CN112714230A - Robust video steganography method and device based on audio side channel - Google Patents
Robust video steganography method and device based on audio side channel Download PDFInfo
- Publication number
- CN112714230A CN112714230A CN202011388479.2A CN202011388479A CN112714230A CN 112714230 A CN112714230 A CN 112714230A CN 202011388479 A CN202011388479 A CN 202011388479A CN 112714230 A CN112714230 A CN 112714230A
- Authority
- CN
- China
- Prior art keywords
- video
- audio
- steganography
- carrier
- embedding
- 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.)
- Pending
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
- H04N1/32—Circuits or arrangements for control or supervision between transmitter and receiver or between image input and image output device, e.g. between a still-image camera and its memory or between a still-image camera and a printer device
- H04N1/32101—Display, printing, storage or transmission of additional information, e.g. ID code, date and time or title
- H04N1/32144—Display, printing, storage or transmission of additional information, e.g. ID code, date and time or title embedded in the image data, i.e. enclosed or integrated in the image, e.g. watermark, super-imposed logo or stamp
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
- H04N1/32—Circuits or arrangements for control or supervision between transmitter and receiver or between image input and image output device, e.g. between a still-image camera and its memory or between a still-image camera and a printer device
- H04N1/32101—Display, printing, storage or transmission of additional information, e.g. ID code, date and time or title
- H04N1/32144—Display, printing, storage or transmission of additional information, e.g. ID code, date and time or title embedded in the image data, i.e. enclosed or integrated in the image, e.g. watermark, super-imposed logo or stamp
- H04N1/32149—Methods relating to embedding, encoding, decoding, detection or retrieval operations
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
- H04N1/32—Circuits or arrangements for control or supervision between transmitter and receiver or between image input and image output device, e.g. between a still-image camera and its memory or between a still-image camera and a printer device
- H04N1/32101—Display, printing, storage or transmission of additional information, e.g. ID code, date and time or title
- H04N1/32144—Display, printing, storage or transmission of additional information, e.g. ID code, date and time or title embedded in the image data, i.e. enclosed or integrated in the image, e.g. watermark, super-imposed logo or stamp
- H04N1/32149—Methods relating to embedding, encoding, decoding, detection or retrieval operations
- H04N1/32154—Transform domain methods
- H04N1/32165—Transform domain methods using cosine transforms
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
- H04N1/32—Circuits or arrangements for control or supervision between transmitter and receiver or between image input and image output device, e.g. between a still-image camera and its memory or between a still-image camera and a printer device
- H04N1/32101—Display, printing, storage or transmission of additional information, e.g. ID code, date and time or title
- H04N1/32144—Display, printing, storage or transmission of additional information, e.g. ID code, date and time or title embedded in the image data, i.e. enclosed or integrated in the image, e.g. watermark, super-imposed logo or stamp
- H04N1/32149—Methods relating to embedding, encoding, decoding, detection or retrieval operations
- H04N1/32267—Methods relating to embedding, encoding, decoding, detection or retrieval operations combined with processing of the image
- H04N1/32277—Compression
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The invention relates to a robust video steganography method and a robust video steganography device based on an audio side channel. The method includes the steps that accompanying audio and video are divided to obtain carrier video and carrier audio; embedding a secret message in a carrier video to obtain a steganographic video; embedding a steganography protocol in the carrier audio to obtain a steganographically-generated audio; and merging and packaging the video subjected to steganography and the audio subjected to steganography into an accompanying sound steganography video. When extracting the secret information, the sound accompaniment steganography video is shunted to obtain a carrier video and a carrier audio, a steganography protocol is extracted from the carrier audio, then a robust area is selected, and the secret information is extracted from the carrier video. The invention establishes two hidden channels, the audio side channel is embedded in a communication protocol, the video channel is embedded in a secret message, the video steganography method has strong robustness, can self-adaptively select a steganography area, can be used for social network hidden communication, and enhances the concealment and the safety of the whole communication process.
Description
Technical Field
The invention belongs to the field of information security, and relates to a robust steganography method and a robust steganography device based on video accompanying sound streams, which can be used for covert communication of social network channels.
Background
The information hiding technology in the field of information security consists of two parts, namely steganography and steganalysis, wherein the steganalysis is mainly used for researching how to embed secret information into carriers such as images, audios and videos to achieve the aim of covert communication, and the steganalysis is used for distinguishing common carriers from files subjected to steganalysis by adopting methods such as machine learning and mode recognition.
