CN109361926B - Lossless reversible information hiding method for H.264/AVC video visual quality - Google Patents
Lossless reversible information hiding method for H.264/AVC video visual quality Download PDFInfo
- Publication number
- CN109361926B CN109361926B CN201811423158.4A CN201811423158A CN109361926B CN 109361926 B CN109361926 B CN 109361926B CN 201811423158 A CN201811423158 A CN 201811423158A CN 109361926 B CN109361926 B CN 109361926B
- Authority
- CN
- China
- Prior art keywords
- cbp
- coefficient
- secret information
- bit
- value
- Prior art date
- Legal status (The legal status 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 status listed.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 46
- 230000002441 reversible effect Effects 0.000 title claims abstract description 33
- 230000000007 visual effect Effects 0.000 title claims abstract description 25
- 230000008569 process Effects 0.000 claims abstract description 16
- 238000012986 modification Methods 0.000 claims abstract description 14
- 230000004048 modification Effects 0.000 claims abstract description 14
- 238000000605 extraction Methods 0.000 claims description 14
- 238000012360 testing method Methods 0.000 claims description 12
- YEPGIYSLLAGBSS-UHFFFAOYSA-M chloro-[3-[(4-iodophenyl)carbamoylamino]-2-methoxypropyl]mercury Chemical compound Cl[Hg]CC(OC)CNC(=O)NC1=CC=C(I)C=C1 YEPGIYSLLAGBSS-UHFFFAOYSA-M 0.000 claims description 9
- 230000009191 jumping Effects 0.000 claims description 6
- 238000013139 quantization Methods 0.000 claims description 5
- 238000012545 processing Methods 0.000 claims description 4
- 230000005540 biological transmission Effects 0.000 claims description 3
- 239000013598 vector Substances 0.000 claims description 3
- 230000006835 compression Effects 0.000 abstract description 2
- 238000007906 compression Methods 0.000 abstract description 2
- 238000013461 design Methods 0.000 abstract description 2
- 238000005516 engineering process Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 6
- 238000013519 translation Methods 0.000 description 5
- 230000007704 transition Effects 0.000 description 4
- 238000011160 research Methods 0.000 description 3
- 230000009466 transformation Effects 0.000 description 3
- 239000000969 carrier Substances 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 239000002360 explosive Substances 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000002427 irreversible effect Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Images
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
- H04N19/46—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
- 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
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/12—Selection from among a plurality of transforms or standards, e.g. selection between discrete cosine transform [DCT] and sub-band transform or selection between H.263 and H.264
- H04N19/122—Selection of transform size, e.g. 8x8 or 2x4x8 DCT; Selection of sub-band transforms of varying structure or type
-
- 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
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/13—Adaptive entropy coding, e.g. adaptive variable length coding [AVLC] or context adaptive binary arithmetic coding [CABAC]
-
- 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
- H04N19/17—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 an image region, e.g. an object
- H04N19/176—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 an image region, e.g. an object the region being a block, e.g. a macroblock
-
- 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
- H04N19/186—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 colour or a chrominance component
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- Discrete Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The invention discloses a lossless reversible information hiding method for H.264/AVC video visual quality, which comprises the steps of carrying out entropy decoding on an H.264 video compression code stream when secret information is embedded, obtaining a CBP coefficient and a macro block prediction mode, selecting I4, P16 x 16, P16 x 8, P8 x 16 and P8 x 8 macro blocks, embedding a brightness CBP coefficient, embedding a chroma CBP coefficient and continuously carrying out entropy coding to obtain a compressed video. The invention designs an embedding and extracting scheme with lossless video visual quality by combining the characteristics of the CBP coefficient and the H.264/AVC video coding principle. The macroblock CBP coefficient is modulated, the video visual quality is completely fidelity after secret information is embedded, and invalid modification does not exist in the embedding process, so that the embedding efficiency of the method is improved, and the video quality and the method safety are guaranteed.
Description
Technical Field
The invention relates to the technical field of steganography of a video compression domain, in particular to a method for hiding lossless and reversible information of H.264/AVC video visual quality, which is a method for hiding lossless and reversible information of H.264/AVC video visual quality based on a macroblock CBP coefficient value.
Background
In the 21 st century, human society has entered the information age, and the explosive development of internet and communication technology has led to an explosive increase in the amount of information. The concept of 'internet +' provides diversified choices for people and brings more and more potential threats. The network is an open platform, and the network with higher security level has the risk of being cracked. The importance of information is increasingly reflected from the problems of civilian life to national security. In the information age with both opportunities and challenges, how to ensure information security becomes a worldwide research topic. Multimedia information steganography is an important branch of the information security field. The digital video is an ideal steganographic carrier due to the characteristics that the data volume is large and the number of the secret messages can be accommodated. H.264/AVC is the most widely used video coding standard at present and is widely applied to various products of the Internet, so that the research on the steganography method taking the H.264/AVC as a carrier has important theoretical significance and practical application value.
Conventional information hiding methods modify some characteristic values of the carrier more or less during the information embedding process, and such changes are usually irreversible, thereby causing permanent distortion of the carrier. In some application scenarios, however, the original data of the carrier needs to be completely preserved, and any distortion causes immeasurable effects, such as embedding patient information in medical images, hiding access rights in confidential documents, and the like. For such information hiding scenarios that cannot accept carrier distortion, a reversible information hiding technique works.
