WO2009036684A1 - Procédé et dispositif pour l'intégration et la détection d'informations en filigrane - Google Patents

Procédé et dispositif pour l'intégration et la détection d'informations en filigrane Download PDF

Info

Publication number
WO2009036684A1
WO2009036684A1 PCT/CN2008/072308 CN2008072308W WO2009036684A1 WO 2009036684 A1 WO2009036684 A1 WO 2009036684A1 CN 2008072308 W CN2008072308 W CN 2008072308W WO 2009036684 A1 WO2009036684 A1 WO 2009036684A1
Authority
WO
WIPO (PCT)
Prior art keywords
watermark information
zero
data
embedding
frame
Prior art date
Application number
PCT/CN2008/072308
Other languages
English (en)
Chinese (zh)
Inventor
Yongping Zhang
Tao HAO
Original Assignee
Huawei Technologies Co., Ltd.
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from CN200810087514XA external-priority patent/CN101389009B/zh
Application filed by Huawei Technologies Co., Ltd. filed Critical Huawei Technologies Co., Ltd.
Publication of WO2009036684A1 publication Critical patent/WO2009036684A1/fr

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
    • H04N19/157—Assigned coding mode, i.e. the coding mode being predefined or preselected to be further used for selection of another element or parameter
    • H04N19/159—Prediction type, e.g. intra-frame, inter-frame or bidirectional frame prediction
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T1/00—General purpose image data processing
    • G06T1/0021—Image watermarking
    • G06T1/0028—Adaptive watermarking, e.g. Human Visual System [HVS]-based watermarking
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T1/00—General purpose image data processing
    • G06T1/0021—Image watermarking
    • G06T1/0085—Time domain based watermarking, e.g. watermarks spread over several images
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
    • H04N19/18—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being a set of transform coefficients
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
    • H04N19/184—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being bits, e.g. of the compressed video stream
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/46—Embedding additional information in the video signal during the compression process
    • H04N19/467—Embedding additional information in the video signal during the compression process characterised by the embedded information being invisible, e.g. watermarking
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/20—Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
    • H04N21/23—Processing of content or additional data; Elementary server operations; Server middleware
    • H04N21/238—Interfacing the downstream path of the transmission network, e.g. adapting the transmission rate of a video stream to network bandwidth; Processing of multiplex streams
    • H04N21/2389—Multiplex stream processing, e.g. multiplex stream encrypting
    • H04N21/23892—Multiplex stream processing, e.g. multiplex stream encrypting involving embedding information at multiplex stream level, e.g. embedding a watermark at packet level
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/80—Generation or processing of content or additional data by content creator independently of the distribution process; Content per se
    • H04N21/83—Generation or processing of protective or descriptive data associated with content; Content structuring
    • H04N21/835—Generation of protective data, e.g. certificates
    • H04N21/8358—Generation of protective data, e.g. certificates involving watermark