Generally, when designing a steganographic algorithm, the following factors need to be considered comprehensively:
1) imperceptibility: the carriers before and after steganography are required to be indistinguishable to human senses, namely whether the carriers are subjected to steganography cannot be judged only by means of human hearing and vision;
2) robustness: the steganographic file is required to reliably and completely recover the embedded secret information after signal processing operation or malicious modification;
3) embedding capacity: when covert communication is required, secret information data embedded in a carrier file should be increased as much as possible under the condition of ensuring certain steganography security;
4) embedding efficiency: the modification degree of the carrier is required to be reduced as much as possible on the premise that the quantity of the embedded information is certain;
5) safety: the steganographic embedding operation is required not to cause obvious disturbance to the statistical characteristics of the carrier file, so that a steganographic analyst (attacker) cannot detect the existence of steganographic behaviors by using simple statistical characteristics.
Through years of research of expert scholars, image steganography is greatly improved and developed, and a plurality of classical steganography algorithms such as F5, HUGO and the like appear. However, with the increasing change of internet technology and the emerging of video coding standards, video has become the most popular and replaceable transmission medium, and the advantages of video over image are that the application scene is richer and the amount of information transmitted is larger. Therefore, video is gradually replacing images and becoming a more suitable carrier in steganography.
Video steganography can be generally divided into the following three main categories according to the embedded domain of secret information:
1) spatial domain video steganography: before video compression coding, original pixel values of video frames or transformation domain coefficients of the original pixel values are directly modified to embed secret information;
2) compressed domain video steganography: steganographic embedding operations are tightly coupled with video compression coding by utilizing certain modules or features in the video compression coding framework to embed secret information.
3) Format video steganography: i.e. secret information has been embedded using reserved (or redundant) fields of the video encapsulation format (also called container).
The compressed domain video steganography algorithm embeds secret messages, such as motion vectors, intra-frame prediction modes, transformation coefficients and the like, into a video code stream by modifying syntax elements of the video code stream, and the algorithm is interfered by noise in a transmission process of a lossy channel, so that a receiver cannot correctly extract the secret messages. Therefore, the compressed domain video steganography algorithm has a large selectivity of an embedded domain and a wide variety of algorithms, but has no robustness.
The format video steganography algorithm embeds the secret information by utilizing the redundant field of the video packaging format, the embedding capacity of the algorithm is low, and the secret information can be erased in the video transcoding process and cannot resist video transcoding.
The spatial domain video steganography algorithm can be divided into the following two categories according to the difference of secret message embedding domains:
1) video steganography algorithm based on pixel domain: the main idea is to embed the secret message by modifying the pixel values or the distribution of pixel values of the video frame. For example: the scholars propose to combine hash indexing with Least Significant Bit (LSB) techniques to embed secret messages by modifying the pixel values of video frames.
2) The video steganography algorithm based on the spatial domain transformation domain comprises the following steps: the main idea is to transform the video frame pixel values to other domains using a matrix transform and then to perform modifications in the transform domain to embed the secret message. For example: the scholars propose to transform the video frame pixel values into the frequency domain using Discrete Wavelet Transform (DWT) and then embed the secret message by modifying the low frequency sub-bands.
The spatial domain video steganography algorithm is simple to operate and has stronger robustness. However, since the original video frame needs to be lossy compressed, the steganographic algorithm generally needs to adopt an error control coding technique or a repeated embedding means to ensure that the receiving end can correctly recover the secret information therein.
Disclosure of Invention
The invention aims to provide a robust video steganography scheme which can be used for a social network lossy transcoding channel, wherein the scheme utilizes an audio construction side channel and is modified in a transform domain to embed a secret message. Meanwhile, the scheme is combined with an error correction coding technology, and the lossy transcoding of the social network can be effectively resisted.
Compared with other spatial domain video steganography algorithms, the invention provides a video frame selection strategy, which can adaptively select a robust region, and improves the algorithm robustness by modifying block statistical characteristics to embed secret information. Meanwhile, the invention embeds the protocol by constructing the audio side channel, so that the receiving end can correctly extract the secret message.
According to research, all the current video steganography algorithms construct a single covert channel to embed secret information, in the embedding process, two covert channels are respectively established across media, an audio side channel is embedded in a communication protocol, and a video channel is embedded in secret information. In the extraction process, protocol information is first extracted from the audio side channel, then a robust region is selected, and secret information is extracted from the video channel.
Specifically, the technical scheme adopted by the invention is as follows:
a robust video steganography method based on an audio side channel comprises an embedding process of secret information, wherein the embedding process of the secret information comprises the following steps:
shunting the accompanying audio and video to obtain a carrier video and a carrier audio;
embedding a secret message in a carrier video to obtain a steganographic video;
embedding a steganography protocol in the carrier audio to obtain a steganographically-generated audio;
and merging and packaging the video subjected to steganography and the audio subjected to steganography into an accompanying sound steganography video.