Reversible information hiding is used for hiding information into a digital medium in a reversible mode, and carrier distortion caused by an information embedding process is reversible through designing a reasonable modulation strategy. The difference expansion and histogram shift algorithm is two main reversible information hiding technologies, and can be applied to carriers such as images and videos.
The reversible information hiding technology develops more mature in the field of image information hiding, wherein the achievement based on the spatial domain is most abundant. Jun Tian in 2003 8 months published article "Reversible data embedding using a differential encoding algorithm" in IEEE Transactions on Circuits and Systems for video Technology, and the first time proposed a differential extension Reversible steganography algorithm mode, and differential embedding confidence of adjacent pixel pairs is extended by using integer Haar wavelet transform. Ni et al published article "Reversible data hiding" in the journal IEEE Transactions on Circuits and Systems for Video Technology in 2006 proposed a Reversible information hiding algorithm based on the histogram translation theory, which has lower computational complexity and better hiding performance compared to other Reversible steganography algorithms.
Reversible information hiding of images many research theories can be generalized to reversible information hiding of videos. The selection of QDCT coefficients to embed secret information is the most common practice in video-supported reversible information hiding algorithms. Lin et al published in the journal of manufacturing academic of Science Engineering & Technology in 2012, in Reversable watermarking for H.264/AVC video, and proposed a Reversible watermarking algorithm based on histogram translation theory, which collects the last non-zero coefficient of all 4 × 4 blocks of QDCT coefficients, predicts the latter coefficient with the former coefficient to obtain a prediction residual, and modifies and embeds watermark information in the histogram of the prediction residual. Liu et al published an article "Reversible Data locating Scheme Based On H.264/AVC with out discovery drive" On Journal of Software in 2012, and proposed a Reversible information Hiding algorithm for eliminating Drift errors aiming at modifying Drift errors introduced by I frames. In 2016, Zhao et al published an article "Anovel two-dimensional histogram modification for reversible data embedding stereo H.264 video" in Multimedia Tools & Applications, and improved the one-dimensional histogram translation technology, and proposed a two-dimensional histogram translation strategy to embed secret letter in the H.264/AVC video QDCT coefficient, and most 3-bit secret letter can be embedded when the coefficient modification amplitude is 1. The algorithm greatly improves the embedding efficiency, improves the embedding capacity and simultaneously can obtain better visual quality.
The current reversible information hiding algorithm based on video generally has the following defects:
1. the modification in the embedding process causes the carrier video pixel value to change, and introduces distortion to cause the video visual quality to be reduced. Although such distortions can be fully recovered in the extraction process by the definition of reversible information hiding, distortions are still present in the carrier before the recipient extracts the secret. Therefore, the security problem caused by the video visual quality reduction still needs to be considered when the secret carrier file is transmitted in a channel.
2. When a modulation strategy is designed, in order to ensure reversible ubiquitous invalid modification, a reversible information hiding algorithm modifies a carrier but does not carry any secret information. For example, in the histogram translation algorithm, all points between a peak point and a zero point need to be translated without embedding secret information; in the difference expansion algorithm, a position map needs to be additionally embedded in a carrier to realize reversibility, and the method also belongs to invalid modification. Invalid modifications reduce algorithm efficiency.
Disclosure of Invention
The invention provides an H.264/AVC video visual quality lossless reversible information hiding method based on a macro block CBP coefficient value aiming at the defects of a reversible steganography technology in the field of video steganography; the method does not introduce any distortion while embedding the secret information, and can obtain the video with completely lossless video visual quality after embedding, thereby greatly improving the safety of the method.
