Video recompression forensics method for H.266 / VVC standard

Abstract

A video recompression forensics method for H.266 / VVC standard, characterized by comprising the steps of extracting primary stream features of video stream, determining the proportion of video stream coding unit classification types, determining the attribute matrix of video stream coding unit, re-compressing the video stream, extracting primary stream features of the re-compressed video stream, determining the attribute matrix of the re-compressed video stream coding unit, determining the attribute mark of the video stream coding unit, determining the proportion of different attribute coding units of the video stream, determining the advanced stream features of the video stream, and determining the re-compression forensics result. The method solves the problem that the existing technology does not effectively consider the impact of the new coding technology in the H.266 / VVC standard and the high complexity of forensics. The present invention has the advantages of high accuracy and low complexity of video re-compression forensics, and can be used in the technical field of forensics of H.266 / VVC standard video streams.

Description

Technical Field

[0001] The present invention belongs to the technical field of video forensics, and in particular relates to video recompression forensics oriented to the H.266 / VVC standard. Background Art

[0002] Video security is an important manifestation of information security and network security. Video forensics methods that aim to detect video content tampering information are an important guarantee for video security. In most current video applications such as video surveillance, video conferencing, radio and television, and streaming media, video signals must go through operations such as acquisition, encoding and compression, transmission, storage, decoding, and display. Tamperers usually tamper with the video content during the transmission and storage of the original video, and the tampered video content must be re-encoded and compressed before it can be effectively transmitted and stored. Therefore, the number of encoding and compression times becomes a necessary condition for determining whether the original video has been tampered with. The method that aims to detect the number of video encoding compression times is called video re-compression forensics method, which has important applications in criminal investigation, forensic science and other fields.

[0003] At present, some video recompression forensics methods have been proposed, but they are mainly aimed at early video coding standards such as MPEG-2, MPEG-4, H.264 / AVC, and H.265 / HEVC. In recent years, in addition to continuing to develop towards high-definition / ultra-high-definition, video is also developing towards high frame rate, high color component sampling depth, high dynamic range, 360-degree panorama, etc. The coding efficiency of early video coding standards has been difficult to meet actual needs. Accordingly, the H.266 / VVC standard came into being. H.266 / VVC is the latest generation of international video coding standards released in 2020. Its compression efficiency is about twice that of the H.265 / HEVC standard and four times that of the H.264 / AVC standard. It also meets the coding requirements of ultra-high-definition, screen content, high dynamic / wide color gamut, 360-degree panorama and other video content. With its excellent compression efficiency and remarkable versatility, the H.266 / VVC standard has broad market application prospects. The H.266 / VVC standard uses a large number of advanced coding technologies, which not only significantly improves the efficiency of video coding, but also significantly changes the syntax, semantics and compression traces of the video bitstream obtained by coding compression. The currently proposed video re-compression forensics methods cannot be effectively applied to the H.266 / VVC standard because they do not effectively consider the impact of new coding technologies in the H.266 / VVC standard, and the accuracy of re-compression forensics is low. In addition, some of the currently proposed video re-compression forensics methods require the video bitstream to be encoded and compressed more than once, which makes the forensics complex. Summary of the invention

[0004] The technical problem to be solved by the present invention is to overcome the shortcomings of the above-mentioned prior art and provide a video recompression forensics method oriented to the H.266 / VVC standard which effectively considers the quantization parameters and the division type, prediction mode, and reference row influence of the intra-frame prediction frame brightness domain coding unit in H.266 / VVC, and only needs to encode and compress the video bit stream to be recompressed forensics once, has high accuracy and low complexity of video recompression forensics.

[0005] The technical solution adopted to solve the above technical problems consists of the following steps:

[0006] (1) Extracting primary stream features of video stream

[0007] The primary bitstream features are extracted from the collected video bitstream using decoding analysis software. The decoding analysis software has been made public at https: / / vcgit.hhi.fraunhofer.de / jvet / VVCSoftware_VTM. The primary bitstream features include quantization parameters and the division type, prediction mode, and reference line of the intra-frame prediction frame brightness domain coding unit.