Definitions

  • the present invention relates to the field of digital image processing and information security technologies, and in particular, to a method and an apparatus for embedding and detecting watermark information.
  • Digital watermarking technology refers to embedding identification signals (ie, watermark information) in digitized multimedia content by means of signal processing, so as to extract the identification signals and compare whether the extracted identification signals are consistent with the embedded identification signals. Achieve copyright tracking, copyright claims and content tampering for multimedia content. Therefore, in the era of rapid development of digital multimedia technology, digital watermarking technology can not be ignored in order to prevent digital video content from being illegally falsified or pirated due to the portability of data media.
  • a digital watermark based video authentication method in the prior art is: First, at the embedding end of the digital watermark, a watermark signal is generated from the I frame, and then, when performing variable length coding (VLC, Variable Length Code), The watermark signal is embedded in the I frame. At the authentication end of the digital watermark, watermark information is generated from the I frame, and then compared with the watermark information extracted from the I frame to further verify whether the received video stream is tampered with.
  • VLC variable length coding
  • the watermark information is embedded in the I frame, only the I frame is protected, and the P frame and the B frame cannot be protected, so the protection range of the watermark information is small; meanwhile, since the I frame is encoded, the P frame and the B are The frame has an effect. Therefore, embedding the watermark information into the I frame will affect the visual quality of the entire video content, and if the visual quality is guaranteed, only the protection range of the watermark information can be reduced.
  • An embodiment of the present invention provides a method for embedding watermark information, including:
  • the watermark information is embedded in the non-critical data.
  • an embodiment of the present invention further discloses a method for detecting watermark information, including:
  • the data in the same image group or the same group of the video is divided into the same key data and non-critical data as the embedded end;
  • an embodiment of the present invention further discloses an apparatus for embedding watermark information, including:
  • the shunting unit divides the data in the same group or the image of the video into key data and non-critical data according to the degree of influence on the content of the video or the image;
  • a watermark information generating unit configured to generate a watermark information embedding unit from the key data obtained by the shunting unit, for embedding the watermark information in the watermark information generating unit into the key data.
  • the embodiment of the present invention further discloses a watermark information detecting apparatus, and a package Includes:
  • the shunting unit divides the data in the received image or the same group of the video into the same key data and non-critical data as the embedding according to the degree of influence on the content of the video or the image;
  • a first watermark information unit configured to generate first watermark information from key data obtained by the shunt unit
  • a second watermark information unit configured to extract second watermark information from the non-critical data obtained by the shunting unit
  • a determining unit configured to determine whether the correlation between the first watermark information and the second watermark information meets a preset threshold, and if yes, determine that the video or image is not falsified; The video or image has been tampered with.
  • the embodiment of the invention has the following advantages:
  • data in the same group of images or images in the video is divided into key data and non-critical data according to the degree of influence on the content of the video or the image, since the watermark information is generated from the key data and embedded in the non-key data.
  • the protection range of the watermark information is expanded, and the quality of the video or image is ensured, and the visual influence of the video or the image is reduced, thereby achieving seamless protection of the video or image content.
  • FIG. 1( a ) is a schematic diagram of an embedding process of watermark information in the prior art
  • FIG. 1 (b) is a schematic diagram of a process of detecting watermark information in the prior art
  • FIG. 2 is a flow chart showing a method for embedding watermark information according to an embodiment of the present invention
  • FIG. 3 is a schematic diagram of a process of embedding video coding and watermark information according to an embodiment of the present invention
  • FIG. 4 is a flow chart showing the process of generating and embedding watermark information in FIG. 3;
  • Figure 5 is a schematic diagram of reference relationships of frames in the same group
  • FIG. 6 is a flow chart of a method for detecting watermark information according to an embodiment of the present invention.
  • FIG. 7 is a schematic diagram of a video decoding and video authentication process according to an embodiment of the present invention.
  • FIG. 8 is a schematic flow chart of a process of detecting watermark information in FIG. 7;
  • FIG. 9 is a structural block diagram of an embodiment of an apparatus for embedding watermark information according to the present invention.
  • FIG. 10 is a structural block diagram of another embodiment of embedding apparatus for watermark information according to the present invention.
  • 11 is a structural block diagram of another embodiment of embedding device for watermark information according to the present invention.
  • 12 is a structural block diagram of an embodiment of a device for detecting watermark information according to the present invention
  • FIG. 13 is a structural block diagram of another embodiment of a device for detecting watermark information according to the present invention.
  • a method for embedding watermark information according to the present invention includes: dividing data in the same group of images or images into video data and non-critical data according to the degree of influence on the content of the video or image; The generated watermark information in the key data is embedded in the non-critical data.
  • the frames in the same group of the video are divided into key frames and non-key frames according to the degree of influence on other data frames during encoding, wherein the data in the key frames is key data.
  • the data in the non-key frame is non-critical data.
  • the relatively high frequency data in the image can be split into non-critical data according to the degree of influence of different frequencies on the content of the image, and the relatively low frequency data is divided into key data.
  • FIG. 2 it is an implementation flow of a method for embedding watermark information according to the present invention.
  • the method embeds watermark information generated from a key frame of a video code stream into a non-key frame, and after embedding the watermark information by the method,
  • the authentication terminal restores the original video stream.
  • FIG. 3 and FIG. 4 the schematic diagram of the generation and embedding process of the watermark information shown in FIG. 3 and FIG. 4 can be combined.
  • the method includes the following steps:
  • Step 201 Divide the original video code stream in the same group of pictures (GOP, Group of Pictures) according to a preset shunting rule to obtain key frames and non-key frames.
  • group of pictures GOP, Group of Pictures
  • the key frame and the non-key frame may be determined according to the degree of influence on other data frames during encoding.
  • an I frame is an intra coded frame whose encoding does not depend on already encoded image data.
  • the P frame is a forward predicted frame
  • the B frame is a bidirectional predicted frame.
  • the P frame needs to refer to the I frame
