JP2004364263A - Information embedding apparatus and method, tampering detecting apparatus and method, and recording medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a tampering detecting technology by which tampering on a time base can be detected in moving image data. <P>SOLUTION: An inherent information calculating part 102 calculates inherent information which is inherent for each of frames configuring moving image data. An information embedding part 104 embeds the inherent information of a specific frame image present in a position different from a frame image of an embedding target on the time base into the frame image of the embedding target as tempering detection information. Tampering on the time base can be detected in the digital moving image data by order information. Further, tampering in a predictive encoding image itself can be detected. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to an information embedding device that embeds information for detecting tampering in moving image data and a related technique.

  Conventionally, recording and reproduction have been performed by analog image processing technology. Examples of the medium include a film and a VHS video tape. In such analog image processing, if tampering is added, an unnatural trace is left at a joint or the like, and it is said that a skilled person can see whether tampering has been added to a moving image. Thus, if a skilled person determines that a VHS videotape has not been tampered with, it will be able to prove that the VHS videotape is authentic.

  In recent years, digital image input devices such as digital cameras and digital recording devices such as hard disk recorders have rapidly spread in place of analog image processing technology.

  Digital moving image data can be easily and easily falsified (edited, processed, etc.) using commercially available image processing software. Moreover, it is also possible to perform falsification on the digital moving image data without leaving any trace. Therefore, it is almost impossible for even a skilled person to determine whether or not certain digital moving image data has been tampered with.

  For example, digital moving image data has been used in the field of, for example, monitoring and recording devices, and there is a need to establish a detection technique that can determine whether digital moving image data has been tampered with.

  The digital moving image data includes a data unit configured to sequentially arrange a plurality of frame images on a time axis, and a header unit that holds information about the data unit. This time axis may be a time axis for reproducing a plurality of frame images or a time axis for storing a plurality of frame images on a recording medium (for example, a disk, a tape, a memory, or the like). The “time axis” in this specification may be any of these time axes.

  In the present specification, the “frame image” may be an image of one frame in moving image data employing a frame structure, or a frame image created based on a field image in moving image data employing a field structure. Images are fine.

  Now, falsification of moving image data, which is the subject of this specification, can be roughly classified into the following two types.

  (First falsification) One of them is to process a frame image itself. For example, there is a case where a criminal erases his / her own image (a partial image of a frame image) from a frame image or replaces it with another image in order to conceal his / her crime.

  (Second Modification) Another modification is to modify a mode in which a plurality of frame images are sequentially arranged on a time axis by a data portion of moving image data. For example, the criminal extracts a frame image (inconvenient frame image) of a scene in which the criminal is captured in frame image units, or extracts the extracted portion in frame image units in a frame image of a scene in which the criminal is not captured. This is the case when replacing.

  Of course, both the first falsification and the second falsification may be performed. In order to prove that the moving image data is authentic and has not been tampered with, it must be possible to cope with not only the first tampering but also the second tampering.

  Conventionally, for the first falsification, a falsification detection method using a digital watermark technique is known. Digital watermarking is a technique for embedding digital information in digital moving image data in a form that cannot be perceived by humans.

  For example, Japanese Patent Application Laid-Open No. 2002-271609 (Patent Document 1) discloses a method for verifying digital image data for tampering.

  Hereinafter, this falsification verification method will be described with reference to FIG. FIG. 15 is a block diagram of a conventional falsification verification device.

  In FIG. 15, a control unit 804 controls each of the elements 801 to 803 and 805 to 813. The key input unit 801 inputs key information, and the pseudo random number generation unit 810 generates a pseudo random number. The image data is input from the image data input unit 809 and stored in the memory 812.

  The block division unit 807 divides the image data stored in the memory 812 into blocks of a specific size, and the frequency conversion unit 802 converts the divided image data into a frequency space. The falsification verification data embedding unit 803 generates falsification verification data from the generated frequency components, and the falsification verification data embedding unit 803 embeds the generated falsification verification data in a part of the generated frequency components. This makes it possible to determine whether or not tampering has occurred based on the tampering verification data. Note that the Huffman coding unit 808 performs Huffman coding on the frequency component in which the data for falsification verification is embedded.