[0008] (2) Determine the proportion of video bitstream coding unit division types

[0009] According to formula (1), the proportion P of the zth intra-frame prediction frame brightness domain coding unit division type in the video bit stream is determined z :

[0010]

[0011] Where I represents the total number of intra-frame prediction frames contained in the video bitstream, I is a finite positive integer, and N i,z represents the number of the zth luminance domain coding unit partition types of the i-th intra-frame prediction frame in the video bitstream, i∈{1,2,…,I}, z∈{1,2,…,α}, α represents the total number of optional partition types of the luminance domain coding unit of the intra-frame prediction frame supported by the H.266 / VVC standard, α is a finite positive integer, D i Indicates the number of coding units contained in the i-th intra-frame prediction frame in the video bitstream, D i is a finite positive integer.

[0012] (3) Determine the video code stream encoding unit attribute matrix

[0013] According to formula (2), the attribute matrix A of the xth luminance domain coding unit of the i-th intra-frame prediction frame in the video bitstream is determined: i,x :

[0014] A i,x =[t i,x ,m i,x ,r i,x ](2)

[0015] Among them, t i,x 、m i,x 、r i,x They represent the partition type, prediction mode, and reference row number of the xth luminance domain coding unit of the i-th intra-frame prediction frame in the video bitstream, respectively. x∈{1,2,…,w}, w represents the total number of luminance domain coding units contained in the intra-frame prediction frame. w is a finite positive integer. i,x ∈{1,2,…,α},m i,x ∈{1,2,…,β}, β represents the total number of optional prediction modes of the intra-frame prediction frame brightness domain coding unit supported by the H.266 / VVC standard, and β is a finite positive integer. Indicates the total number of optional reference lines for intra-frame prediction luminance domain coding units supported by the H.266 / VVC standard. is a finite positive integer.

[0016] (4) Recompress the video stream

[0017] The video bitstream is decoded into a decoded video using a decoder supporting the H.266 / VVC standard, and the decoded video is re-encoded and compressed using an encoder supporting the H.266 / VVC standard to obtain a re-compressed video bitstream.

[0018] (5) Extracting primary stream features of the recompressed video stream

[0019] Extract primary code stream features from the re-compressed video code stream, where the primary code stream features include the division type, prediction mode, and reference row of the intra-frame prediction frame brightness domain coding unit.

[0020] (6) Determine the attribute matrix of the encoding unit of the video code stream after recompression

[0021] According to formula (3), the attribute matrix B of the xth luminance domain coding unit of the i-th intra-frame prediction frame in the re-compressed video code stream is determined: i,x :

[0022] B i,x =[u i,x ,v i,x ,q i,x ](3)

[0023] Among them, u i,x 、v i,x ,q i,x They represent the division type, prediction mode, and reference row number of the xth luminance domain coding unit of the i-th intra-frame prediction frame in the recompressed video bitstream, respectively. i,x ∈{1,2,…,α},v i,x ∈{1,2,…,β},

[0024] (7) Determine the video code stream encoding unit attribute mark

[0025] Determine the attribute label θ of the xth luminance domain coding unit of the i-th intra-frame prediction frame in the video bitstream according to formula (4): i,x :

[0026]

[0027] Among them A i,x ≠B i,x The coding units are different attribute coding units.

[0028] (8) Determine the proportion of coding units with different attributes in the video bitstream

[0029] According to formula (5), the proportion of different attribute coding units of the zth type of brightness domain coding unit division type in the video bit stream is determined. z :

[0030]

[0031] (9) Determine the high-level stream characteristics of the video stream

[0032] According to formula (6), the high-level bitstream feature matrix F of the video bitstream is determined as follows:

[0033] F=[Q,S1,S2,…,S α ,P1,P2,…,P α ](6)

[0034] Wherein, Q represents a quantization parameter extracted from a video bitstream, and Q is a finite positive integer.

[0035] (10) Determine the results of the recompression forensics

[0036] The high-level bitstream feature matrix F is input into a software package for support vector machine pattern recognition and regression. According to the binary classification data output by the software package for support vector machine pattern recognition and regression, the heavy compression forensics result corresponding to the video bitstream is determined. The software package for support vector machine pattern recognition and regression has been made public at https: / / www.csie.ntu.edu.tw / ~cjlin / libsvm / .