  • the B frame needs to refer to both the I frame and the P frame or the two P frames. It can be seen that the I frame change will directly affect the subsequent P frame and B frame; while the P frame changes, it will affect the B frame encoding before and after.
  • the I frame is a key frame for the P frame and the B frame; and the P frame is a key frame for the B frame, whereas the B frame and the P frame are non-key frames for the I frame, and the B frame is a non-key frame for the P frame.
  • Step 202 Extract a quantization coefficient of an intra prediction residual of each macro block (Macro Block) in the key frame I frame.
  • the intra prediction residual may be obtained by intra prediction according to an intra prediction mode, and then subjected to integer transform and quantization to obtain a quantized coefficient of the intra prediction residual.
  • Step 203 The quantized coefficients of the intra prediction residuals of the macroblocks in the I frame extracted from the step 202 are processed to generate watermark information (abbreviated as wm_signal).
  • the watermark information can be generated as follows:
  • Method 1 Compare the size of the quantized coefficients of the intra prediction residuals of the corresponding blocks in the adjacent two macroblocks in the I frame, and obtain 0 or 1 according to the comparison result, for example, when the comparison result is greater than 1 , otherwise, get 0.
  • the watermark information can then be obtained by an encryption process such as a hash function (i.e., a hash function).
  • the size of the quantization coefficient of the intra prediction residual of the corresponding one of the two macroblocks in the I frame may be implemented by the following method, but is not limited thereto, for example: First, randomly selecting from the I frame A macroblock MBr and a macroblock MBe form a macroblock pair, and a non-zero AC coefficient is randomly selected from the quantized coefficients of the intra prediction residuals of each of the two macroblocks.
  • the block may be an 8x8 block and may include a chroma block.
  • the comparison result is quantized into a binary sequence of 0 or 1. For example: If the macroblock MB r is greater than MB e, then the comparison result quantized 0, otherwise, quantized to 1.
  • the size of the quantized coefficients of the intra prediction residuals in the adjacent two macroblocks may not be compared, and the quantized coefficients of the intra prediction residuals of the macroblocks in the I frame may be directly subjected to encryption processing such as a hash function. Get the watermark information.
  • Method 2 Convert the sequence number information of the data frame to be embedded into a binary number, and perform the modulo-2 addition processing on the bit by bit with the quantized coefficient of the intra prediction residual of the I frame extracted in step 202.
  • the result is watermark information.
  • the method for generating watermark information of the present invention is not limited to the above methods, but the watermark information generated based on the quantized coefficients of the intra prediction residuals of each macroblock in the I frame does not affect the implementation of the present invention.
  • Step 204 Extract each of the non-key frames in the group (P frame and B frame in this embodiment)
  • the inter-frame prediction may be implemented by motion estimation and motion compensation, and the obtained inter-frame prediction residual is subjected to integer transform and quantization to obtain a quantized coefficient of the inter-frame prediction residual, which belongs to common knowledge in the field, and therefore no longer Praise.
  • Step 205 Extract a non-zero AC coefficient (referred to as CoefAC) from the Coef of the inter prediction residual of each block as a candidate embedding domain.
  • CoefAC non-zero AC coefficient
  • Step 206 Select a non-zero AC coefficient as the embedded domain in the CoefAC by using a user password (Key).
  • the specific method for selecting a non-zero AC coefficient may be: converting the user password into a binary number, where the binary number corresponds to the number of bits of CoefAC in the current partition; and selecting a value according to each digit in the binary.
  • the corresponding bit in the CoefAC corresponding to 1 is used as the embedded domain. For example, if the 5th digit value in the binary is 1 , the 5th bit in the CoefAC is selected as the embedded domain.
  • the selection method may also be implemented by other equivalent transformation methods.
  • the decimal point may be used, and the corresponding bit in the CoefAC corresponding to the even number is selected as the embedded domain.
  • Step 207 Extract the wm_signal of the corresponding position in the corresponding macroblock of the I frame according to the location of the macroblock where the CoefAC is selected in the above step 206, and embed the extracted wm_signal into the step 206. The least significant bit in the selected CoefAC. It is then encoded to obtain a video stream of the watermarked information.
  • the method of embedding the wm_signal into the least significant bit of the CoefAC may be implemented according to the formula (1), that is, the CoefAC is taken as an absolute value and multiplied by 2 (ie, shifted one bit to the left) to stay. The lowest bit is embedded for the wm_signal, then the wm_signal is embedded in the lowest bit, and finally, the embedded value is assigned the same sign as the CoefAC before embedding.
  • the embodiment of the present invention embeds the watermark information generated in the key frame I frame into all non-key frames in the same group as the I frame, so that all the video code streams except the key frame
  • the data frame is embedded with watermark information, thus not only expanding the protection range of the watermark information, but also improving
  • the watermark information can protect the video stream, and can also effectively reduce the visual impact of the watermark information on the video, thereby achieving seamless protection of the video content.
  • the embodiment of the present invention uses the user password to control the embedding position of the watermark information, when extracting the authentication, it is also necessary to use the same user password to improve the watermark information of the location, thereby further increasing the security of the video stream. Sex, and can locate the tamper-resistant macroblock level.
  • the embedding of the watermark information is performed after the entropy coding, that is, the encapsulation outside the coding loop, so that the embedding of the watermark information does not affect the video content, thereby further ensuring the video quality.
  • the embodiment of the present invention may first shift the CoefAC to be embedded into an absolute value and then shift it to the left by one bit, and then embed the watermark information into the lowest bit.
  • the original video stream can be restored, so that the distortion of the original video caused by embedding the watermark information can be minimized.
  • the macroblock needs to be selected from the macroblock.
  • a block of zero AC coefficients is used as a block of watermark information to be embedded, and then in step 205, a candidate embedding field is selected from the selected blocks of watermark information to be embedded.
  • the method for selecting a block to be embedded with the watermark information from the macroblock for example: selecting a brightness 8x8 block having a non-zero AC coefficient as the to-be-embedded in the macro block according to the user password in the step 206 Blocking of watermark information.
  • the selection method may be: calculating a remainder of the user password and the number of luma blocks having a non-zero AC coefficient, and the result of adding 1 to the remainder is the brightness of all non-zero AC coefficients of the macroblock to be embedded in the macroblock. The position in the block.
  • the user password may be adjusted to the password related to the chrominance information in the above step 206, for example: calculating the user password, The current frame code, the current macroblock number, and the XOR value of the non-zero AC coefficient of the selected chrominance U and V blocks (ie, the XOR value) are the adjusted password.