The purpose of the invention can be realized by the following technical scheme:
the lossless and reversible information hiding method for the visual quality of the H.264/AVC video comprises a secret letter embedding part and a secret letter extracting part, and is characterized in that the secret letter embedding part comprises the following steps:
1) entropy decoding the H.264 video compressed code stream, reading a first macro block of a current decoding frame, and obtaining a macro block prediction mode of the current decoding frame;
2) determining whether the current macroblock prediction mode is I4, P16 × 16, P16 × 8, P8 × 16, or P8 × 8: if yes, reading the CBP coefficient value of the current macro block, and executing the step 3); otherwise, reading the next macro block and re-executing the step 2);
3) sequentially reading the lower 4 bits of the CBP coefficient, if the current bit is equal to 1, skipping the bit and not embedding secret information; if the current bit is equal to 0, 1-bit secret information is embedded; the embedding rule is that the current bit is modified to be consistent with the value of the secret information to be embedded;
4) reading the upper 2 bits of the CBP coefficient and processing the CBP coefficient according to the following three conditions: firstly, if the value is 10, skipping the two bits without embedding secret information; second, if the value is 00, 2-bit or 1-bit secret information can be embedded, and the embedding rule is: when the secret information is 1, modifying the current two bits into 10; when the secret information is 00, no modification is made; when the secret information is 01, modifying the current two digits into 01; thirdly, if the value is 01, 1-bit secret information can be embedded; the embedding rule is: when the secret information is 1, modifying the current two bits into 10; when the secret information is 0, no modification is made;
5) reading the next macro block to obtain the macro block prediction mode, and jumping to the step 2); if the current frame is traversed completely, entropy coding is carried out on the frame again, the next frame is read, the step 1) is skipped, if all frames in the video sequence are processed completely or the secret information is completely embedded, the embedding process is finished, and the embedded video code stream is output;
the secret information extraction process comprises the following steps:
(1) entropy decoding the H.264 video compressed code stream, reading a first macro block of a current decoding frame, and obtaining a macro block prediction mode of the current decoding frame;
(2) determining whether the current macroblock prediction mode is I4, P16 × 16, P16 × 8, P8 × 16, or P8 × 8: if yes, reading the CBP coefficient value of the current macro block, and executing the step (3); otherwise, reading the next macro block and re-executing the step (2);
(3) sequentially reading the lower 4 bits of the CBP coefficient, and if the current bit is equal to 0, extracting 1-bit secret information, wherein the value of the secret information is 0; if the current bit is equal to 1, further reading the 8 × 8 luminance sub-block corresponding to the bit, and determining whether all coefficients contained in the luminance sub-block are 0: if yes, extracting 1-bit secret information, wherein the value of the secret information is 1, and simultaneously resetting the current CBP coefficient position to 0; otherwise, the bit is considered to be hidden without secret information and is not extracted;
(4) reading the high 2 bits of the CBP coefficient, and processing according to the following three conditions; A) if the value is 00, 2-bit secret information is extracted, the value is 00, and the current 2-bit CBP coefficient does not need to be modified; B) if the value is 01, the chrominance components of the macro block need to be read, if the DC coefficients are all 0, 2-bit secret information can be extracted, the value is 01, and the current 2-bit CBP coefficient is reset to 00; if the DC coefficients are not all 0, 1-bit secret information can be extracted, the value is 0, and the current 2-bit CBP coefficient does not need to be modified; C) if the value is 10, the chrominance component of the macro block needs to be read, if all the coefficients are 0, 1-bit secret information can be extracted, the value is 1, and the current 2-bit CBP coefficient is reset to 00; if the AC coefficients are all 0 but the DC coefficients are not all 0, 1-bit secret information can be extracted, the value is 1, and the current 2-bit CBP coefficient is reset to 01; if the AC coefficients are not all 0, skipping the current macro block without secret information extraction, and modifying the current 2-bit CBP coefficient;
(5) reading the next macro block to obtain the macro block prediction mode, jumping to the step (2), if the current frame is traversed, entropy coding the frame again, reading the next frame, jumping to the step (1), if all frames in the video sequence are processed or all secret information is extracted, ending the extraction process, and outputting the video code stream after secret information extraction.
To further achieve the object of the present invention, preferably, in step 1), the resolution size of the h.264 video is 352 × 288.
Preferably, in step 1), the frame structure of the test sequence of the macroblock prediction mode is IPPPP, and each GOP group contains 5 frames; when the test sequence is coded, the I frame and the P frame both adopt fixed quantization parameters 28, the coding frame rate is 25 frames/second, and 100 frames are coded; the macroblock prediction mode is held in the currMB- > mb _ type variable, and the CBP coefficient is held in the currMB- > CBP variable.
Preferably, in step 1), the CBP coefficient values of the macroblock have 6 bits in total, and the highest two bits are used to indicate the chroma component; when the chroma CBP coefficient value is 00, all the chroma coefficients are 0, and no transmission coefficient is used; when the chroma CBP coefficient value is 01, the AC coefficients in the cb and cr components are all 0, but at least one DC coefficient is not 0, and only the DC coefficient needs to be transmitted; when the chroma CBP coefficient value is 10, at least one AC coefficient in cb and cr components is not 0, and both the DC coefficient and the AC coefficient are transmitted; the lower 4 bits of the CBP coefficient values are used to represent the case of luminance components, the 4-bit coefficients corresponding from low to high to 4 8 × 8 luminance blocks from top to bottom and left to right, respectively; if the coefficient of a certain bit is 1, it indicates that at least one quantized DCT coefficient in the 8 x 8 luminance block corresponding to the bit is not 0; otherwise, it means that all the quantized DCT coefficients in the 8 × 8 luminance block are 0.
Preferably, in step 1), the macroblock prediction modes are I4, I16, IPCM, P16 × 16, P16 × 8, P8 × 16, P8 × 8 and Pskip modes; wherein the CBP coefficient values for the I16 macroblock are jointly coded with intra prediction mode, the I16 macroblock is not embeddable with secret information; the IPCM macro block directly transmits the pixel value of the image without QDCT coefficient and CBP coefficient value; the Pskip mode is a copy mode, a predicted pixel value is obtained through a motion vector, and a reconstructed pixel value of a macro block is directly equal to the predicted pixel value; pskip macroblocks have no QDCT coefficients and no CBP coefficient values; neither IPCM macroblocks nor Pskip macroblocks use CBP coefficient values to embed secret information.
Preferably, the sequence number of the 8 × 8 luminance sub-macroblock is represented by block8 × 8, and CBP _ block8 × 8 represents the coefficient value of the second block8 × 8 bits in the CBP coefficient; when embedding, reading a CBP _ block8 × 8 value corresponding to the 8 × 8 luminance sub-macro block, and if the CBP _ block8 × 8 is equal to 1, not embedding secret information; if CBP _ block8 x 8 is equal to 0, embedding 1 bit secret information, modifying CBP _ block8 x 8 to 1 if the secret information value is 1, otherwise, not modifying.