[0037] Completed the video re-compression forensics method for H.266 / VVC standard.

[0038] In the formula (1) of the step (2) of determining the proportion of the video bitstream coding unit division types of the present invention, the I represents the total number of intra-frame prediction frames contained in the video bitstream, and the minimum value of I is 1. α represents the total number of optional division types of the intra-frame prediction frame luminance domain coding unit supported by the H.266 / VVC standard, and α∈[1,17].

[0039] In the formula (2) of the step (3) of determining the attribute matrix of the video code stream coding unit of the present invention, the β represents the total number of optional prediction modes of the intra-frame prediction frame brightness domain coding unit supported by the H.266 / VVC standard, β∈[1,67], Indicates the total number of optional reference lines for intra-frame prediction luminance domain coding units supported by the H.266 / VVC standard.

[0040] In the formula (6) of the step (9) of determining the high-level bitstream characteristics of the video bitstream of the present invention, the Q represents the quantization parameter extracted from the video bitstream, Q∈[0,63].

[0041] Since the present invention adopts the steps of determining the proportion of video code stream coding unit division types, re-compressing the video code stream, determining the proportion of different attribute coding units of the video code stream, and determining the advanced code stream characteristics of the video code stream, the problem that the existing technology does not effectively consider the impact of new coding technology in the H.266 / VVC standard and the high complexity of evidence collection is solved. The present invention has the advantages of high accuracy and low complexity of video re-compression evidence collection, and can be used in the field of evidence collection technology of H.266 / VVC standard video code streams. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 This is a flow chart of Example 1 of the present invention. DETAILED DESCRIPTION

[0043] The present invention will be further described in detail below in conjunction with the accompanying drawings and examples, but the present invention is not limited to the following embodiments.

[0044] Example 1

[0045] The video recompression evidence collection method for the H.266 / VVC standard of this embodiment consists of the following steps (see Figure 1 ):

[0046] (1) Extracting primary stream features of video stream

[0047] The primary stream features are extracted from the collected video stream using decoding analysis software. The primary stream features include quantization parameters and the division type, prediction mode, and reference line of the intra-frame prediction frame brightness domain coding unit.

[0048] (2) Determine the proportion of video bitstream coding unit division types

[0049] According to formula (1), the proportion P of the zth intra-frame prediction frame brightness domain coding unit division type in the video bit stream is determined z :

[0050]

[0051] Where I represents the total number of intra-frame prediction frames contained in the video bitstream, and I is a finite positive integer. In this embodiment, the value of I is 5, and N i,z represents the number of the zth luminance domain coding unit partition types of the i-th intra-frame prediction frame in the video bitstream, i∈{1,2,…,I}, z∈{1,2,…,α}, α represents the total number of optional partition types of the luminance domain coding unit of the intra-frame prediction frame supported by the H.266 / VVC standard, α∈[1,17], and the value of α in this embodiment is 9, D i Indicates the number of coding units contained in the i-th intra-frame prediction frame in the video bitstream, D i is a finite positive integer.

[0052] (3) Determine the video code stream encoding unit attribute matrix

[0053] According to formula (2), the attribute matrix A of the xth luminance domain coding unit of the i-th intra-frame prediction frame in the video bitstream is determined: i,x :

[0054] A i,x =[t i,x ,m i,x ,r i,x ](2)

[0055] Among them, t i,x 、m i,x 、r i,x They represent the partition type, prediction mode, and reference row number of the xth luminance domain coding unit of the i-th intra-frame prediction frame in the video bitstream, respectively. x∈{1,2,…,w}, w represents the total number of luminance domain coding units contained in the intra-frame prediction frame. w is a finite positive integer. i,x ∈{1,2,…,α},m i,x ∈{1,2,…,β}, β represents the total number of optional prediction modes of the intra-frame prediction frame brightness domain coding unit supported by the H.266 / VVC standard, β∈[1,67], and the value of β in this embodiment is 34, Indicates the total number of optional reference lines for intra-frame prediction luminance domain coding units supported by the H.266 / VVC standard. This embodiment of The value is 2.