Further, the splitting the audio and video to obtain the carrier video and the carrier audio includes: transcoding the carrier audio into WAV format audio, decoding the carrier video to an airspace, and acquiring a video frame brightness component sequence, namely a Y component sequence.
Further, the embedding the secret message in the carrier video to obtain the steganographically written video includes:
encoding the secret message with BCH: encoding the secret message by using an error correcting code BCH code to generate an encoded string m of the secret message to be embedded1;
For the Y component sequence, performing the following operations:
1) selecting a robust frame and generating a steganographic protocol: selecting a frame with a low half error rate in a video as a robust frame and other frames as non-robust frames by utilizing a pre-embedding process, and generating a mark sequence as a steganography protocol P;
2) extracting a video embedding domain: discrete cosine transform is carried out on each pixel block to generate four sub-bands of LL, LH, HL and HH, and then Singular Value Decomposition (SVD) is carried out on the LL sub-band to obtain a diagonal matrixObtaining the maximum singular value alpha1As a video embedding domain;
3) quantization index modulation embedding: for maximum singular value alpha1Carrying out quantization index modulation and embedding a string m of information to be embedded1;
4) Generating a sequence of luminance components: maximum singular value alpha 'after QIM modulation'1Writing back the diagonal matrix sigma, recalculating to generate a new pixel block, merging all the modified pixel blocks, and writing the brightness frame back to the Y component sequence, thereby obtaining the video subjected to steganography.
Further, embedding a steganographic protocol in the carrier audio to obtain a steganographic audio includes:
utilizing BCH coding steganography protocol: encoding the steganographic protocol P by using an error correcting code BCH code to generate an encoded string f of secret information to be embedded1;
For the WAV format audio, the following operations are executed:
1) sampling point segmentation: segmenting a sampling point sequence in the WAV format audio according to a fixed step length to obtain a plurality of sampling segments;
2) extracting an audio embedded domain: carrying out DWT conversion on each sampling point segment to generate a low-frequency sub-band L and a high-frequency sub-band H; then, the L subband is subjected to SVD decomposition to obtain a singular value array U ═ (μ)1 μ2…μn) Obtaining the maximum singular value mu1As an audio embedded domain;
3) quantization index modulation embedding: for maximum singular value mu1Performing QIM modulation, and embedding a string f to be embedded with secret information1;
4) Writing the maximum singular value after QIM modulation back to the array U, recalculating to generate new sampling point segments, combining all modified sampling point segments, writing the sampling point segments back to the sampling point sequence, and generating the WAV format audio after steganography.
A robust video steganography method based on an audio side channel comprises a secret information extraction process, wherein the secret information extraction process extracts secret information from sound steganography video generated by the method; the secret information extraction process comprises the following steps:
shunting the audio steganography video to obtain a carrier video and a carrier audio;
the steganographic protocol is extracted from the carrier audio and then the robust region is selected and the secret information is extracted from the carrier video.
Further, the secret information extraction process includes the following steps:
1) pretreatment: shunting the audio and video subjected to steganography to obtain a carrier video and a carrier audio, transcoding the carrier audio into WAV format audio, decoding the carrier video to an airspace, and acquiring a video frame brightness component sequence, namely a Y component sequence;
2) executing the following operations on the WAV format audio obtained in the step 1):
2.1) segmentation of sampling points: segmenting a sampling point sequence in the WAV format audio according to a fixed step length to obtain a plurality of sampling segments;
2.2) extracting an audio embedded domain: carrying out DWT conversion on each sampling point segment to generate a low-frequency subband L and a high-frequency subband H, and then carrying out SVD (singular value decomposition) on the L subband to obtain a singular value array U ═ (mu)1 μ2 … μn) Obtaining the maximum singular value mu1As an audio embedded domain;
2.3) quantization index modulation extraction: using QIM method to measure the maximum singular value mu1Extracting to generate a secret string f'1;
3) Secret information string f 'is matched by using error correcting code BCH code'1Decoding to generate a steganographic protocol P;
4) performing the following operations on the Y component sequence obtained in the step 1):
4.1) selecting robust frames: selecting a robust frame in the Y component sequence by using the steganographic protocol P obtained in the step 3);
4.2) extracting a video embedding domain: discrete cosine transform (DWT) is carried out on each pixel block to generate four sub-bands of LL, LH, HL and HH, and then Singular Value Decomposition (SVD) is carried out on the LL sub-band to obtain a diagonal matrixObtaining the maximum singular value alpha1As a video embedding domain;
4.3) quantization index modulation extraction: using QIM algorithm to measure the maximum singular value alpha1Extracting to obtain secret string m'1;
5) Secret information string m 'is paired by using error correction code BCH code'1Decoding is carried out to generate a secret message, and the secret message is decrypted to generate a message plaintext.