  Japanese Patent Application Laid-Open No. H11-341268 (Patent Document 2) discloses the following falsification detection method. That is, the digital watermark is embedded in the digital image by embedding the bits of the digital signature of the hash function of the digital image in the frequency function of the digital image. On the other hand, a vulnerable watermark embedded at the time of watermark insertion is extracted from the digital image, a hash function of the digital image is calculated in the same way as at the time of inserting the digital watermark, and the extracted watermark is converted into a valid signature of the hash value using the public key. By verifying whether falsification is performed, falsification is detected. If the signature is valid, it is determined that the digital image has not been tampered with; otherwise, it is determined that the digital image has been tampered with.

  However, in the related art, when moving image data conforms to a format using inter-frame prediction (for example, MPEG), a digital watermark used for falsification detection cannot be embedded in a predicted encoded image.

  For example, when the moving image data records a scene with little motion and almost a still image, the difference between the frame images becomes almost zero, and the prediction coded image (P-picture or B-picture in MPEG) is used. ) Will have substantially no blocks being coded. In this case, both Patent Documents 1 and 2 assume that there is a coded block in which a digital watermark is to be embedded. Therefore, it is impossible to embed information itself. Therefore, in the related art, tampering detection cannot be performed on a prediction coded image.

  Further, the conventional technology cannot cope with the first falsification but can cope with the second falsification, which is insufficient.

  For the second falsification, it is conceivable to add a time display (visible) to the frame image and use a technology for recording. However, in this case, a human checks the time display one by one in order to detect the presence or absence of tampering. In this case, when the moving image data is extended for a long time, a large amount of labor is required for the check, which is not realistic.

More importantly, if the person making the tampering creates a frame image with the time display added to make it consistent, it is not possible to expose the fraud and increase the credibility of the moving image data. Can not. As a result, the conventional technique cannot take sufficient measures against falsification of moving image data.
JP-A-2002-271609 JP-A-11-341268 JP 2000-165248 A U.S. Pat. No. 6,064,764

  A first object of the present invention is to provide a technique capable of detecting tampering on a time axis in digital moving image data.

  A second object of the present invention is to provide a technique capable of detecting tampering of a prediction coded image itself.

  The information embedding device according to claim 1 embeds information in moving image data including a data section configured to sequentially arrange a plurality of frame images on a time axis and a header section holding information on the data section. An apparatus, comprising: an order information input unit that generates order information of a plurality of frame images; and an information embedding unit that embeds the order information generated by the order information input unit in moving image data so that the order information is not visible in the plurality of frame images. And

  The falsification detection device according to claim 5, wherein the presence / absence of falsification of the moving image data including a data portion configured to sequentially arrange a plurality of frame images on a time axis and a header portion holding information on the data portion. An input unit for inputting moving image data in which order information is embedded, a detecting unit for detecting order information from the moving image data input to the input unit, and an order detected by the detecting unit. A falsification determining unit that determines whether or not the moving image data has been falsified based on the information;

  In these configurations, the falsification on the time axis can be detected by the order information, and the automatic detection thereof is also possible. In addition, since the order information is embedded in the moving image data in a manner invisible in the frame image, the order information can be protected from malicious rewriting.

  In the information embedding device, the information embedding unit embeds the order information in a plurality of frame images as a digital watermark.

  With this configuration, the order information can be extracted from the frame image, and the image quality degradation of the frame image can be reduced. Further, an important partial image of the frame image (for example, a partial image in which the criminal's face is captured) is not hidden behind the sequence information and becomes unclear.

  In the information embedding device according to the third aspect, the information embedding section embeds the order information in the header section.

  With this configuration, the sequence information can be embedded in the moving image data without performing any processing on the frame image.

  In the information embedding device, the order information is a number indicating the order of the plurality of frame images.

  In the falsification detection device according to the sixth aspect, the order information is a number indicating the order of a plurality of frame images.

  With these configurations, the order can be easily defined and interpreted by the number.

  In the falsification detection device, the falsification determination unit determines that the moving image data has been falsified when the order of the plurality of frame images is changed, dropped, or added.

  With this configuration, various forms of tampering on the time axis can be detected.

  The information embedding device according to claim 8 embeds information in moving image data including a data portion configured to sequentially arrange a plurality of frame images on a time axis and a header portion holding information on the data portion. The apparatus, which is included in the moving image data, a specific frame image of the plurality of frame images, a single or a plurality of other frame images in a certain relationship on the time axis, a plurality of frame images And an information embedding unit that embeds the calculated unique information in a specific frame image.