[0056] (4) Recompress the video stream

[0057] The video bitstream is decoded into a decoded video using a decoder supporting the H.266 / VVC standard, and the decoded video is re-encoded and compressed using an encoder supporting the H.266 / VVC standard to obtain a re-compressed video bitstream.

[0058] (5) Extracting primary stream features of the recompressed video stream

[0059] Extract primary code stream features from the re-compressed video code stream, where the primary code stream features include the division type, prediction mode, and reference row of the intra-frame prediction frame brightness domain coding unit.

[0060] (6) Determine the attribute matrix of the encoding unit of the video code stream after recompression

[0061] According to formula (3), the attribute matrix B of the xth luminance domain coding unit of the i-th intra-frame prediction frame in the re-compressed video code stream is determined: i,x :

[0062] B i,x =[u i,x ,v i,x ,q i,x ](3)

[0063] Among them, u i,x 、v i,x ,q i,x They represent the division type, prediction mode, and reference row number of the xth luminance domain coding unit of the i-th intra-frame prediction frame in the recompressed video bitstream, respectively. i,x ∈{1,2,…,α},v i,x ∈{1,2,…,β},

[0064] (7) Determine the video code stream encoding unit attribute mark

[0065] Determine the attribute label θ of the xth luminance domain coding unit of the i-th intra-frame prediction frame in the video bitstream according to formula (4): i,x :

[0066]

[0067] Among them A i,x ≠B i,x The coding units are different attribute coding units.

[0068] (8) Determine the proportion of coding units with different attributes in the video bitstream

[0069] According to formula (5), the proportion of different attribute coding units of the zth type of brightness domain coding unit division type in the video bit stream is determined. z :

[0070]

[0071] (9) Determine the high-level stream characteristics of the video stream

[0072] According to formula (6), the high-level bitstream feature matrix F of the video bitstream is determined as follows:

[0073] F=[Q,S1,S2,…,S α ,P1,P2,…,P α ](6)

[0074] Wherein, Q represents a quantization parameter extracted from a video bitstream, Q∈[0,63]. In this embodiment, the value of Q is 32, and Q can also be arbitrarily selected between 0 and 63.

[0075] Since the present invention adopts the steps of determining the proportion of video code stream coding unit division types, re-compressing the video code stream, determining the proportion of different attribute coding units of the video code stream, and determining the advanced code stream characteristics of the video code stream, the problem that the existing technology does not effectively consider the impact of the new coding technology in the H.266 / VVC standard and the high complexity of evidence collection is solved. The present invention has the advantages of high accuracy and low complexity of video re-compression evidence collection.

[0076] (10) Determine the results of the recompression forensics

[0077] The advanced bitstream feature matrix F is input into a software package for support vector machine pattern recognition and regression, and the heavy compression forensics result corresponding to the video bitstream is determined according to the 2-classification data output by the software package for support vector machine pattern recognition and regression.

[0078] Completed the video re-compression forensics method for H.266 / VVC standard.

[0079] Example 2

[0080] The video recompression evidence collection method for the H.266 / VVC standard of this embodiment consists of the following steps:

[0081] (1) Extracting primary stream features of video stream

[0082] This step is the same as in Example 1.

[0083] (2) Determine the proportion of video bitstream coding unit division types

[0084] According to formula (1), the proportion P of the zth intra-frame prediction frame brightness domain coding unit division type in the video bit stream is determined z :

[0085]

[0086] Where I represents the total number of intra-frame prediction frames contained in the video bitstream, and the minimum value of I is 1. In this embodiment, the value of I is 4. i,z represents the number of the zth luminance domain coding unit partition types of the i-th intra-frame prediction frame in the video bitstream, i∈{1,2,…,I}, z∈{1,2,…,α}, α represents the total number of optional partition types of the luminance domain coding unit of the intra-frame prediction frame supported by the H.266 / VVC standard, α∈[1,17], and the value of α in this embodiment is 1, D i Indicates the number of coding units contained in the i-th intra-frame prediction frame in the video bitstream, D i is a finite positive integer.