A robust video steganography device based on an audio side channel comprises a secret information embedding unit and/or a secret information extracting unit; the secret information embedding unit embeds the secret information by adopting the embedding process of the secret information in the method to generate the sound steganography video; the secret information extraction unit extracts the secret information from the sound steganography video by adopting the secret information extraction process in the method.
The robust video steganography method has the following beneficial effects in the related technical field:
1) and the robustness is strong. At present, all video steganography methods based on a compressed domain have no robustness, and the main reason is that video transcoding can greatly change video syntax elements such as transformation coefficients, quantization parameters, motion vectors and the like, so that a steganography embedded domain is damaged. The invention is a video steganography algorithm based on airspace, and the change of the video syntax element does not influence the airspace characteristics, so the invention has strong robustness and can resist the recompression and transcoding of the video. In particular, the invention is still applicable when the transmission channel is a lossy transcoding channel.
2) The steganographic areas may be adaptively selected. The steganography area of the existing video steganography algorithm is only related to the algorithm and is not related to the content of a video carrier. During the transmission of the lossy channel, different videos may be introduced with different degrees of noise, thereby affecting the robustness of the steganographic algorithm. The invention can select the steganographic area according to the content of the video carrier, adaptively select the robust frame to embed the secret message, and improve the robustness of the algorithm.
3) Can be used for social network covert communication. Since the social network transcoding mechanism is a closed black box, the noise introduced by the social network transcoding channel is uncertain. The existing video steganography algorithm cannot carry out covert communication through a social network, and the steganography video is often transmitted through a manual or lossless channel, so that the risk of communication exposure is increased. The invention can utilize the social network to carry out covert communication, thereby enhancing the concealment and the safety of the whole communication process.
Drawings
FIG. 1 is a schematic diagram of secret information embedding of the present invention;
FIG. 2 is a schematic diagram of secret information extraction of the present invention;
FIG. 3 is a flow chart of the pre-embedding of the present invention;
FIG. 4 is a flow chart of the audio side channel embedding of the present invention;
FIG. 5 is a flow chart of the video channel embedding of the present invention;
FIG. 6 is a flow chart of the audio side channel extraction of the present invention;
fig. 7 is a flow chart of video channel extraction in accordance with the present invention.
Detailed Description
The invention is described in one step below with reference to the following examples and figures.
In the robust video steganography method based on the audio side channel of the embodiment, the embedding process of the secret information is shown in fig. 1, and the method includes the following steps:
1) pretreatment: and (4) shunting the accompanying audio and video to obtain a carrier video and a carrier audio. Transcoding the carrier audio into WAV format audio. And decoding the carrier video to an airspace to obtain a video frame brightness component sequence, namely a Y component sequence.
2) Encoding the secret message with BCH: encoding the secret message by using an error correcting code BCH code to generate an encoded string m of the secret message to be embedded1。
3) Performing the following operations on the Y component sequence obtained in the step 1):
3.1) selecting robust frames and generating steganographic protocols: by utilizing a pre-embedding process (figure 3), a frame with a low half error rate in a video is selected as a robust frame, other frames are selected as non-robust frames, and a mark sequence is generated to serve as a steganographic protocol P.
3.2) extracting a video embedding domain: discrete cosine transform (DWT) is performed for each pixel block to generate four subbands LL, LH, HL, and HH. Then, Singular Value Decomposition (SVD) is performed on the LL subband to obtain a diagonal matrixObtaining the maximum singular value alpha1As a video embedding domain.
3.3) quantization index modulation embedding: for maximum singular value alpha1Performing Quantization Index Modulation (QIM) to embed the string m to be embedded1。
3.4) generating a luminance component sequence: maximum singular value alpha 'after QIM modulation'1Writing back the diagonal matrix sigma and recalculating to generate a new pixel block. And merging all the modified pixel blocks, and writing the brightness frame back to the Y component sequence to obtain the video sequence after steganography.
4) Utilizing BCH coding steganography protocol: encoding the steganographic protocol P by using an error correcting code BCH code to generate an encoded string f of secret information to be embedded1。
5) Executing the following operations on the WAV format audio obtained in the step 1):
5.1) segmentation of sampling points: and segmenting the sampling point sequence in the WAV format audio according to the fixed step length to obtain a plurality of sampling segments.
