terminal, 602 ... Data solution 603 ... inverse quantizer, 604 ... inverse transformer, 605 ... adder, 606 ... image output terminal, 607 ... frame memory, 608 ... prediction signal generator, 609 ... motion detector, 610 ... information extractor, 611: Information output terminal.

Claims (12)

A moving image is divided into a plurality of target regions, a motion vector information indicating a prediction region and a position of the prediction region is determined from a restored reference image for the target region, a difference signal between the prediction region and the target region, and Input means for inputting a bitstream including data obtained by encoding the motion vector information;
By decoding the difference signal and the motion vector information from the input bitstream, and restoring the moving image from the prediction region signal specified by the motion vector information and the difference signal from the reference image. Image decoding means for generating a playback moving image;
A reference motion vector for a target region in the playback video to which the decoded motion vector information belongs is detected from the generated playback video and the reference image by a predetermined method, and the motion vector information and the reference video An information extracting unit that compares a reference motion vector and extracts information embedded in the moving image based on the comparison result;
A moving picture decoding apparatus comprising:   The information extraction unit detects the reference motion vector by comparing an adjacent area including reproduced pixels adjacent to a target area in the reproduction moving image to which the motion vector information belongs and a signal of the reference image. The moving picture decoding apparatus according to claim 1, wherein the motion vector information is compared with the reference motion vector, and the embedded information is extracted based on the comparison result. Among a plurality of images constituting a moving image, an input means for inputting a target image to be encoded;
A motion detection unit that divides the target image into a plurality of target regions and detects a plurality of pieces of motion vector information for specifying a prediction signal from a reference image reproduced for the target region;
Prediction signal generating means for generating a prediction signal for the target region based on one motion vector information among the plurality of motion vector information;
Encoding means for encoding a difference between the target region and the prediction signal and generating an encoded difference signal;
Reproducing means for reconstructing the encoded differential signal and adding the reconstructed encoded differential signal and the prediction signal to generate a reproduced moving image;
Storage means for storing the reproduced moving image used as a reference image when generating a prediction signal;
Information embedding means for executing processing for embedding information to be embedded in the moving image in the moving image,
The motion detection means detects a first motion vector by a first motion detection method from a reproduced reference image for the target area, and detects a motion based on the reproduced moving image and the reference image. Detecting a second motion vector different from the first motion vector,
The information embedding means selects either the first motion vector or the second motion vector based on the information to be embedded,
The prediction signal generation means uses the selected first motion vector or the second motion vector as the one motion vector information, and generates a prediction signal for the target region based on the one motion vector information. A moving picture coding apparatus characterized by that.   The motion detection means, as the second motion detection method, compares an adjacent region composed of reproduced pixels adjacent to a target region in the reproduced moving image with a signal of the reference image, and determines the comparison result. 4. The moving image encoding apparatus according to claim 3, wherein the second motion vector is detected based on the second motion vector.   As the second motion detection method, the motion detection means detects the second motion vector that gives a prediction signal having an error value that is more than a predetermined width away from an error value of the prediction signal specified by the first motion vector. The moving picture coding apparatus according to claim 3. A moving picture decoding method executed by a moving picture decoding apparatus,
A moving image is divided into a plurality of target regions, a motion vector information indicating a prediction region and a position of the prediction region is determined from a restored reference image for the target region, a difference signal between the prediction region and the target region, and An input step of inputting a bitstream including data obtained by encoding the motion vector information;
By decoding the difference signal and the motion vector information from the input bitstream, and restoring the moving image from the prediction region signal specified by the motion vector information and the difference signal from the reference image. An image decoding step for generating a playback moving image;
A reference motion vector for a target region in the playback moving image to which the decoded motion vector information belongs is detected from the generated playback moving image and the reference image by a predetermined method, and the motion vector information and the reference image are detected. An information extraction step of comparing a reference motion vector and extracting information embedded in the moving image based on the comparison result;
A moving picture decoding method comprising:   In the information extraction step, the reference motion vector is detected by comparing the reference image signal with an adjacent region composed of reproduced pixels adjacent to the target region in the reproduced moving image to which the motion vector information belongs. 7. The moving picture decoding method according to claim 6, wherein the motion vector information is compared with the reference motion vector, and the embedded information is extracted based on the comparison result. A video encoding method executed by a video encoding device,
An input step of inputting a target image to be encoded among a plurality of images constituting the moving image;
A motion detection step of dividing the target image into a plurality of target regions, and detecting a plurality of motion vector information for specifying a prediction signal from a reference image reproduced for the target region;
A prediction signal generating step for generating a prediction signal for the target region based on one motion vector information among the plurality of motion vector information;
An encoding step of encoding a difference between the target region and the prediction signal and generating an encoded difference signal;
Regenerating the encoded differential signal, and adding the reconstructed encoded differential signal and the prediction signal to generate a reproduction moving image; and
A storing step of storing the reproduced moving image used as a reference image when generating a prediction signal;
An information embedding step for executing processing for embedding information to be embedded in the moving image in the moving image,
In the motion detection step, a second motion detection is performed in which a first motion vector is detected from a reference image that has been reproduced for the target region by a first motion detection method, and motion is detected based on the reproduced moving image and the reference image. Detecting a second motion vector different from the first motion vector,
In the information embedding step, either the first motion vector or the second motion vector is selected based on the information to be embedded,
In the prediction signal generation step, the selected first motion vector or the second motion vector is used as the one motion vector information, and a prediction signal for the target region is generated based on the one motion vector information. A moving image encoding method characterized by:   In the motion detection step, as the second motion detection method, an adjacent region composed of reproduced pixels adjacent to a target region in the playback moving image is compared with a signal of the reference image, and the comparison result is obtained. 9. The moving image encoding method according to claim 8, wherein the second motion vector is detected based on the second motion vector.   In the motion detection step, as the second motion detection method, the second motion vector that provides a prediction signal having an error value separated by a predetermined width or more from an error value of the prediction signal specified by the first motion vector is detected. 9. The moving picture encoding method according to claim 8, wherein:   A moving picture decoding program for causing a computer to function as the moving picture decoding apparatus according to claim 1.   A moving picture coding program for causing a computer to function as the moving picture coding apparatus according to any one of claims 3 to 5.

Images