  The falsification detection device according to claim 13, wherein the presence or absence of falsification of the moving image data includes a data portion configured to sequentially arrange a plurality of frame images on a time axis, and a header portion holding information on the data portion. An input unit for inputting moving image data in which order information is embedded, and in the moving image data input to the input unit, the input unit is embedded in a specific frame image of a plurality of frame images. In addition to detecting the unique information and obtaining the first unique information, one or more different frame images having a fixed relationship with respect to a specific frame image on the time axis and included in the plurality of frame images A unique information calculation unit that calculates unique information to obtain second unique information, a comparison unit that compares the first unique information with the second unique information, and a comparison result of the comparison unit Based on, and a determining tamper determining unit whether the alteration of the moving image data.

  In these configurations, since the specific information is used, a specific frame image can be associated with another frame image. If the association is broken, it can be determined that there is tampering on the time axis. Further, when the frame image itself is tampered with, the association is broken, so that the first tampering can be dealt with at the same time.

  In the information embedding device according to the ninth aspect, the unique information is a hash value related to another frame image.

  With this configuration, unique information can be accurately represented using a hash value.

  In the information embedding device according to the tenth aspect, the other frame image is a frame image adjacent to a specific frame image on a time axis.

  With this configuration, adjacent frame images can be associated with each other to counter tampering on the time axis.

  In the information embedding device according to claim 11, the moving image data is MPEG-compressed moving image data, the specific frame image is an intra-coded image, and the one or more other frame images are one or more. It is a prediction encoding image.

  With this configuration, it is possible to cope with a case where the data of the prediction coded image is small.

  In the information embedding device according to the twelfth aspect, the intra-coded image and one or more predictive-coded images belong to one GOP (group of pictures).

  With this configuration, processing can be performed in a unit by using the correlation.

  According to the present invention, tampering on the time axis can be detected in digital moving image data based on the order information. In addition, it is possible to detect tampering of the prediction coded image itself.

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

(Embodiment 1)
FIG. 1 is a block diagram of an information embedding device according to Embodiment 1 of the present invention.

  As shown in FIG. 1, the information embedding device according to the present embodiment includes an order information input unit 101, a unique information calculation unit 102, a unique information holding unit 103, and an information embedding unit 104.

  Here, as shown in FIG. 10, the moving image data includes a data section 20 configured to sequentially arrange a plurality of frame images F <b> 1 to F <b> 5 on a time axis, and a header section holding information on the data section 20. 10 is included.

  1, the order information input unit 101 generates order information of a plurality of frame images F1 to F5. The information embedding unit 104 embeds the order information generated by the order information input unit 101 in the moving image data so that the order information is not visible in the plurality of frame images F1 to F5.

  The information embedding unit 104 of the present embodiment can embed the order information in a plurality of frame images as a digital watermark, or embed the order information in the extension area 10a of the header unit 10. The order information of the present embodiment is a number indicating the order of the plurality of frame images F1 to F5.

  The unique information calculation unit 102 includes one or more different frame images having a fixed relationship on the time axis with respect to a specific one of the plurality of frame images F1 to F5, and includes the plurality of frame images. The unique information holding unit 103 holds the calculated unique information and outputs it to the information embedding unit 104. The information embedding unit 104 embeds the unique information held by the unique information holding unit 103 in a specific frame image.

  Hereinafter, the information embedding device of the present embodiment will be described with further reference to FIG. FIG. 2 is a flowchart of the information embedding apparatus of FIG. In the present embodiment, the moving image data for embedding the tampering detection information includes an intra-coded image (I picture), a forward prediction coded image (P picture), and a bidirectional predicted coded image (B picture). MPEG compressed moving image data is used.

  First, the unique information calculation unit 102 calculates unique information based on a frame image of input moving image data (Step 201).

  When the frame image is an I-picture (intra-coded image), the unique information calculation unit 102 performs hashing on all the component values of all the blocks constituting the frame image except for the component in which the digital watermark is embedded. The unique information (unique information I) is calculated using the function.

  When the frame image is a P picture or a B picture (predicted coded image), the unique information calculation unit 102 calculates a hash function from the frequency components of all the blocks in all the P pictures or B pictures included in the unit section. Is used to calculate unique information (unique information BP).