[0087] (3) Determine the video code stream encoding unit attribute matrix

[0088] According to formula (2), the attribute matrix A of the xth luminance domain coding unit of the i-th intra-frame prediction frame in the video bitstream is determined: i,x :

[0089] A i,x =[t i,x ,m i,x ,r i,x ](2)

[0090] Among them, t i,x 、m i,x 、r i,x They represent the partition type, prediction mode, and reference row number of the xth luminance domain coding unit of the i-th intra-frame prediction frame in the video bitstream, respectively. x∈{1,2,…,w}, w represents the total number of luminance domain coding units contained in the intra-frame prediction frame. w is a finite positive integer. i,x ∈{1,2,…,α},m i,x ∈{1,2,…,β}, β represents the total number of optional prediction modes of the intra-frame prediction frame brightness domain coding unit supported by the H.266 / VVC standard, β∈[1,67], and the value of β in this embodiment is 1, Indicates the total number of optional reference lines for intra-frame prediction luminance domain coding units supported by the H.266 / VVC standard. This embodiment of The value is 1.

[0091] (4) Recompress the video stream

[0092] This step is the same as in Example 1.

[0093] (5) Extracting primary stream features of the recompressed video stream

[0094] This step is the same as in Example 1.

[0095] (6) Determine the attribute matrix of the encoding unit of the video code stream after recompression

[0096] This step is the same as in Example 1.

[0097] (7) Determine the video code stream encoding unit attribute mark

[0098] This step is the same as in Example 1.

[0099] (8) Determine the proportion of coding units with different attributes in the video bitstream

[0100] This step is the same as in Example 1.

[0101] (9) Determine the high-level stream characteristics of the video stream

[0102] According to formula (6), the high-level bitstream feature matrix F of the video bitstream is determined as follows:

[0103] F=[Q,S1,S2,…,S α ,P1,P2,…,P α ](6)

[0104] Wherein, Q represents a quantization parameter extracted from a video bitstream, Q∈[0,63]. In this embodiment, the value of Q is 0, and Q can also be arbitrarily selected between Q∈[0,63].

[0105] The other steps are the same as those in Example 1. The video recompression evidence collection method for the H.266 / VVC standard is completed.

[0106] Example 3

[0107] The video recompression evidence collection method for the H.266 / VVC standard of this embodiment consists of the following steps:

[0108] (1) Extracting primary stream features of video stream

[0109] This step is the same as in Example 1.

[0110] (2) Determine the proportion of video bitstream coding unit division types

[0111] According to formula (1), the proportion P of the zth intra-frame prediction frame brightness domain coding unit division type in the video bit stream is determined z :

[0112]

[0113] Where I represents the total number of intra-frame prediction frames contained in the video bitstream, and the minimum value of I is 1. In this embodiment, the value of I is 7. i,zrepresents the number of the zth luminance domain coding unit partition types of the i-th intra-frame prediction frame in the video bitstream, i∈{1,2,…,I}, z∈{1,2,…,α}, α represents the total number of optional partition types of the luminance domain coding unit of the intra-frame prediction frame supported by the H.266 / VVC standard, α∈[1,17], and the value of α in this embodiment is 17, D i Indicates the number of coding units contained in the i-th intra-frame prediction frame in the video bitstream, D i is a finite positive integer.

[0114] (3) Determine the video code stream encoding unit attribute matrix

[0115] According to formula (2), the attribute matrix A of the xth luminance domain coding unit of the i-th intra-frame prediction frame in the video bitstream is determined: i,x :

[0116] A i,x =[t i,x ,m i,x ,r i,x ](2)

[0117] Among them, t i,x 、m i,x 、r i,x They represent the partition type, prediction mode, and reference row number of the xth luminance domain coding unit of the i-th intra-frame prediction frame in the video bitstream, respectively. x∈{1,2,…,w}, w represents the total number of luminance domain coding units contained in the intra-frame prediction frame. w is a finite positive integer. i,x ∈{1,2,…,α},m i,x ∈{1,2,…,β}, β represents the total number of optional prediction modes of the intra-frame prediction frame brightness domain coding unit supported by the H.266 / VVC standard, β∈[1,67], and the value of β in this embodiment is 67. Indicates the total number of optional reference lines for intra-frame prediction luminance domain coding units supported by the H.266 / VVC standard. This embodiment of The value is 3.