5.2) extracting an audio embedded domain: and carrying out DWT conversion on each sampling point segment to generate a low-frequency subband L and a high-frequency subband H. Then, the L subband is subjected to SVD decomposition to obtain a singular value array U ═ (μ)1 μ2 μn) Obtaining the maximum singular value mu1As the audio embedded domain.
5.3) quantization index modulation embedding: for maximum singular value mu1Performing QIM modulation, and embedding a string f to be embedded with secret information1。
And 5.4) writing the maximum singular value after QIM modulation back to an array U, and recalculating to generate new sampling point segments. And combining all the modified sampling points in a segmented manner, and writing the sampling points back to the sampling point sequence to generate the WAV format audio subjected to steganography.
6) And (3) post-treatment: and (3) encoding the steganographic video sequence generated in the step (3.4) and the steganographic WAV format audio generated in the step (5.4), and packaging the combined stream into an accompanying sound steganographic video.
Further, the method further includes a secret information extraction process, as shown in fig. 2, including the following steps:
1) pretreatment: and shunting the audio and video subjected to steganography to obtain a carrier video and a carrier audio. Transcoding the carrier audio into WAV format audio. And decoding the carrier video to a space domain to obtain a video frame brightness component sequence (namely a Y component sequence).
2) Executing the following operations on the WAV format audio obtained in the step 1):
2.1) segmentation of sampling points: and segmenting the sampling point sequence in the WAV format audio according to the fixed step length to obtain a plurality of sampling segments.
2.2) extracting an audio embedded domain: and carrying out DWT conversion on each sampling point segment to generate a low-frequency subband L and a high-frequency subband H. Then, the L subband is subjected to SVD decomposition to obtain a singular value array U ═ (μ)1 μ2…μn) Obtaining the maximum singular value mu1As the audio embedded domain.
2.3) quantization index modulation extraction: using QIM method to measure the maximum singular value mu1Extracting to generate a secret string f'1。
3) Secret information string f 'is matched by using error correcting code BCH code'1Decoding is carried out to generate the steganographic protocol P.
4) Performing the following operations on the Y component sequence obtained in the step 1):
4.1) selecting robust frames: selecting a robust frame in the Y component sequence by using the steganographic protocol P obtained in the step 3).
4.2) extracting a video embedding domain: discrete cosine transform (DWT) is performed for each pixel block to generate four subbands LL, LH, HL, and HH. Then, Singular Value Decomposition (SVD) is performed on the LL subband to obtain a diagonal matrixObtaining the maximum singular value alpha1As a video embedding domain.
4.3) quantization index modulation extraction: using QIM algorithm to measure the maximum singular value alpha1Extracting to obtain secret string m'1。
5) Secret information string m 'is paired by using error correction code BCH code'1Decoding is performed to generate a secret message. And decrypting the secret message to generate a message plaintext.
The above method is described in further detail below.
Before the invention is adopted to carry out the message steganography embedding, the message to be embedded can be encrypted firstly to obtain random binary data stream. The invention provides a robust video steganography method based on an audio side channel, wherein the embedding process of a secret message can be divided into two parts: one is the pre-embedding process and the other is the secret message embedding process.
The pre-embedding process (fig. 3) generates a video frame sequence flag bit array P (i.e., steganographic protocol) to help select robust frames for secret message embedding. The specific operation details are as follows:
1) generating a random bit string s of fixed length1。
2) And decoding the carrier video to an airspace to obtain a video brightness component sequence.
3) One frame is taken out of the luminance component sequence and divided into non-overlapping blocks of pixels of size 16 × 16.
4) One block is taken from the obtained data, and DWT transformation is carried out to obtain four sub-bands of LL, LH, HL and HH. Taking LL sub-band, SVD decomposing to obtain diagonal matrix of singular value, taking maximum singular value alpha of first row and first column1。
5) Bit string s from step 1)1Taking one bit, and utilizing QIM algorithm to make singular value alpha1Is modulated to complete the embedding to generate alpha'1。
6) Alpha 'is prepared'1The diagonal matrix is written back and a new subband LL' is generated by recalculation with SVD. And carrying out DWT inverse transformation on the LL', LH, HL and HH sub-bands to generate a new pixel block.
7) And repeating the steps 4), 5) and 6) on the rest of the blocks until all the blocks of the frame are completely embedded.
8) Repeating steps 3) to 7) for the frames in the remaining luminance component sequence until all frames of the luminance sequence are embedded.
9) All luminance frames are combined into a sequence of luminance components and the entire spatial sequence is encoded into a video bitstream. And transcoding the video code stream by using a transcoder with fixed parameters to generate a transcoded video.