  In the present embodiment, a unit section on the time axis is a section from the I picture to the immediately preceding image of the next I picture in the storage order. Of course, a section of a frame image belonging to one GOP (group of pictures) may be set as a unit section.

  Further, the unique information I and the unique information BP are each 128-bit data. FIG. 3 shows the relationship between the unique information and the picture for which the unique information is calculated.

  Next, the unique information holding unit 103 holds the unique information calculated in Step 201 (Step 202).

  Next, the order information input unit 101 determines whether or not the processing of the unit section has been completed (step 203). When the processing has been completed, the order information input unit 101 inputs the order information according to a predetermined rule. (Step 204). In the present embodiment, as the order information, a number that increases by “1” in order of “0”, “1”, and “2” is used. The order information is 16-bit data.

  Next, as shown in FIG. 3, the information embedding unit 104 uses the unique information I and unique information BP held in step 202 and the number input in step 204 as tampering detection information, and Is embedded as a digital watermark in the first I picture (step 205).

  Next, an embedding method will be described with reference to FIG. FIG. 4 shows a method of assigning tampering detection information in block units. Each rectangle in the frame to be embedded in FIG. 4 indicates a block of 8 pixels × 8 pixels.

  The tampering detection information in the present embodiment is obtained by simply connecting the unique information I, the unique information BP, and the number, and the total number of bits is 272 bits.

  The information embedding unit 104 selects a block of several bits of the falsification detection information information, and sets the selected block as an embedding target block. In FIG. 4, the embedding target block is indicated by a shaded rectangle. The information embedding unit 104 assigns one bit of falsification detection information to each of the selected blocks.

  FIG. 5 shows each frequency component of the embedding target block. The information embedding unit 104 selects one component of the frequency component to be configured in the embedding target block, and embeds the assigned bits. In the present embodiment, the information embedding unit 104 selects the frequency component AC3 in FIG. If the assigned bit is “0”, the information embedding unit 104 changes the value of the frequency component AC3 to an even number, and if “1”, the information embedding unit 104 changes the value of the frequency component AC3 to an odd number. change.

  Finally, the information embedding unit 104 determines whether or not the input of the moving image data has been completed (step 206). If the input has been completed, the process ends. If not, the process returns to step 201.

(Embodiment 2)
In the second embodiment, the falsification detection information is detected from moving image data in which the falsification detection information is embedded as a digital watermark in the first embodiment, and the presence or absence of falsification is determined.

  Hereinafter, a second embodiment of the present invention will be described with reference to the drawings. FIG. 6 is a block diagram of a falsification detection device according to Embodiment 2 of the present invention.

  As shown in FIG. 6, the falsification detection device includes an input unit 501, a unique information calculation unit 502, a unique information storage unit 503, a detection unit 504, a detection information storage unit 505, a comparison unit 506, a continuity determination unit 507, and a falsification determination. A unit 508 is provided.

  The input unit 501 inputs moving image data in which order information is embedded. The detection unit 504 detects order information from the moving image data input to the input unit 501. The detection information holding unit 505 holds the order information detected by the detection unit 504.

  The continuity determining unit 507 determines whether the continuity of the frame image is maintained based on the order information held by the detection information holding unit 505. More specifically, the continuity determining unit 507 determines that there is no continuity when the order of the frame images is changed, dropped, or added, and otherwise determines that there is continuity.

  The alteration determination unit 508 determines whether the moving image data has been altered based on the determination result of the continuity determination unit 507.

  The unique information calculation unit 502 detects the unique information embedded in a specific frame image among the plurality of frame images in the moving image data input to the input unit 501, and obtains first unique information. In addition, the unique information calculation unit 502 calculates unique information regarding one or more different frame images having a fixed relationship on the time axis with respect to a specific frame image and included in the plurality of frame images. Obtain second unique information. These first and second unique information are held in the unique information holding unit 503.

  The comparing unit 506 compares the first unique information and the second unique information held in the unique information holding unit 503. The falsification determination unit 508 determines whether the moving image data has been falsified based on the comparison result of the comparison unit 506.

  Here, the falsification determination unit 508 determines that the moving image data has not been falsified when the comparison result that the first unique information and the second unique information are equal is input from the comparison unit 506.