Preferably, the sequence number of the 8 × 8 luminance sub-macroblock is represented by block8 × 8, and CBP _ block8 × 8 represents the coefficient value of the second block8 × 8 bits in the CBP coefficient; when extracting, reading a CBP _ block8 × 8 value corresponding to an 8 × 8 brightness sub-macroblock, and if the CBP _ block8 × 8 is equal to 0, extracting 1-bit secret information, wherein the secret information value is 0; if CBP _ block8 is equal to 1, judging whether the coefficient values of the current 8 multiplied by 8 brightness sub-macro blocks are all 0; if yes, extracting 1-bit secret information, wherein the secret information value is 1, and recovering CBP _ block8 multiplied by 8 to be 0; if the non-zero coefficient value exists in the 8 x 8 luminance sub-macroblock, secret information is not extracted.
The invention integrates the entropy decoding module and the entropy coding module, executes embedding and extracting operations after entropy decoding and before entropy coding, skips complicated prediction and transformation processes, and greatly reduces the time complexity of the method.
In the step 1) and the step 2), only the macroblocks with the prediction modes of I4, P16 × 16, P16 × 8, P8 × 16 and P8 × 8 are selected for secret information embedding or extracting operation, and the macroblocks of other prediction modes do not meet the embedding/extracting conditions of the method.
In the step 3) and the step 4), the low 4-bit coefficient corresponding to the luminance component in the CBP coefficient of the macroblock is first subjected to secret letter embedding, and then the high 2-bit coefficient corresponding to the chrominance component in the CBP coefficient is subjected to secret letter embedding.
In the step (3) and the step (4), the low 4-bit coefficient corresponding to the luminance component in the CBP coefficient of the macroblock is subjected to secret letter extraction, and then the high 2-bit coefficient corresponding to the chrominance component in the CBP coefficient is subjected to secret letter extraction.
Compared with the prior art, the invention has the following advantages and beneficial effects:
1. the invention innovatively selects the CBP coefficient of the H.264/AVC video macro block for embedding. The CBP coefficient value determines whether the quantized DCT coefficient needs to be coded, and the invention carries out embedding when the quantized DCT coefficient does not need to be coded by modulating the CBP coefficient. After the embedding is complete, some zero coefficients that otherwise would not need to be transmitted are additionally encoded, but the quantized DCT coefficient values are not affected. The video pixel values are reconstructed from the quantized DCT coefficients, so the video visual quality is not affected at all.
2. The invention designs the steganography strategy by combining the characteristics of the CBP coefficient. The CBP coefficient value of the macro block has 6 bits in total, and the highest two bits are used for representing the situation of chrominance components; the lower 4 bits are used to represent the case of luminance components, and the 4-bit coefficients correspond from low to high to 4 8 × 8 luminance blocks from top to bottom and left to right, respectively. The strategy embeds secret information in a brightness CBP coefficient and a chroma CBP coefficient respectively, wherein 1-bit secret information is embedded when the brightness CBP coefficient is 0; when the brightness CBP coefficient is 1, secret information is not embedded; when the chroma CBP coefficient is 00, 1-2 bit secret information is embedded; when the chroma CBP coefficient is 01, 1-bit secret information is embedded; when the chroma CBP coefficient is 10, no secret information is embedded. The strategy only modifies the carrier when secret information is embedded, invalid modification does not exist, and the embedding efficiency is high.
3. The invention integrates the entropy decoding module and the entropy coding module, executes embedding and extracting operations after entropy decoding and before entropy coding, skips complicated prediction and transformation processes, and greatly reduces the time complexity of the method.
Drawings
FIG. 1 is a block diagram of the embedding and extracting process flow of the H.264/AVC video visual quality lossless and reversible information hiding method based on the macroblock CBP coefficient value.
Fig. 2(a) is a state transition diagram of the luminance CBP coefficient (lower 4 bits) embedding strategy.
Fig. 2(b) chroma CBP coefficients (2 bits high) are embedded in the policy state transition diagram.
FIG. 3 is a flow chart of a method for verifying validity of the method of the present invention.
Detailed Description
For a better understanding of the present invention, the present invention will be further described with reference to the following examples and drawings, but the embodiments of the present invention are not limited thereto.
Example (b):
the embodiment provides a method for hiding lossless and reversible information of H.264/AVC video visual quality based on a macroblock CBP coefficient value, wherein a block diagram of an embedding (or extracting) process is shown in FIG. 1, and the embedding (or extracting) process is mainly divided into five steps, including obtaining a CBP coefficient and a macroblock prediction mode, selecting a macroblock meeting requirements, embedding (or extracting) a luminance CBP coefficient, embedding (or extracting) a chrominance CBP coefficient, and continuing entropy coding to obtain a compressed video. The following describes the implementation of the present invention in detail by taking a video library composed of 9 CIF videos as an example. In the embodiment, an H.264/AVC official test model JM-18.6 is adopted as a codec, and the resolution size of all test videos is 352 multiplied by 288. The frame structure of the test sequence is IPPPP, and each GOP group contains 5 frames. When the test sequence is coded, I, P frames adopt fixed quantization parameters 28, the coding frame rate is 25 frames/second, and 100 frames are coded.