10) Decoding the video obtained in the step 9) to a space domain to obtain a video brightness component sequence (Y component sequence).
11) Taking a frame from the sequence of luminance components and dividing it into 16 x 16 blocks of pixels
12) One block is taken from the obtained data, and DWT transformation is carried out to obtain four sub-bands of LL, LH, HL and HH. Taking LL sub-band, SVD decomposing to obtain diagonal matrix of singular value, taking maximum singular value alpha of first row and first column1。
13) Singular value alpha using QIM algorithm1And (4) extracting.
14) And repeating the step 12) and the step 13) on the rest of the blocks until all the blocks of the frame are completely extracted.
15) Steps 11) to 14) are repeatedly performed for the frames in the remaining luminance component sequence until all frames of the luminance sequence have been extracted.
16) Merging the messages obtained in the step 15) into a bit string s'1And the original random bit string s generated in step 1)1And comparing and calculating the error rate of each frame. Selecting from videoAnd taking the frame with the low error rate as a robust frame, and taking other frames as non-robust frames. Setting the zone bit of the robust frame as 0 and the zone bit of the non-robust frame as 1, and generating a zone bit sequence as a steganographic protocol P.
The secret message embedding process includes two parts: one is the audio side channel embedding process (fig. 4) and the second is the video channel embedding process (fig. 5). The audio side channel embeds the steganographic protocol P and the video channel embeds the secret message.
The audio side channel embedding steganography protocol, the specific operation details of which are shown in fig. 4, includes the following steps:
1) an AAC audio stream separated from video is transcoded into WAV format and a sequence of sample points therein is extracted.
2) Encoding the steganographic protocol P by using BCH (63,7,15) code to generate a string f of information to be embedded1。
3) The sequence of sample points is partitioned into a series of fixed-length audio segments.
4) One audio segment is taken from the audio segment, and DWT conversion is carried out on the audio segment to obtain a low-frequency subband L and a high-frequency subband H. Taking the low-frequency sub-band L to carry out SVD (singular value decomposition) to obtain the maximum singular value mu1。
5) From the string f of secret information to be embedded1Taking one bit, and using QIM algorithm to obtain singular value mu1Modulation is performed to complete message bit embedding, and mu 'is generated'1。
6) Prepared from mu'1Write back, re-compute with SVD to generate new low frequency subband L'. And carrying out DWT inverse transformation on the low-frequency sub-band L' and the high-frequency sub-band H to generate a new audio segment.
7) Carrying out steps 4) to 6) on the rest sampling points in a segmented manner until a string f to be embedded with secret information is formed1The embedding is completed completely.
8) And splicing the audio segments subjected to steganography into a complete audio sampling point sequence to generate a WAV format audio file subjected to steganography.
The process of embedding the secret message in the video channel is similar to the pre-embedding process, and the main difference is that only the robust frame is selected as a carrier for embedding, and other non-robust frames are not processed. The specific operation details are shown in fig. 5, and the method comprises the following steps:
1) and decoding the carrier video to an airspace to obtain a video brightness component sequence.
2) The secret message is encoded by using BCH (63.7,15) code to generate a string m of the secret message to be embedded1。
3) A frame of robust frames is taken from the sequence of luminance components and divided into non-overlapping blocks of size 16 x 16.
4) One block is taken from the obtained data, and DWT transformation is carried out to obtain four sub-bands of LL, LH, HL and HH. Taking LL sub-band, SVD decomposing to obtain diagonal matrix of singular value, taking maximum singular value alpha of first row and first column1。
5) From the string m of secret information to be embedded1Taking one bit, and utilizing QIM algorithm to make singular value alpha1Is modulated to complete the embedding to generate alpha'1。
6) Alpha 'is prepared'1The diagonal matrix is written back and a new subband LL' is generated by recalculation with SVD. And carrying out DWT inverse transformation on the LL', LH, HL and HH sub-bands to generate a new pixel block.
7) Repeating the steps 4), 5) and 6) for the rest of the blocks until all the blocks of the frame are embedded.
8) Repeating the steps 3) to 7) on the robust frame in the rest luminance component sequence until the string m to be embedded is finished1Is embedded.
9) And writing all the frames back to the brightness component sequence, and coding the video spatial domain sequence into a video code stream.
10) And (3) encoding the encrypted WAV format audio stream embedded with the steganographic protocol and the video stream generated in the step 6), and packaging the combined stream into an accompanying audio steganographic video.
The invention provides a robust video steganography method based on an audio side channel, wherein the extraction process of secret information can also be divided into two parts: one is the steganographic protocol extraction process (fig. 6) and the other is the secret message extraction process (fig. 7).