[0118] (4) Recompress the video stream

[0119] This step is the same as in Example 1.

[0120] (5) Extracting primary stream features of the recompressed video stream

[0121] This step is the same as in Example 1.

[0122] (6) Determine the attribute matrix of the encoding unit of the video code stream after recompression

[0123] This step is the same as in Example 1.

[0124] (7) Determine the video code stream encoding unit attribute mark

[0125] This step is the same as in Example 1.

[0126] (8) Determine the proportion of coding units with different attributes in the video bitstream

[0127] This step is the same as in Example 1.

[0128] (9) Determine the high-level stream characteristics of the video stream

[0129] According to formula (6), the high-level bitstream feature matrix F of the video bitstream is determined as follows:

[0130] F=[Q,S1,S2,…,S α ,P1,P2,…,P α ] (6)

[0131] Wherein, Q represents a quantization parameter extracted from a video bitstream, Q∈[0,63]. In this embodiment, the value of Q is 63, and Q may also be arbitrarily selected between Q∈[0,63].

[0132] The other steps are the same as those in Example 1. The video recompression evidence collection method for the H.266 / VVC standard is completed.

[0133] In order to verify the beneficial effects of the present invention, the inventors conducted experiments on the test video using the method of Embodiment 3 of the present invention. The experimental results are as follows:

[0134] 29 standard videos recommended by the coding standards formulation organization are selected as test videos, and the color space and resolution of the videos are YCbCr4:2:0 and 720P respectively. The reference software VTM-11.0 recommended by the H.266 / VVC standard is used to encode and compress the test video once and twice to obtain the corresponding video stream, and the video stream is input into the method of Example 3 of the present invention for recompression and evidence collection. The encoding level is main_10, the encoding structure is a low-delay encoding structure, and the quantization parameters are set to 22, 27, 32, 37, 42, and 47 respectively. For other encoding parameters, refer to the default settings in the configuration file encoder_lowdelay_P_vtm.cfg. The video recompression and evidence collection accuracy g corresponding to the method of Example 3 of the present invention is calculated according to formula (7):

[0135] g=(c+f) / 2 (7)

[0136] Wherein, c represents the accuracy of the present invention in identifying single compressed videos, and f represents the accuracy of the present invention in identifying heavily compressed videos. The results are shown in Table 1.

[0137] Table 1 Video recompression forensics accuracy g results of the method in Example 3 of the present invention

[0138]

[0139] It can be seen from Table 1 that under each test quantization parameter, the video recompression forensics accuracy of the method in Example 3 of the present invention is greater than or equal to 91.5%. For all test quantization parameters, the video recompression forensics accuracy of the method in Example 3 of the present invention is 96.4% on average, indicating that the video recompression forensics accuracy of the method in Example 3 of the present invention is high.

[0140] In addition, in the steps of the technical solution of the present invention, the video bitstream to be re-compressed for evidence only needs to be encoded and compressed once in the step of obtaining the video bitstream after re-compression, which shows that the complexity of video re-compression forensics of the method of embodiment 3 of the present invention is low.