First, a CBP coefficient and a macroblock prediction mode are obtained.
And carrying out entropy decoding operation on the H.264/AVC video file to obtain a CBP coefficient and a macroblock prediction mode. In JM-18.6, the prediction mode of the macroblock is held in the currMB- > mb _ type variable, and the CBP coefficient is held in the currMB- > CBP variable. The CBP coefficient values of the macroblock have 6 bits in total, and the highest two bits are used to indicate the case of the chrominance components. When the chroma CBP coefficient value is 00, all the chroma coefficients are 0, and no transmission coefficient is used; when the chroma CBP coefficient value is 01, the AC coefficients in the cb and cr components are all 0, but at least one DC coefficient is not 0, and only the DC coefficient needs to be transmitted; a chroma CBP coefficient value of 10 indicates that at least one of the cb and cr components has an AC coefficient other than 0, and that both the DC coefficient and the AC coefficient are transmitted. The lower 4 bits of the CBP coefficient values are used to represent the case of luminance components, and the 4-bit coefficients correspond from low to high to 4 8 × 8 luminance blocks from top to bottom and left to right, respectively. If a bit is 1, it indicates that at least one quantized DCT coefficient in the 8 x 8 luminance block corresponding to the bit is not 0; otherwise, it means that all the quantized DCT coefficients in the 8 × 8 luminance block are 0. In the encoding process, when a certain bit of the CBP coefficient value is 0, the encoding process of the corresponding block quantization DCT coefficient can be directly skipped, and only the CBP coefficient value needs to be stored. Correspondingly, if a certain bit of the CBP coefficient value is found to be 0 during the decoding process, all the quantized DCT coefficients of the corresponding block may be set to be 0 directly.
And secondly, selecting the macro blocks meeting the requirements.
Considering the case of coding using only I-frames and P-frames, macroblocks have several basic prediction types as follows: i4, I16, IPCM, P16 × 16, P16 × 8, P8 × 16, P8 × 8 and Pskip patterns. Among them, the CBP coefficient values of the I16 macroblock are jointly coded with the intra prediction mode, so the I16 macroblock cannot embed secret information. The IPCM macroblock directly transfers the pixel values of the image without prediction and transformation processes and without QDCT coefficients and CBP coefficient values. The Pskip mode, also called copy mode, obtains the predicted pixel values from the motion vectors, making the reconstructed pixel values of the macroblock directly equal to the predicted pixel values. The Pskip macroblock also has no QDCT coefficients and no CBP coefficient values. Neither IPCM macroblock nor Pskip macroblock can use CBP coefficient values to embed secret information. The invention selects I4, P16 × 16, P16 × 8, P8 × 16, P8 × 8 type macro blocks as carriers to embed secret information.
And thirdly, embedding and extracting the brightness CBP coefficient.
The 4 lower bits of the macroblock CBP coefficient values are luminance component CBP coefficient values, which represent coefficient value cases in four luminance blocks of 8 × 8 size, respectively. The number of 8 × 8 luminance sub-macroblocks is represented by block8 × 8, and CBP _ block8 × 8 represents the coefficient value of the second block8 × 8 bits of CBP coefficients.
When embedded, the state transition diagram is shown in fig. 2 (a). Reading a CBP _ block8 x 8 value corresponding to the 8 x 8 luminance sub-macro block, and if the CBP _ block8 x 8 is equal to 1, not embedding secret information; if CBP _ block8 x 8 is equal to 0, embedding 1 bit secret information, modifying CBP _ block8 x 8 to 1 if the secret information value is 1, otherwise, not modifying.
When extracting, reading a CBP _ block8 × 8 value corresponding to an 8 × 8 brightness sub-macroblock, and if the CBP _ block8 × 8 is equal to 0, extracting 1-bit secret information, wherein the secret information value is 0; if CBP _ block8 × 8 is equal to 1, it is determined whether the current luminance sub-macroblock coefficient values of 8 × 8 are all 0. If yes, extracting 1-bit secret information, wherein the secret information value is 1, and recovering CBP _ block8 multiplied by 8 to be 0; if the non-zero coefficient value exists in the 8 multiplied by 8 brightness sub-macro block, secret information is not extracted;
and fourthly, embedding and extracting the chroma CBP coefficient.
The 2 bits of the macroblock CBP coefficient values are chroma component CBP coefficient values, and the two chroma components share a chroma CBP variable.
When embedded, the state transition diagram is shown in fig. 2 (b). The chroma CBP coefficient value is read, and if the chroma CBP coefficient value is equal to 00, 2 bits or 1 bit secret information can be embedded. When the secret information is 00, no modification is made; when the secret information is 01, modifying the chroma CBP coefficient value to be 01; when the secret is 1, the chroma CBP coefficient value is modified to 10. If the chroma CBP coefficient value is equal to 01, 1-bit secret information can be embedded. When the secret is 1, the chroma CBP coefficient value is modified to 10, otherwise, no modification is made. If the chroma CBP coefficient value is equal to 10, no secret information is embedded, and the chroma CBP coefficient value is not modified.