In the process of covert communication, a receiving end cannot distinguish a robust frame from a non-robust frame when extracting a secret message. Therefore, before extracting the secret message, the receiving end needs to extract the steganographic protocol, and the specific operation details thereof are shown in fig. 6, which includes the following steps:
1) an AAC audio stream separated from video is transcoded into WAV format and a sequence of sample points therein is extracted.
2) The sampling point is divided into a series of audio segments with fixed length according to fixed step length.
3) One audio segment is taken from the audio segment, and DWT conversion is carried out on the audio segment to obtain a low-frequency subband L and a high-frequency subband H. Taking the low-frequency sub-band L to carry out SVD (singular value decomposition) to obtain the maximum singular value mu1。
4) Singular value mu by using QIM algorithm1And extracting to obtain a secret information bit.
5) And repeating the step 3) and the step 4) on the rest audio segments until all secret information bits are completely extracted.
6) Utilizing BCH (63,7,15) codes to match extracted secret information strings f'1Decoding is carried out to generate the steganographic protocol P.
After extracting the steganographic protocol, the receiving end extracts the secret message only from the robust frame in the video luminance sequence, and the specific operation details thereof are shown in fig. 7, which includes the following steps:
1) and decoding the steganographic video to a space domain to obtain a video brightness component sequence.
2) According to the steganographic protocol mark, a frame of robust frame is taken from the brightness component sequence and is divided into pixel blocks which are 16 multiplied by 16 and do not overlap with each other.
3) One block is taken from the obtained data, and DWT transformation is carried out to obtain four sub-bands of LL, LH, HL and HH. Taking LL subband, andSVD is carried out to obtain a singular value diagonal matrix, and the maximum singular value alpha of the first row and the first column is taken1。
4) Singular value alpha using QIM algorithm1And extracting to obtain a secret information bit.
5) And repeating the step 3) and the step 4) on the rest of the blocks until all the blocks of the frame finish extraction of the secret information bits.
6) And repeating the step 2) to the step 5) on the robust frame in the rest luminance component sequence until all the secret information bits are completely extracted.
7) Using BCH (63,7,15) code to extract secret information string m'1Decoding is performed to generate a secret message. And decrypting the secret message to generate a message plaintext.
It can be seen from the above specific embodiments that, in the embedding process of the present invention, a robust embedding area can be selected for steganography according to the specific situation of each carrier video, so that the robustness of the steganography algorithm is improved. Meanwhile, the invention introduces an error correction coding technology into the algorithm, improves the error correction capability through the BCH code, and reduces the error rate. The invention is reasonably designed from three aspects of algorithm, embedded area and error correcting code, and the robustness is greatly improved.
The method can effectively resist the lossy transcoding operation of the social network, and can be used for covert communication based on the social network. In order to verify the effectiveness of the present invention in covert communication based on social network, 10 videos with resolution of 720p (1280 × 720) were selected and tested for load rate and bit error rate on YouTube channel, as shown in table 1.
TABLE 1 robustness evaluation results of the present invention in the YouTube channel
Based on the same inventive concept, another embodiment of the present invention provides a robust video steganography apparatus based on an audio side channel, comprising a secret information embedding unit and/or a secret information extraction unit; the secret information embedding unit embeds the secret information by adopting the embedding process of the secret information in the method to generate the sound steganography video; the secret information extraction unit extracts the secret information from the sound steganography video by adopting the secret information extraction process in the method.
Based on the same inventive concept, another embodiment of the present invention provides an electronic device (computer, server, smartphone, etc.) comprising a memory storing a computer program configured to be executed by the processor, and a processor, the computer program comprising instructions for performing the steps of the inventive method.
Based on the same inventive concept, another embodiment of the present invention provides a computer-readable storage medium (e.g., ROM/RAM, magnetic disk, optical disk) storing a computer program, which when executed by a computer, performs the steps of the inventive method.
The above examples are only for illustrating the technical solutions of the present invention and not for limiting the same, and those skilled in the art can make modifications or equivalent substitutions to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and the scope of the present invention should be determined by the claims.
Claims (10)
1. A robust video steganography method based on an audio side channel comprises an embedding process of secret information, and is characterized in that the embedding process of the secret information comprises the following steps:
shunting the accompanying audio and video to obtain a carrier video and a carrier audio;
embedding a secret message in a carrier video to obtain a steganographic video;
embedding a steganography protocol in the carrier audio to obtain a steganographically-generated audio;
and merging and packaging the video subjected to steganography and the audio subjected to steganography into an accompanying sound steganography video.