  • the remainder of the number of non-zero AC coefficients in the user password and the chrominance U and V blocks can be calculated, and the result obtained by adding 1 to the remainder is the selected non-zero AC coefficient in the chrominance U and V blocks.
  • the non-zero AC coefficients of the chrominance U and V blocks are selected from the locations in the chrominance U and V blocks.
  • the method of selecting the embedded domain in step 206 may further: calculate a remainder of the adjusted password and the number of non-zero AC coefficients of the watermark partition to be embedded, and the result of adding 1 to the remainder is an embedded domain.
  • step 207 after extracting the wm_signal of the corresponding position in the corresponding macroblock of the I frame, the wm_signal and the The XOR value of the high order value in the CoefAC selected in the step 206.
  • the calculated result is then embedded in the least significant bit of the CoefAC selected in step 206. It is then encoded to obtain a video stream of water-printed information.
  • step 202 of the foregoing embodiment if the I frame does not adopt the intra prediction technique, the quantized coefficients of each macroblock in the I frame may be directly extracted, that is, the macroblocks of the I frame are directly directly.
  • the pixel values in the integer are transformed to obtain coefficients, and then the coefficients are quantized to obtain quantized coefficients of the macroblocks.
  • the quantized coefficients of the macroblocks in the I frame may be used to process the watermark information.
  • the method for detecting watermark information generates a first watermark information from a key frame of a video code stream, and then compares it with a second watermark information extracted in a non-key frame to determine Whether the current video stream has been tampered with.
  • the key frame and the non-key frame are obtained according to the same shunting rule as the embedded end.
  • a schematic diagram of the extraction and detection process of the watermark information shown in FIG. 7 and FIG. 8 may be combined, and the method includes:
  • Step 601 After the video code stream of the received watermark information is entropy decoded, the key frame and the non-key frame are obtained according to the same shunting rule as the embedded end.
  • the I frame is used as the key frame in the same group, and the P frame and the B frame are used as the non-key frames as an example.
  • Step 602 Extract the quantization coefficient of the intra prediction residual of each macroblock in the key frame (ie, I frame).
  • Step 603 The watermark information generating method is the same as that of the embedded end, and the quantized coefficients of the intra prediction residuals of the macroblocks in the I frame extracted from the step 602 are processed to generate a first watermark.
  • Information this embodiment is abbreviated as wm_signal').
  • method 1 uses the method 1 to generate first watermark information, that is, first randomly select two macroblocks from the I frame, respectively Randomly selecting a non-zero AC coefficient among the quantized coefficients of the intra prediction residuals of each block of the macroblocks, and respectively calculating the sum of the selected non-zero AC coefficients in each macroblock, and then comparing the two macros The size of the sum of the non-zero AC coefficients of the block.
  • first watermark information that is, first randomly select two macroblocks from the I frame, respectively Randomly selecting a non-zero AC coefficient among the quantized coefficients of the intra prediction residuals of each block of the macroblocks, and respectively calculating the sum of the selected non-zero AC coefficients in each macroblock, and then comparing the two macros The size of the sum of the non-zero AC coefficients of the block.
  • Step 604 Extract a quantization coefficient (referred to as Coef) of the inter prediction residual of each of the macroblocks in each non-key frame in the picture group (P frame and B frame in this embodiment).
  • Coef a quantization coefficient
  • Step 605 extracting a prediction residual nonzero AC coefficients Coef (abbreviated as CoefAC ') 0 from each of the inter-block step 604
  • Step 606 Select a non-zero AC coefficient in the CoefAC to extract the second watermark information by using a user password (Key) using the same selection method as the embedded end.
  • the user password is the same as the user password of the embedded end. Therefore, in the embodiment, the user password is converted into a binary number, and according to the value of each digit in the binary, the CoefAC corresponding to the value 1 is selected. The corresponding bit in 'the second watermark information is extracted. For example, if the 5th digit value in the binary is 1 , the second watermark information is extracted from the 5th bit in the CoefAC.
  • Step 607 Extract second watermark information (abbreviated as wm_signal in this embodiment) from the least significant bit in the selected CoefAC'.
  • the method for extracting the watermark information can be implemented according to the formula (2), that is, after the selected CoefAC' takes the absolute value, the modulo is taken for 2, and the obtained remainder is the second watermark information.
  • Step 608 Calculate the correlation value of wm_signal and wm-extract of the corresponding block in the same macroblock, and obtain the relevant result. Determining whether the correlation result is greater than or equal to a preset threshold, and if so, outputting "the video stream is not tampered with”; otherwise, "the video stream is tampered with” may be output.
  • the correlation value of the wm_signal' and the wm-extract may be calculated according to the formula (3). If the correlation value is 1, the output "the video stream has no tampering"; if the correlation value is less than 1, the output is " The video stream is tampered with.”
  • the watermark information of each macroblock can be used to determine the Whether the macroblock has been tampered with, when a macroblock is tampered with, it can also output "X frames X macroblocks have been tampered with". In addition, it is also possible to change all the pixels in the falsified macroblock to 255 when the video is displayed to identify the tampering area. In addition, if there is a non-zero AC coefficient in the macroblock, the method in the foregoing method needs to select a non-zero AC coefficient in the macroblock from the macroblock before performing step 605. The block is used as a block of the watermark information to be measured, and then the step 605 is performed to select a candidate detection domain from the selected block of the watermark information to be detected.
  • the method for selecting a block of the watermark information to be detected from the macroblock may select, according to the user password in the step 606, a brightness 8x8 block having a non-zero AC coefficient as the to-be-detected in the macro block.
  • the selection method may be: calculating a remainder of the user password and the number of luma blocks having a non-zero AC coefficient, and the result of adding 1 to the remainder is the brightness of all non-zero AC coefficients of the water block to be detected in the macroblock. The position in the block.
  • the adjusted password selection detection field is used in the embedding step 206, in the above step 606, it is also necessary to select a non-zero AC coefficient in the CoefAC' using the adjusted password, and the password adjustment is performed.
  • the method of selecting a non-zero AC coefficient method according to the adjusted password and the embedding end for example: adjusting the preset user password to a password associated with the chroma information in the same manner as the embedding end, and then adjusting the The number of non-zero coefficients of the block of the watermark information to be detected in each macroblock of the non-key frame is increased by 1 and the result is the position of the non-zero AC coefficient to be detected.