During extraction, reading a chroma CBP coefficient value, and if the chroma CBP coefficient value is equal to 00, extracting 2-bit secret information, wherein the secret information value is 00; if the chroma CBP coefficient value is equal to 01, 2-bit or 1-bit secret can be extracted. When the chroma component DC coefficients are all 0, extracting secret information 01, and restoring the chroma CBP coefficient value to 00; when the chroma component DC coefficients are not all 0, extracting secret information 0 without modifying chroma CBP coefficient values; if the chroma CBP coefficient value is equal to 10, reading the chroma component of the macro block, extracting 1-bit secret information when all coefficients of the chroma component are 0, and recovering the chroma CBP coefficient value to 00 if the value of the secret information is 1; when the chrominance component AC coefficients are all 0 but the DC coefficients are not all 0, extracting 1-bit secret information, wherein the value of the secret information is 1, and recovering the chrominance CBP coefficient value to be 01; when the chroma component AC coefficients are not all 0, skipping the current macro block without secret information extraction and modifying the chroma CBP coefficient value;
and fifthly, continuously entropy coding to obtain a compressed video.
And entropy coding to obtain the compressed video embedded with the secret information. In the embodiment, 9 video sequences of akiyo, bus, city, crop, foreman, ice, mobile, paris and students are selected as the test video. The method comprises various typical video sequences with single and complex background, slow moving picture, fast motion and the like, and the resolution of all test videos is 352 multiplied by 288. The frame structure of the test sequence is IPPPP, and each GOP group contains 5 frames. When the test sequence is coded, I, P frames adopt fixed quantization parameters 28, the coding frame rate is 25 frames/second, and 100 frames are coded. The steganography process has no influence on the subjective visual quality of the video and has excellent visual concealment. Taking the luminance component as an example, as shown in fig. 3, the present invention only embeds when the CBP coefficient is 0, even if the CBP is modified to 1, the encoder only encodes all 0 luminance sub-block coefficients at once, and the quantized DCT coefficients of all 0 can be obtained in the same way when decoding, and the chrominance components are the same. Therefore, the invention can obtain completely lossless video after embedding.
The embodiments of the present invention are not limited thereto, and those skilled in the art can substitute or change the technical solution and idea of the present invention within the scope of the present invention disclosed by the present invention, and belong to the protection scope of the present invention.
Claims (7)
- The H.264/AVC video visual quality lossless reversible information hiding method comprises two parts of secret letter embedding and secret letter extraction, and is characterized in that the secret letter embedding comprises the following steps:1) entropy decoding the H.264 video compressed code stream, reading a first macro block of a current decoding frame, and obtaining a macro block prediction mode of the current decoding frame;2) determining whether the current macroblock prediction mode is I4, P16 × 16, P16 × 8, P8 × 16, or P8 × 8: if yes, reading the CBP coefficient value of the current macro block, and executing the step 3); otherwise, reading the next macro block and re-executing the step 2);3) sequentially reading the lower 4 bits of the CBP coefficient, if the current bit is equal to 1, skipping the bit and not embedding secret information; if the current bit is equal to 0, 1-bit secret information is embedded; the embedding rule is that the current bit is modified to be consistent with the value of the secret information to be embedded;4) reading the upper 2 bits of the CBP coefficient and processing the CBP coefficient according to the following three conditions: firstly, if the value is 10, skipping the two bits without embedding secret information; second, if the value is 00, 2-bit or 1-bit secret information can be embedded, and the embedding rule is: when the secret information is 1, modifying the current two bits into 10; when the secret information is 00, no modification is made; when the secret information is 01, modifying the current two digits into 01; thirdly, if the value is 01, 1-bit secret information can be embedded; the embedding rule is: when the secret information is 1, modifying the current two bits into 10; when the secret information is 0, no modification is made;5) reading the next macro block to obtain the macro block prediction mode, and jumping to the step 2); if the current frame is traversed completely, entropy coding is carried out on the frame again, the next frame is read, the step 1) is skipped, if all frames in the video sequence are processed completely or the secret information is completely embedded, the embedding process is finished, and the embedded video code stream is output;the secret information extraction process comprises the following steps:(1) entropy decoding the H.264 video compressed code stream, reading a first macro block of a current decoding frame, and obtaining a macro block prediction mode of the current decoding frame;(2) determining whether the current macroblock prediction mode is I4, P16 × 16, P16 × 8, P8 × 16, or P8 × 8: if yes, reading the CBP coefficient value of the current macro block, and executing the step (3); otherwise, reading the next macro block and re-executing the step (2);(3) sequentially reading the lower 4 bits of the CBP coefficient, and if the current bit is equal to 0, extracting 1-bit secret information, wherein the value of the secret information is 0; if the current bit is equal to 1, further reading the 8 × 8 luminance sub-block corresponding to the bit, and determining whether all coefficients contained in the luminance sub-block are 0: if yes, extracting 1-bit secret information, wherein the value of the secret information is 1, and simultaneously resetting the current CBP coefficient position to 0; otherwise, the bit is considered to be hidden without secret information and is not extracted;(4) reading the high 2 bits of the CBP coefficient, and processing according to the following three conditions; A) if the value is 00, 2-bit secret information is extracted, the value is 00, and the current 2-bit CBP coefficient does not need to be modified; B) if the value is 01, the chrominance components of the macro block need to be read, if the DC coefficients are all 0, 2-bit secret information can be extracted, the value is 01, and the current 2-bit CBP coefficient is reset to 00; if the DC coefficients are not all 0, 1-bit secret information can be extracted, the value is 0, and the current 2-bit CBP coefficient does not need to be modified; C) if the value is 10, the chrominance component of the macro block needs to be read, if all the coefficients are 0, 1-bit secret information can be extracted, the value is 1, and the current 2-bit CBP coefficient is reset to 00; if the AC coefficients are all 0 but the DC coefficients are not all 0, 1-bit secret information can be extracted, the value is 1, and the current 2-bit CBP coefficient is reset to 01; if the AC coefficients are not all 0, skipping the current macro block without secret information extraction, and modifying the current 2-bit CBP coefficient;(5) reading the next macro block to obtain the macro block prediction mode, jumping to the step (2), if the current frame is traversed, entropy coding the frame again, reading the next frame, jumping to the step (1), if all frames in the video sequence are processed or all secret information is extracted, ending the extraction process, and outputting the video code stream after secret information extraction.