Claims

1. A video recompression forensics method for H.266 / VVC standard, characterized in that It consists of the following steps: (1) Extracting primary stream features of video stream The primary bitstream features are extracted from the collected video bitstream using decoding analysis software. The primary bitstream features include quantization parameters and the division type, prediction mode, and reference line of the intra-frame prediction frame brightness domain coding unit; (2) Determine the proportion of video bitstream coding unit division types According to formula (1), the proportion P of the zth intra-frame prediction frame brightness domain coding unit division type in the video bit stream is determined z : Where I represents the total number of intra-frame prediction frames contained in the video bitstream, I is a finite positive integer, and N i,z represents the number of the zth luminance domain coding unit partition types of the i-th intra-frame prediction frame in the video bitstream, i∈{1,2,…,I}, z∈{1,2,…,α}, α represents the total number of optional partition types of the luminance domain coding unit of the intra-frame prediction frame supported by the H.266 / VVC standard, α is a finite positive integer, D i Indicates the number of coding units contained in the i-th intra-frame prediction frame in the video bitstream, D i is a finite positive integer; (3) Determine the video code stream encoding unit attribute matrix According to formula (2), the attribute matrix A of the xth luminance domain coding unit of the i-th intra-frame prediction frame in the video bitstream is determined: i,x : A i,x =[t i,x ,m i,x ,r i,x ] (2) Among them, t i,x 、m i,x 、r i,x They represent the partition type, prediction mode, and reference row number of the xth luminance domain coding unit of the i-th intra-frame prediction frame in the video bitstream, respectively. x∈{1,2,…,w}, w represents the total number of luminance domain coding units contained in the intra-frame prediction frame. w is a finite positive integer. i,x ∈{1,2,…,α},m i,x ∈{1,2,…,β}, β represents the total number of optional prediction modes of the intra-frame prediction frame brightness domain coding unit supported by the H.266 / VVC standard, and β is a finite positive integer. Indicates the total number of optional reference lines for intra-frame prediction luminance domain coding units supported by the H.266 / VVC standard. is a finite positive integer; (4) Recompress the video stream Decode the video bitstream into a decoded video using a decoder supporting the H.266 / VVC standard, and re-encode and compress the decoded video using an encoder supporting the H.266 / VVC standard to obtain a re-compressed video bitstream; (5) Extracting primary stream features of the recompressed video stream Extracting primary code stream features from the re-compressed video code stream, the primary code stream features including the division type, prediction mode, and reference row of the intra-frame prediction frame brightness domain coding unit; (6) Determine the attribute matrix of the encoding unit of the video code stream after recompression According to formula (3), the attribute matrix B of the xth luminance domain coding unit of the i-th intra-frame prediction frame in the re-compressed video code stream is determined: i,x : B i,x =[u i,x ,v i,x ,q i,x ] (3) Among them, u i,x 、v i,x ,q i,x They represent the division type, prediction mode, and reference row number of the xth luminance domain coding unit of the i-th intra-frame prediction frame in the recompressed video bitstream, respectively. i,x ∈{1,2,…,α},v i,x ∈{1,2,…,β}, (7) Determine the attribute mark of the video code stream encoding unit Determine the attribute label θ of the xth luminance domain coding unit of the i-th intra-frame prediction frame in the video bitstream according to formula (4): i,x : Among them A i,x ≠B i,x The coding units are different attribute coding units; (8) Determine the proportion of coding units with different attributes in the video bitstream According to formula (5), the proportion of different attribute coding units of the zth type of brightness domain coding unit division type in the video bit stream is determined. z : (9) Determine the high-level stream characteristics of the video stream According to formula (6), the high-level bitstream feature matrix F of the video bitstream is determined as follows: F=[Q,S1,S2,…,S α ,P1,P2,…,P α ] (6) Wherein, Q represents the quantization parameter extracted from the video bitstream, and Q is a finite positive integer; (10) Determine the results of the recompression forensics The advanced bitstream feature matrix F is input into a software package for support vector machine pattern recognition and regression, and the heavy compression forensics result corresponding to the video bitstream is determined according to the 2-classification data output by the software package for support vector machine pattern recognition and regression.

2. The video recompression evidence collection method for the H.266 / VVC standard according to claim 1, characterized in that: In formula (1) of the step (2) of determining the proportion of the video bitstream coding unit division types, I represents the total number of intra-frame prediction frames contained in the video bitstream, and the minimum value of I is 1. α represents the total number of optional division types of the intra-frame prediction frame luminance domain coding unit supported by the H.266 / VVC standard, α∈[1,17].

3. The video recompression evidence collection method for the H.266 / VVC standard according to claim 1, characterized in that: In the formula (2) of the step (3) of determining the attribute matrix of the video bitstream coding unit, the β represents the total number of optional prediction modes of the intra-frame prediction frame brightness domain coding unit supported by the H.266 / VVC standard, β∈[1,67], Indicates the total number of optional reference lines for intra-frame prediction luminance domain coding units supported by the H.266 / VVC standard.

4. The video recompression evidence collection method for the H.266 / VVC standard according to claim 1, characterized in that: In equation (6) of the step (9) of determining the high-level bitstream features of the video bitstream, Q represents a quantization parameter extracted from the video bitstream, Q∈[0,63].

Images