  • the watermark information embedded in the embedded end is an exclusive OR value of the wm_signal extracted from the corresponding macroblock of the I frame and the high value of the selected CoefAC to be embedded, then the first watermark information is generated in the above step 603. After wm_signal, it is also necessary to calculate an exclusive OR value of the wm_signal' and the high-order value in the CoefAC' selected in step 605.
  • step 608 calculating the XOR value of the corresponding block in the same macroblock and the correlation value of the wm-extract, and obtaining a correlation result, thereby determining whether the correlation result is greater than or equal to a preset threshold. value.
  • the CoefAC of each macroblock in the frame to be embedded takes an absolute value and then shifts to the left by one bit, and then embeds the watermark information at the lowest bit. Therefore, the received video stream can be restored to the original pixel at the authentication end.
  • CoefAC [biliary] [i] sign(abs(CoefAC [Blk [ ]) » Key ⁇ i ,CoefAC[Blk ⁇ i ) ( 4 ) where, in the formula (4), sign ( x, y ) is y The sign of positive and negative is assigned to x, Blk is the serial number of the block in the macro block, i is the serial number of the coefficient in the block, and "" means shifting one bit to the right.
  • the watermark information is generated from the key data, thereby not only expanding the protection scope of the watermark information, but also improving
  • the watermark information has the ability to protect the video stream, and can also effectively reduce the visual impact of the watermark information on the video, thereby achieving seamless protection of the video content.
  • the present invention requires the use of the same user password as the embedded end to extract the watermark information to the corresponding location, the security of the video stream can be further increased, and the falsified macroblock level can be located.
  • the CoefAC' after extracting the second watermark information, the CoefAC' can be restored to the original video by right shifting, so that the distortion of the original video caused by embedding the watermark information can be minimized.
  • the I frame is used as the key frame
  • the P frame and the B frame are used as the non-key frames.
  • an I frame and a P frame may be used as a key frame
  • a B frame may be used as a non-key frame
  • an I frame and a partial P frame may be used as a key frame
  • another part of the P frame and the B frame may be regarded as a non-key frame
  • I frame is a key frame for the P frame and the B frame.
  • the P frame is applicable to any combination of the B frame being a key frame principle.
  • the B frame is obtained as the size of the non-off quantization coefficient, and the inter prediction residuals of the corresponding blocks in the two macroblocks of the P frame are compared.
  • the size of the quantized coefficient is obtained. For example, if the quantized coefficient of an intra prediction residual in the A partition is larger than the corresponding coefficient in the B partition, the current watermark information bit is set to 1, otherwise it is 0; wherein, the number of bits of the watermark information It can be determined according to the block size and security requirements.
  • the block size of H.264 is 4x4, and the number of bits of the watermark information of each macro block can be set to 6 bits; the block size of AVS is 8x8, then each bit can be The number of bits of the watermark information of each macroblock is set to 16 bits.
  • the method for comparing the size of the quantized coefficients of the inter prediction residuals of the corresponding blocks in the two macroblocks of the P frame and the corresponding points of the two macroblocks in the I frame are compared in step 203 of the method embodiment shown in FIG.
  • the method of the size of the quantized coefficient of the inter-frame prediction residual of the block Similarly, for example: first randomly selecting two macroblocks from the P frame, and randomly selecting a non-zero AC coefficient from the quantized coefficients of the inter prediction residuals of each of the two macroblocks;
  • the partition may be an 8x8 block and may include a chroma block.
  • the sum of the non-zero AC coefficients selected in each macroblock is separately calculated, and the sum of the non-zero AC coefficient sums of the two macroblocks is compared, and the comparison result is quantized into a binary sequence of 0 or 1.
  • a plurality of watermark information is generated from the I frame and the P frame in the same GOP, it may be embedded as follows: Two reference frames to be referenced when encoding the B frame The generated watermark information is processed by adding by bit modulo 2 to obtain a watermark information, and then the obtained watermark information is embedded in the B frame.
  • the watermark information can be obtained by comparing the size of the quantized coefficients of the intra prediction residuals of the corresponding blocks in the two macroblocks of the I frame. And for a partial P frame as a key frame, obtained by comparing the size of the quantization coefficient of the inter prediction residual of the corresponding block of the two macroblocks in the frame. Among them, how many P frames are taken as key frames, which can be determined according to the actual situation. For example, the first half of the P frame in the same GOP can be used as the key frame, and the second half of the P frame is used as the non-key frame.
  • the obtained watermark information is sequentially embedded into the subsequent non-key frame P frame; or, the obtained watermark information can be simply calculated by bitwise modulo 2 addition to obtain a watermark information.
  • the resulting watermark information is then embedded into subsequent non-keyframe P frames.
  • the second method can reduce the amount of memory stored.
  • the foregoing embodiments are all based on the method of embedding and detecting the watermark information of the video code stream.
  • the embodiment of the present invention may also be applied to the static image, and the method for embedding the watermark information at the embedding end includes: The degree of influence of the content of the image, the relatively high frequency data in the image is divided into non-critical data, and the relatively low frequency data is divided into key data; the watermark information is generated from the key data, and the non-critical data is non-zero in the quantized coefficient of the non-critical data.
  • the AC coefficient a non-zero AC coefficient to be embedded is selected according to a preset user password; the watermark information is embedded in the lowest bit of the selected non-zero AC coefficients.
  • the method for detecting the watermark information includes: dividing the data in the image into key data and non-critical data according to the same shunting rule as the embedded end; generating watermark information from the key data, in the non-critical data Among the non-zero AC coefficients in the quantized coefficients, the same user secret as the embedded end is used. Selecting a non-zero AC coefficient to be detected; extracting second watermark information from a lowest bit of the selected non-zero AC coefficients; and finally determining a correlation between the first watermark information and the second watermark information Whether the preset threshold is met, if it is satisfied, it is judged that there is no tampering; otherwise, the judgment is falsified.
  • the method for generating and embedding the watermark information at the embedded end, and the method for generating the first watermark information at the authentication end, the method for extracting the second watermark information, and the method for determining the same are the same as those of the video stream, and will not be described again.
  • the steps in implementing the above embodiments may be performed by a program to instruct related hardware, and the program may be stored in a computer readable storage.
  • the software corresponding to the program can also be sold or used as a stand-alone product.