- 2. The h.264/AVC video visual quality lossless invertible information hiding method according to claim 1, wherein: in step 1), the resolution of the h.264 video is 352 × 288.
- 3. The h.264/AVC video visual quality lossless invertible information hiding method according to claim 1, wherein: in step 1), the frame structure of the test sequence of the macro block prediction mode is IPPPP, and each GOP group contains 5 frames; when the test sequence is coded, the I frame and the P frame both adopt fixed quantization parameters 28, the coding frame rate is 25 frames/second, and 100 frames are coded; the macroblock prediction mode is held in the currMB- > mb _ type variable, and the CBP coefficient is held in the currMB- > CBP variable.
- 4. The h.264/AVC video visual quality lossless invertible information hiding method according to claim 1, wherein: in step 1), the CBP coefficient values of the macro blocks have 6 bits in total, and the highest two bits are used for representing the conditions of chrominance components; when the chroma CBP coefficient value is 00, all the chroma coefficients are 0, and no transmission coefficient is used; when the chroma CBP coefficient value is 01, the AC coefficients in the cb and cr components are all 0, but at least one DC coefficient is not 0, and only the DC coefficient needs to be transmitted; when the chroma CBP coefficient value is 10, at least one AC coefficient in cb and cr components is not 0, and both the DC coefficient and the AC coefficient are transmitted; the lower 4 bits of the CBP coefficient values are used to represent the case of luminance components, the 4-bit coefficients corresponding from low to high to 4 8 × 8 luminance blocks from top to bottom and left to right, respectively; if the coefficient of a certain bit is 1, it indicates that at least one quantized DCT coefficient in the 8 x 8 luminance block corresponding to the bit is not 0; otherwise, it means that all the quantized DCT coefficients in the 8 × 8 luminance block are 0.
- 5. The h.264/AVC video visual quality lossless invertible information hiding method according to claim 1, wherein: in step 1), the macroblock prediction modes are I4, I16, IPCM, P16 × 16, P16 × 8, P8 × 16, P8 × 8 and Pskip modes; wherein the CBP coefficient values for the I16 macroblock are jointly coded with intra prediction mode, the I16 macroblock is not embeddable with secret information; the IPCM macro block directly transmits the pixel value of the image without QDCT coefficient and CBP coefficient value; the Pskip mode is a copy mode, a predicted pixel value is obtained through a motion vector, and a reconstructed pixel value of a macro block is directly equal to the predicted pixel value; pskip macroblocks have no QDCT coefficients and no CBP coefficient values; neither IPCM macroblocks nor Pskip macroblocks use CBP coefficient values to embed secret information.
- 6. The h.264/AVC video visual quality lossless invertible information hiding method according to claim 1, wherein: the sequence number of the 8 × 8 luminance sub-macroblock is represented by block8 × 8, and CBP _ block8 × 8 represents the coefficient value of the second block8 × 8 bits in the CBP coefficients; when embedding, reading a CBP _ block8 × 8 value corresponding to the 8 × 8 luminance sub-macro block, and if the CBP _ block8 × 8 is equal to 1, not embedding secret information; if CBP _ block8 x 8 is equal to 0, embedding 1 bit secret information, modifying CBP _ block8 x 8 to 1 if the secret information value is 1, otherwise, not modifying.