  Hereinafter, the falsification detection device of the present embodiment will be described with further reference to FIG. FIG. 7 is a flowchart of the tampering detection device of FIG.

  First, the input unit 501 inputs a frame image constituting moving image data in which falsification detection information is embedded (step 601).

  Next, the unique information calculation unit 502 calculates unique information from the input frame image (step 602). The processing in this step is the same as the processing in step 201 (see FIG. 2) in the first embodiment, and thus the description is omitted.

  Next, the unique information holding unit 503 holds the unique information calculated in Step 201 (Step 603).

  Next, the unique information calculation unit 502 determines whether or not the processing of the unit section has been completed (step 604). When the processing has been completed, the detection unit 504 determines the information embedded from the first I picture of the unit section. Detect (Step 605). Here, the unique information calculation unit 502 first selects an embedding target block using the same selection method as that used at the time of embedding.

  Next, in the selected block, the bit value is set to “0” when the value of the frequency component selected at the time of embedding is an even number, and the bit value is set to “1” when the value is odd. Detected information. That is, the unique information calculation unit 502 performs a process reverse to step 205 in the first embodiment.

  Next, the detection information holding unit 505 holds the information detected in Step 605 (Step 606).

  Next, the continuity determining unit 507 determines whether the bit string corresponding to the number in the detection information held in step 606 is continuous by comparing the bit string corresponding to the number with a predetermined rule indicating continuity (step 607). ). Here, the rule indicating the continuity is the same as that at the time of embedding. For example, if the numbers to be consecutively detected from the I pictures in the unit section are always increasing by “1” in the order of “0”, “1”, and “2”, the continuity determining unit 507 determines It is determined that there is, and otherwise, it is determined that there is no continuity.

  Next, the comparison unit 506 compares and matches the unique information I and unique information BP held in step 603 with the bit strings corresponding to the unique information I and unique information BP among the detection information held in step 606. Then, the difference is output (step 608).

  Next, the falsification determination unit 508 determines that there is no falsification if it is determined that there is continuity in step 607 and the difference output in step 608 is “0”, and if not, falsification. It is determined that there is (step 609).

  Finally, the falsification determination unit 508 determines whether or not the input of the moving image data has been completed (step 610). If the input has been completed, the process ends. If not, the process returns to step 601.

  As described above, according to the information embedding device according to the first embodiment and the tampering detection device according to the second embodiment, the embedded order information is detected and collated with the rule indicating the same continuity as at the time of embedding. Thus, tampering on the time axis can be detected.

  Further, at the time of embedding information, the unique information BP of the P picture and the B picture is embedded in another I picture, and at the time of detecting a digital watermark, the unique information BP calculated by the same method as at the time of embedding is compared with the detected information. This makes it possible to detect spatial tampering of the P picture and the B picture. Even when there are almost no coded blocks in the P picture and the B picture, it is possible to detect spatial tampering of the P picture and the B picture by embedding all the blocks in the coded I picture. .

  Further, by embedding the unique information of all the P pictures and B pictures included in the unit section in another I picture, tampering on the time axis of the P picture and the B picture can be detected. For example, if several P pictures and B pictures are omitted, the detected information and the unique information BP calculated at this time are different at the time of detecting a digital watermark, and tampering can be found.

  Further, the unique information I of the I picture itself, which is the embedding target image, is embedded in the I picture, and the unique information I calculated by the same method as that at the time of digital watermark detection is compared with the detected information. It is possible to detect tampering in the space of the picture itself from one picture alone.

  In the present embodiment, MPEG encoded data is used as moving image data. However, the present invention is not limited to this, and can be similarly applied to, for example, motion JPEG.

  In the information embedding device according to the first embodiment, the falsification detection information is embedded as a digital watermark in the frequency component of the selected block representing the pixel, but the falsification detection information is directly stored in the header section 10 of the moving image data. It may be added. For example, in step 205, the tampering detection information is added to the extension area 10a included in the header section 10 of the MPEG encoded data, and in step 605 of the tampering detection device according to the second embodiment, the information added from the extension area 10a is added. The same effect can be obtained by detecting.