  • the storage medium may be a ROM/RAM, a magnetic disk, an optical disk, or the like.
  • the embodiment of the present invention further discloses a device for embedding watermark information.
  • the device includes: a splitting unit 901, a watermark information generating unit 902, and an embedding unit 903.
  • the offloading unit 901 is configured to divide data in the same group of images or in the image into key data and non-critical data according to the degree of influence on the content of the video or the image;
  • the watermark information generating unit 902 is configured to generate watermark information from the key data obtained by the shunting unit 901;
  • the embedding unit 903 is configured to embed the watermark information generated in the watermark information generating unit 902 into the non-critical data.
  • the above embodiment divides the data in the same group or the image of the video into the key data and the non-critical data according to the degree of influence on the content of the video or the image by the shunting unit 901. Then, the embedding unit 903 embeds the watermark information generated from the key data into other non-critical data except the key data, which not only expands the protection scope of the watermark information, but also improves the protection capability of the watermark information to the video or the image. Moreover, the visual impact of the watermark information on the video or the image can be effectively reduced, and the quality of the video or image is ensured, thereby achieving seamless protection of the video or image content.
  • the embodiment of the present invention further discloses a device for embedding watermark information.
  • the shunting unit 901 includes:
  • the video offloading unit 9011 is configured to divide the frames in the same group of the video into key frames and non-key frames according to the degree of influence on other data frames during encoding, where the data in the key frames is Key data, the data in the non-key frame is non-critical data.
  • the device further includes:
  • the selecting unit 1001 is configured to select, to be embedded, a non-zero AC coefficient in the quantized coefficient of the inter prediction residual of each macroblock in the non-key frame obtained by the video offloading unit 9011 according to the preset user password.
  • Non-zero AC coefficient
  • the embedding unit 903 includes: a first embedding unit 9031, configured to embed the watermark information into a lowest bit of a non-zero AC coefficient selected by the selecting unit 1201.
  • the embedding unit 903 of the above embodiment embeds the watermark information into the lowest bit of the non-zero AC coefficient selected by the selecting unit 1001 based on the user password. Not only can the effect on video or image content be embedded at the time of embedding, but also the user's password can be used to control the embedding position of the watermark information, thereby further increasing the security of the video stream and enabling the tampering macroblock level. Positioning.
  • the embodiment of the present invention further discloses a device for embedding watermark information. As shown in FIG. 11, on the basis of the device embodiment shown in FIG. 10, the device further includes:
  • the non-zero AC coefficient processing unit 1101 is configured to multiply the non-zero AC coefficient selected by the selecting unit 1001 by an absolute value, and then notify the first embedding unit 9031 to embed the watermark information into the processed The lowest bit of the coefficient.
  • the device may further include: a watermark information processing unit 1102, configured to add a plurality of watermark information to a watermark information through modulo-2 addition processing, and notify the first embedded unit 9031 to use the obtained The watermark information is embedded.
  • a watermark information processing unit 1102 configured to add a plurality of watermark information to a watermark information through modulo-2 addition processing, and notify the first embedded unit 9031 to use the obtained The watermark information is embedded.
  • the non-zero AC coefficient processing unit 1101 takes the selected non-zero AC coefficient by an absolute value and multiplies it by 2, that is, shifts one bit to the left, and then embeds the watermark information into the lowest bit of the processed coefficient. on.
  • the original video can be restored by the method corresponding to the embedding, so that the distortion of the original video caused by embedding the watermark information can be minimized.
  • the embodiment of the present invention further discloses a device for detecting watermark information.
  • the device includes: a splitting unit 1201, a first watermark information unit 1202, a second watermark information unit 1203, and a judgment. Unit 1204, wherein
  • the shunting unit 1201 adopts the same shunting rule as the embedding end, according to the video or image
  • the extent of the impact of the content, the data in the same image group in the received image or video is divided into key data and non-critical data;
  • the first watermark information unit 1202 is configured to generate first watermark information from key data obtained by the shunting unit 1201;
  • the second watermark information unit 1203 is configured to extract second watermark information from the non-critical data obtained by the shunting unit 1201;
  • the determining unit 1204 is configured to determine whether the correlation between the first watermark information and the second watermark information meets a preset threshold, and if yes, determine that the video or image is not falsified; The video or image is obtained from being tampered with.
  • the above embodiment is divided by the shunting unit 1201 into data of the same group of pictures or images in the video according to the degree of influence on the content of the video or image into key data and non-critical data. Since the watermark information is generated from the key data and embedded in the non-critical data, it not only expands the protection scope of the watermark information, but also improves the protection capability of the watermark information to the video or image, and can also effectively reduce the watermark information.
  • the visual impact on the video or image guarantees the quality of the video or image, enabling seamless protection of video or image content.
  • the embodiment of the present invention further discloses a device for verifying watermark information.
  • the shunting unit 1201 includes:
  • the video offloading unit 12011 is configured to divide the frames in the same group of the video into key frames and non-key frames according to the degree of influence on other data frames during encoding, where the data in the key frames is key data.
  • the data in the non-key frame is non-critical data.
  • the device also includes:
  • the selecting unit 1301 is configured to select, in the non-zero AC coefficient in the quantized coefficient of the inter prediction residual of each macroblock in the non-key frame obtained by the video offloading unit 12011, using the same user password as the embedded end. Embedded non-zero AC coefficient;
  • the second watermark information unit 1203 includes:
  • the second watermark information sub-unit 12031 is configured to extract the second watermark information from the lowest bit of the non-zero AC coefficients selected by the selecting unit 1301.
  • the above embodiment is based on the watermark information being located at the lowest bit of the non-zero AC coefficient selected by the selection unit 1301 based on the user password. Not only can the watermark information be further reduced to the video or image The effect of the content generation, and since the selection unit 1301 needs to select the extraction location of the watermark information by using the same user password as the embedded end, the security of the video code stream can be further increased, and the tamper-resistant macroblock level can be performed. Positioning.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Security & Cryptography (AREA)
  • Compression Or Coding Systems Of Tv Signals (AREA)
  • Editing Of Facsimile Originals (AREA)