- 7. The h.264/AVC video visual quality lossless invertible information hiding method according to claim 1, wherein: the sequence number of the 8 × 8 luminance sub-macroblock is represented by block8 × 8, and CBP _ block8 × 8 represents the coefficient value of the second block8 × 8 bits in the CBP coefficients; when extracting, reading a CBP _ block8 × 8 value corresponding to an 8 × 8 brightness sub-macroblock, and if the CBP _ block8 × 8 is equal to 0, extracting 1-bit secret information, wherein the secret information value is 0; if CBP _ block8 is equal to 1, judging whether the coefficient values of the current 8 multiplied by 8 brightness sub-macro blocks are all 0; if yes, extracting 1-bit secret information, wherein the secret information value is 1, and recovering CBP _ block8 multiplied by 8 to be 0; if the non-zero coefficient value exists in the 8 x 8 luminance sub-macroblock, secret information is not extracted.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811423158.4A CN109361926B (en) | 2018-11-27 | 2018-11-27 | Lossless reversible information hiding method for H.264/AVC video visual quality |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811423158.4A CN109361926B (en) | 2018-11-27 | 2018-11-27 | Lossless reversible information hiding method for H.264/AVC video visual quality |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109361926A CN109361926A (en) | 2019-02-19 |
| CN109361926B true CN109361926B (en) | 2020-04-28 |
Family
ID=65343194
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811423158.4A Expired - Fee Related CN109361926B (en) | 2018-11-27 | 2018-11-27 | Lossless reversible information hiding method for H.264/AVC video visual quality |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109361926B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110881128B (en) * | 2019-11-27 | 2021-01-19 | 华南理工大学 | JPEG image reversible data hiding method |
| CN117354529B (en) * | 2023-11-28 | 2024-03-12 | 广东匠芯创科技有限公司 | Image processing method based on video coding system, electronic equipment and medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104038761A (en) * | 2014-06-05 | 2014-09-10 | 宁波工程学院 | Encrypted domain H.264/AVC video data hiding method based on CABAC binary string mapping |
| US8848791B2 (en) * | 2010-09-29 | 2014-09-30 | Futurewei Technologies, Inc. | Compressed domain video watermarking |
| CN106817586A (en) * | 2016-12-22 | 2017-06-09 | 宁波工程学院 | A kind of H.264/AVC video reversible data concealing method |
| CN107888925A (en) * | 2017-11-03 | 2018-04-06 | 清华大学深圳研究生院 | A kind of embedding grammar and detection method of digital video hiding information |
-
2018
- 2018-11-27 CN CN201811423158.4A patent/CN109361926B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8848791B2 (en) * | 2010-09-29 | 2014-09-30 | Futurewei Technologies, Inc. | Compressed domain video watermarking |
| CN104038761A (en) * | 2014-06-05 | 2014-09-10 | 宁波工程学院 | Encrypted domain H.264/AVC video data hiding method based on CABAC binary string mapping |
| CN106817586A (en) * | 2016-12-22 | 2017-06-09 | 宁波工程学院 | A kind of H.264/AVC video reversible data concealing method |
| CN107888925A (en) * | 2017-11-03 | 2018-04-06 | 清华大学深圳研究生院 | A kind of embedding grammar and detection method of digital video hiding information |
Non-Patent Citations (1)
| Title |
|---|
| 修改标志位的大容量无损HEVC信息隐藏方法;胡永健，龚文斌，刘琲贝，刘烁炜，朱美能;《华南理工大学学报》;20180831;第46卷(第5期);第1-8页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109361926A (en) | 2019-02-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Yang et al. | An information hiding algorithm based on intra-prediction modes and matrix coding for H. 264/AVC video stream | |
| CN108366259B (en) | High-fidelity H.264/AVC video three-coefficient reversible steganography method | |
| JP7085009B2 (en) | Methods and devices for merging multi-sign bit concealment and residual sign prediction | |
| KR101808327B1 (en) | Video encoding/decoding method and apparatus using paddding in video codec | |
| Liu et al. | A robust and improved visual quality data hiding method for HEVC | |
| JP2006187024A (en) | Image data decoding method and computer readable medium recording program therefor | |
| CN109819260B (en) | Video steganography method and device based on multi-embedded domain fusion | |
| CN109361926B (en) | Lossless reversible information hiding method for H.264/AVC video visual quality | |
| WO2023065891A1 (en) | Multimedia data processing method and apparatus, device, computer-readable storage medium and computer program product | |
| US11973976B2 (en) | Systems and methods for performing padding in coding of a multi-dimensional data set | |
| Dong et al. | Adaptive HEVC steganography based on steganographic compression efficiency degradation model | |
| CN113612898B (en) | Robust covert communication device for resisting JPEG image downsampling | |
| TWI533670B (en) | Method for decoding picture in form of bit-stream | |
| Ma et al. | A cross channel context model for latents in deep image compression | |
| WO2015028098A1 (en) | Method and system for embedding information and authenticating a h.264 video using a digital watermark | |
| CN113329229A (en) | High-capacity hiding method for H.265 video information with high-efficiency fidelity | |
| CN108353192B (en) | Hierarchical deblocking filtering in video processing systems and methods | |
| Koya et al. | Analysis of application of arithmetic coding on dct and dct-dwt hybrid transforms of images for compression | |
| Peng et al. | An optimized algorithm based on generalized difference expansion method used for HEVC reversible video information hiding | |
| Gujjunoori et al. | Reversible Data Embedding for MPEG-4 Video Using Middle Frequency DCT Coefficients. | |
| Kunte | Framework for reversible data hiding using cost-effective encoding system for video steganography. | |
| Wang et al. | Video reversible data hiding: A systematic review | |
| CN116320471B (en) | Video information hiding method, system, equipment and video information extracting method | |
| CN104469377A (en) | H.264 file authentication playing method based on video steganography | |
| CN115150627B (en) | DST-based video compression robustness-resistant blind watermarking method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20200428 |