  In step 201, a 128-bit hash value calculated from the frequency component is used as the unique information, but the present invention is not limited to this. The same effect can be obtained if the information is unique to each image. Further, one piece of unique information is calculated from all P pictures and B pictures included in the unit section. However, unique information may be calculated independently, or unique information may be calculated in units of a plurality of images. Further, the section from the I picture to the immediately preceding picture of the next I picture in the storage order is set as the unit section, but the present invention is not limited to this.

  Further, in step 203, as the information indicating the continuity, a 16-bit number that is incremented by “1” is used. The number is not limited to a number and may be, for example, alphabetical order, a character of a meaningful sentence, an appropriate numerical sequence, or the like, as long as the rule conforms to a common rule at the time of digital watermark embedding and at the time of digital watermark detection.

  Also, in step 205, as a method of embedding the tampering detection information, a method has been described in which one bit is assigned to a selected block and one frequency component of the block is changed, but this is only an example. It is desirable to use an embedding method that suppresses image quality degradation that does not affect the original moving image data as much as possible.

  In step 205, embedding is performed on the head I picture of the unit section. However, the embedding is not limited to the I picture. If possible, embedding may be performed on a P picture or a B picture.

  Also, in step 609, if it is determined in step 607 that the data is continuous, and if the difference output in step 608 is “0”, it is determined that there is no tampering. By making the determination, it is also possible to distinguish between mere irreversible image conversion and malicious tampering.

  Embodiments 1 and 2 can be further variously modified. Hereinafter, each mode of information embedding and falsification detection will be described with reference to FIGS. 9 to 14.

(Embodiment 3)
FIG. 9 is an explanatory diagram of information embedding and falsification detection according to Embodiment 3 of the present invention. In this embodiment, the order information is embedded as an electronic watermark in each of the frame images F1 to F5 constituting the data section 20 of the moving image data. Also, at the time of falsification detection, an embedded digital watermark is detected from each of the frame images F1 to F5, and order information is obtained.

  If there is an abnormality in the obtained order information, it is determined that there is tampering on the time axis, otherwise, it is determined that there is no tampering on the time axis.

(Embodiment 4)
FIG. 10 is an explanatory diagram of information embedding and falsification detection according to Embodiment 4 of the present invention.

  In the present embodiment, unlike the third embodiment, the order information of the frame images F1 to F5 is stored in the extension area 10a of the header section 10, whereby the falsification is determined. Even in this case, the order information cannot be visually observed.

(Embodiment 5)
FIG. 11 is an explanatory diagram of information embedding in the fifth embodiment of the present invention, and FIG. 12 is an explanatory diagram of tampering detection in the fifth embodiment of the present invention.

  In the present embodiment, as shown in FIG. 11, based on a previous frame image F1 (another frame image) located immediately before the current frame image F2 (a specific frame image) of interest on the time axis, Specific information is calculated. Then, the calculated unique information is embedded as a digital watermark in the current frame image F2.

  At the time of falsification detection, as shown in FIG. 12, the digital watermark embedded in the current frame image F2 is detected, and the first unique information is obtained.

  Further, unique information is calculated from the previous frame image F1, and second unique information is obtained.

  Then, the first and second unique information are compared, and if they are equal, it is determined that there is no tampering on the time axis, otherwise, it is determined that there is tampering on the time axis. Further, even when the frame image itself has been tampered with (first tampering), the same relationship holds as described above, and therefore, according to this embodiment, both the first and second tampering can be detected at once. .

  In this way, the previous frame image F1 and the current frame image F2 have a so-called chained relationship. If this is repeated for all frames, all frames will be connected by chains. In that case, it is possible to counter all tampering on the time axis. This is because any time the tampering is performed, any part of the chain is cut, and this cut can be easily detected.

(Embodiment 6)
FIG. 13 is an explanatory diagram of information embedding according to the sixth embodiment of the present invention, and FIG. 14 is an explanatory diagram of falsification detection according to the sixth embodiment of the present invention.

  In the present embodiment, unlike Embodiment 5, the unique information is not embedded in the frame image as a digital watermark, but is stored in the extension area 10a of the header section 10. That is, in the present embodiment, processing is performed on the fifth embodiment so that the third embodiment and the fourth embodiment have a relationship.

  Needless to say, Embodiments 3 to 6 can be implemented alone or can be implemented in appropriate combinations, and any aspect is included in the present invention.