Abstract

La présente invention concerne un procédé et un dispositif pour l'intégration et la détection d'informations en filigrane. Le procédé d'intégration d'informations en filigrane comprend les étapes consistant à : diviser les données du même groupe d'images de la vidéo ou les données d'image en données clés et données non clés en fonction du degré d'influence sur le contenu de la vidéo ou de l'image ; générer des informations en filigrane à partir des données clés ; intégrer les informations en filigrane dans les données non clés. Selon le mode de réalisation de la présente invention, il est possible de diviser les données du même groupe d'images de la vidéo ou les données d'image en données clés et données non clés en fonction du degré d'influence sur le contenu de la vidéo ou de l'image. Et par l'intégration des informations en filigrane générées à partir des données clés dans les données non clés, il est possible non seulement d'étendre la plage de protection des informations en filigrane pour la vidéo ou l'image, mais aussi de réduire efficacement l'influence visuelle des informations en filigrane pour la vidéo ou l'image.
PCT/CN2008/072308 2007-09-14 2008-09-09 Procédé et dispositif pour l'intégration et la détection d'informations en filigrane WO2009036684A1 (fr)

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
CN200710152136.4 2007-09-14
CN200710152136 2007-09-14
CN200810087514XA CN101389009B (zh) 2007-09-14 2008-03-19 一种水印信息的嵌入、检测方法及装置
CN200810087514.X 2008-03-19

Publications (1)

Publication Number Publication Date
WO2009036684A1 true WO2009036684A1 (fr) 2009-03-26

Family

ID=40467516

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2008/072308 WO2009036684A1 (fr) 2007-09-14 2008-09-09 Procédé et dispositif pour l'intégration et la détection d'informations en filigrane

Country Status (1)

Country Link
WO (1) WO2009036684A1 (fr)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3361747A1 (fr) * 2017-02-13 2018-08-15 Markany Inc. Appareil et procédé d'incorporation de filigrane par conversion de structure d'image
CN115002475A (zh) * 2022-05-12 2022-09-02 中国人民武装警察部队工程大学 一种兼顾可逆性和鲁棒性的视频密文域可逆信息隐藏方法
CN116095331A (zh) * 2023-03-03 2023-05-09 浙江大华技术股份有限公司 编码方法以及解码方法

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1455579A (zh) * 2002-04-23 2003-11-12 夏普株式会社 图像处理装置及其系统与方法和电子信息装置
JP2004015626A (ja) * 2002-06-10 2004-01-15 Nippon Hoso Kyokai <Nhk> 電子透かし埋込装置、その検出装置および電子透かし埋込プログラム
CN1893663A (zh) * 2005-09-02 2007-01-10 华为技术有限公司 多媒体通信的传输保护方法
CN101005615A (zh) * 2006-01-18 2007-07-25 华中科技大学 一种图像数据水印信息的嵌入、检测方法及系统
CN101022550A (zh) * 2006-12-26 2007-08-22 北京闪动科技有限公司 流媒体的数字水印信息嵌入方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1455579A (zh) * 2002-04-23 2003-11-12 夏普株式会社 图像处理装置及其系统与方法和电子信息装置
JP2004015626A (ja) * 2002-06-10 2004-01-15 Nippon Hoso Kyokai <Nhk> 電子透かし埋込装置、その検出装置および電子透かし埋込プログラム
CN1893663A (zh) * 2005-09-02 2007-01-10 华为技术有限公司 多媒体通信的传输保护方法
CN101005615A (zh) * 2006-01-18 2007-07-25 华中科技大学 一种图像数据水印信息的嵌入、检测方法及系统
CN101022550A (zh) * 2006-12-26 2007-08-22 北京闪动科技有限公司 流媒体的数字水印信息嵌入方法

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3361747A1 (fr) * 2017-02-13 2018-08-15 Markany Inc. Appareil et procédé d'incorporation de filigrane par conversion de structure d'image
US10134100B2 (en) 2017-02-13 2018-11-20 Markany Inc. Watermark embedding apparatus and method through image structure conversion
CN115002475A (zh) * 2022-05-12 2022-09-02 中国人民武装警察部队工程大学 一种兼顾可逆性和鲁棒性的视频密文域可逆信息隐藏方法
CN116095331A (zh) * 2023-03-03 2023-05-09 浙江大华技术股份有限公司 编码方法以及解码方法

Similar Documents

Publication Publication Date Title
CN101389009B (zh) 一种水印信息的嵌入、检测方法及装置
Xu et al. Steganography in compressed video stream
Zhang et al. Video watermark technique in motion vector
JP3315413B2 (ja) データ・ハイディング方法及びデータ抽出方法
CN100584020C (zh) 数字水印的嵌入、提取认证方法及装置
Sherly et al. A compressed video steganography using TPVD
Su et al. A practical design of digital video watermarking in H. 264/AVC for content authentication
Du et al. Lossless authentication of MPEG-2 video
JPH11284843A (ja) ウォ―タマ―クをディジタル画像シ―ケンスのビットストリ―ム表現に組み込むための方法及び装置
Liu et al. A robust data hiding algorithm for H. 264/AVC video streams
Liao et al. Double H. 264/AVC compression detection using quantized nonzero AC coefficients
EP1529262A1 (fr) Filigranage numerique et dactyloscopie de contenu numerique a l&#39;aide de blocs alternatifs pour incorporer des informations
Yin et al. A semi-fragile watermarking system for MPEG video authentication
CN105100959A (zh) 一种取证标记方法、装置及数字家庭影院
JP2005341525A (ja) ビデオウォーターマーキング方法及び装置と、ビデオウォーターマーキングを用いたビデオコンテンツの保護方法及び装置
JP2004241869A (ja) 透かし埋め込み及び画像圧縮部
US7418110B2 (en) Method and apparatus for compressed-domain watermarking
WO2009036684A1 (fr) Procédé et dispositif pour l&#39;intégration et la détection d&#39;informations en filigrane
EP1221816A2 (fr) Système de transmission vidéo, unité de transmission vidéo et méthode pour l&#39;insertion d&#39;un filigrane
Saadi et al. H. 264/AVC video authentication based video content
Lo et al. Generic attacks on content-based video stream authentication
Pei et al. A video watermarking scheme based on motion vectors and mode selection
WO2018177522A1 (fr) Filigrane vidéo
JP3566924B2 (ja) 電子透かし埋め込み方法，検出方法，電子透かし埋め込み装置，検出装置および電子透かし埋め込みプログラムを記録した記録媒体，検出プログラムを記録した記録媒体
KR100768669B1 (ko) 움직임 벡터의 최소변경 기법을 이용한 블라인드 비디오워터마킹의 방법

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 08800820

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 08800820

Country of ref document: EP

Kind code of ref document: A1