  Typically, as shown in FIG. 8, each function realized by the information embedding device and the falsification detecting device according to the first and second embodiments is a storage device 702 (ROM, ROM, This is realized by a CPU 701 (central processing unit) that executes the program data. In this case, each program data may be introduced via a storage medium 705 such as a CD-ROM or a flexible disk.

  INDUSTRIAL APPLICABILITY The information embedding device according to the present invention can be suitably used, for example, in a field (monitoring record, copyright management) in which moving image data is handled while ensuring the presence or absence of falsification with high credibility.

1 is a block diagram of an information embedding device according to Embodiment 1 of the present invention. Flowchart of information embedding device according to Embodiment 1 of the present invention Explanatory drawing of information embedding in Embodiment 1 of the present invention. Explanatory drawing showing the relationship between falsification detection information and an embedding target block according to Embodiment 1 of the present invention. Explanatory drawing of an embedding target block in Embodiment 1 of the present invention. FIG. 4 is a block diagram of a falsification detection device according to Embodiment 2 of the present invention. Flowchart of falsification detection device according to Embodiment 2 of the present invention 1 is a block diagram of a recording medium according to Embodiments 1 and 2 of the present invention. Explanatory drawing of information embedding and falsification detection in Embodiment 3 of the present invention. Explanatory drawing of information embedding and falsification detection in Embodiment 4 of the present invention. Explanatory drawing of information embedding in Embodiment 5 of the present invention. Explanatory drawing of falsification detection in Embodiment 5 of the present invention. Explanatory drawing of information embedding in Embodiment 6 of the present invention Explanatory drawing of falsification detection in Embodiment 6 of the present invention. Block diagram of conventional tampering detection device

Explanation of reference numerals

101 Order information input unit 102 Unique information calculation unit 103 Unique information storage unit 104 Information embedding unit 501 Input unit 502 Unique information calculation unit 503 Unique information storage unit 504 Detection unit 505 Detection information storage unit 506 Comparison unit 507 Continuity determination unit 508 Tampering Determination unit 701 CPU
702 storage device 703 bus 704 drive 705 recording medium 801 key input unit 802 frequency conversion unit 803 falsification verification data embedding unit 804 control unit 805 external storage device 806 quantization unit 807 block division unit 808 Huffman encoding unit 809 image data input unit 810 pseudo-random number generator 811 falsification verification data generator 812 memory 813 data display

Claims (25)

An apparatus for embedding information in moving image data including a data unit configured to be able to sequentially arrange a plurality of frame images on a time axis and a header unit holding information on the data unit,
An order information input unit that generates order information of the plurality of frame images,
An information embedding device, comprising: an information embedding unit that embeds the order information generated by the order information input unit in the moving image data so that the order information is not visible in the plurality of frame images. The information embedding device according to claim 1, wherein the information embedding unit embeds the order information as an electronic watermark in the plurality of frame images. The information embedding device according to claim 1, wherein the information embedding unit embeds the order information in the header portion. 4. The information embedding device according to claim 1, wherein the order information is a number indicating an order of the plurality of frame images. A data unit configured to be capable of sequentially arranging a plurality of frame images on a time axis, and a device that detects whether there is tampering of moving image data including a header unit that holds information on the data unit,
An input unit for inputting the moving image data in which the order information is embedded,
A detection unit that detects the order information from the moving image data input to the input unit,
An alteration detection device comprising: an alteration determination unit that determines whether the moving image data has been altered based on the order information detected by the detection unit. The falsification detection device according to claim 5, wherein the order information is a number indicating an order of the plurality of frame images. 7. The falsification detection device according to claim 5, wherein the falsification determination unit determines that the moving image data has been falsified when the order of the plurality of frame images is changed, dropped, or added. An apparatus for embedding information in moving image data including a data unit configured to be able to sequentially arrange a plurality of frame images on a time axis and a header unit holding information on the data unit,
In the moving image data, for a specific frame image of the plurality of frame images, a single or a plurality of different frame images in a certain relationship on the time axis, the plurality of frame images A unique information calculation unit for calculating unique information on what is included,
An information embedding device comprising: an information embedding unit that embeds the calculated unique information in the specific frame image. 9. The information embedding device according to claim 8, wherein the unique information is a hash value regarding the another frame image. 10. The information embedding apparatus according to claim 8, wherein the another frame image is a frame image adjacent to the specific frame image on the time axis. The moving image data is MPEG-compressed moving image data, the specific frame image is an intra-coded image, and the one or more different frame images are one or more prediction-coded images. Item 10. The information embedding device according to Item 8 to 9. The information embedding device according to claim 11, wherein the intra-coded image and the one or more predictive-coded images belong to one GOP (group of pictures). A data unit configured to be capable of sequentially arranging a plurality of frame images on a time axis, and a device that detects whether there is tampering of moving image data including a header unit that holds information on the data unit,
An input unit for inputting the moving image data in which the order information is embedded,
In the moving image data input to the input unit, first information is obtained by detecting unique information embedded in a specific frame image of the plurality of frame images, and the first specific information is obtained. A unique information calculating unit that calculates unique information on one or more different frame images having a fixed relationship on the time axis and is included in the plurality of frame images to obtain second unique information,
A comparing unit that compares the first unique information with the second unique information;
An alteration detection device comprising: an alteration determination unit that determines whether the moving image data has been altered based on a comparison result of the comparison unit. 14. The falsification detection device according to claim 13, wherein the falsification determination unit determines that the moving image data has not been falsified when the first unique information and the second unique information are equal. 15. The falsification detection device according to claim 13, wherein the unique information is a hash value related to the another frame image. 16. The falsification detection device according to claim 13, wherein the another frame image is a frame image adjacent to the specific frame image on the time axis. The moving image data is MPEG-compressed moving image data, the specific frame image is an intra-coded image, and the one or more different frame images are one or more prediction-coded images. Item 16. A falsification detection device according to Item 13 to Item 15. 18. The falsification detection device according to claim 17, wherein the intra-coded image and the one or more predictive-coded images belong to one GOP (group of pictures). A data part configured to be able to sequentially arrange a plurality of frame images on a time axis, and a method for detecting the presence or absence of tampering of moving image data including a header part holding information on the data part,
Embedding the order information of the plurality of frame images in the moving image data so as to be invisible in the plurality of frame images,
Detecting the order information from the moving image data;
Determining whether the moving image data has been tampered with, based on the detected order information. A data part configured to be able to sequentially arrange a plurality of frame images on a time axis, and a method for detecting the presence or absence of tampering of moving image data including a header part holding information on the data part,
For a specific frame image of the plurality of frame images, a step of calculating unique information about another frame image having a fixed relationship on the time axis and included in the plurality of frame images,
Embedding the calculated unique information in the specific frame image;
Detecting unique information embedded in the specific frame image to obtain first unique information;
Calculating unique information about the another frame image to obtain second unique information;
Comparing the first unique information with the second unique information to determine whether the moving image data has been tampered with. A method for embedding information in moving image data including a data portion configured to be able to sequentially arrange a plurality of frame images on a time axis, and a header portion holding information on the data portion,
An information embedding method including a step of embedding the order information of the plurality of frame images in the moving image data so that the order information cannot be visually recognized in the plurality of frame images. A data part configured to be able to sequentially arrange a plurality of frame images on a time axis, and a method for detecting the presence or absence of tampering of moving image data including a header part holding information on the data part,
Inputting the moving image data in which the order information is embedded;
Detecting the order information from the input moving image data;
Determining whether the moving image data has been tampered with, based on the detected order information. A method for embedding information in moving image data including a data portion configured to be able to sequentially arrange a plurality of frame images on a time axis, and a header portion holding information on the data portion,
Calculate unique information on one or more different frame images having a fixed relationship on the time axis with respect to a specific frame image among the plurality of frame images, which is included in the plurality of frame images. Steps to
Embedding the calculated unique information in the specific frame image. A data part configured to be able to sequentially arrange a plurality of frame images on a time axis, and a method for detecting the presence or absence of tampering of moving image data including a header part holding information on the data part,
Inputting the moving image data in which the order information is embedded;
Detecting unique information embedded in a specific frame image of the plurality of frame images to obtain first unique information;
For the specific frame image, calculating unique information on one or more different frame images having a fixed relationship on the time axis and included in the plurality of frame images, and calculating second unique information. The steps to get
Comparing the first unique information with the second unique information to determine whether the moving image data has been tampered with. 25. A recording medium storing a processing program for realizing the method according to claim 19 to 24 in a computer-readable manner.

Images