2. The method of claim 1, wherein splitting audio video to obtain carrier video and carrier audio comprises: transcoding the carrier audio into WAV format audio, decoding the carrier video to an airspace, and acquiring a video frame brightness component sequence, namely a Y component sequence.
3. The method of claim 2, wherein embedding the secret message in the carrier video to obtain the steganographically written video comprises:
encoding the secret message with BCH: encoding the secret message by using an error correcting code BCH code to generate an encoded string m of the secret message to be embedded1;
For the Y component sequence, performing the following operations:
1) selecting a robust frame and generating a steganographic protocol: selecting a frame with a low half error rate in a video as a robust frame and other frames as non-robust frames by utilizing a pre-embedding process, and generating a mark sequence as a steganography protocol P;
2) extracting a video embedding domain: discrete cosine transform is carried out on each pixel block to generate four sub-bands of LL, LH, HL and HH, and then Singular Value Decomposition (SVD) is carried out on the LL sub-band to obtain a diagonal matrixObtaining the maximum singular value alpha1As a video embedding domain;
3) quantization index modulation embedding: for maximum singular value alpha1Carrying out quantization index modulation and embedding a string m of information to be embedded1;
4) Generating a sequence of luminance components: maximum singular value alpha 'after QIM modulation'iWriting back the diagonal matrix sigma, recalculating to generate a new pixel block, merging all the modified pixel blocks, and writing the brightness frame back to the Y component sequence, thereby obtaining the video subjected to steganography.
4. The method of claim 3, wherein the pre-embedding process of step 1) comprises:
1.1) generating a random bit string s of fixed length1;
1.2) decoding the carrier video to an airspace to obtain a video brightness component sequence;
1.3) taking one frame from the luminance component sequence and dividing the frame into non-overlapping pixel blocks with the size of 16 multiplied by 16;
1.4) taking one block from the two blocks, carrying out DWT (discrete wavelet transform) to obtain four sub-bands of LL, LH, HL and HH, taking the LL sub-band, carrying out SVD (singular value decomposition) to obtain a singular value diagonal matrix, and taking the maximum singular value alpha of the first row and the first column of the first row1;
1.5) bit string s from step 1)1Taking one bit, and utilizing QIM algorithm to make singular value alpha1Is modulated to complete the embedding to generate alpha'1;
1.6) mixing alpha'1Writing back the diagonal matrix, recalculating by using SVD to generate a new sub-band LL ', and carrying out DWT inverse transformation on the LL', LH, HL and HH sub-bands to generate a new pixel block;
1.7) repeatedly executing the step 1.4), the step 1.5) and the step 1.6) on the rest blocks till all the blocks of the frame are completely embedded;
1.8) repeating the steps 1.3) to 1.7) on the frames in the rest luminance component sequence until all the frames in the luminance sequence are embedded;
1.9) combining all the brightness frames into a brightness component sequence, coding the whole airspace sequence into a video code stream, and transcoding the video code stream by using a transcoder with fixed parameters to generate a transcoded video;
1.10) decoding the video obtained in the step 1.9) to an airspace to obtain a video brightness component sequence;
1.11) taking a frame from the luminance component sequence and dividing it into 16 × 16 pixel blocks;
1.12) taking one block from the two blocks, carrying out DWT (discrete wavelet transform) to obtain four sub-bands of LL, LH, HL and HH, taking the LL sub-band, carrying out SVD (singular value decomposition) to obtain a singular value diagonal matrix, and taking the maximum singular value alpha of the first row and the first column of the first row1;
1.13) singular value α using QIM algorithm1Carrying out extraction;
1.14) repeatedly executing the step 1.12) and the step 1.13) on the rest blocks until all the blocks of the frame are completely extracted;
1.15) repeating the steps 1.11) to 1.14) on the frames in the rest luminance component sequence until all the frames in the luminance sequence are completely extracted;
1.16) merging the messages obtained in step 1.15) into a bit string s'1And the original random bit string s generated in step 1.1)1Comparing, and calculating the error rate of each frame; selecting from videoTaking the frame with lower error rate as a robust frame, and taking other frames as non-robust frames; setting the zone bit of the robust frame as 0 and the zone bit of the non-robust frame as 1, and generating a zone bit sequence as a steganographic protocol P.
5. The method according to claim 3 or 4, wherein embedding a steganographic protocol in the carrier audio to obtain steganographically written audio comprises:
utilizing BCH coding steganography protocol: encoding the steganographic protocol P by using an error correcting code BCH code to generate an encoded string f of secret information to be embedded1;
For the WAV format audio, the following operations are executed:
1) sampling point segmentation: segmenting a sampling point sequence in the WAV format audio according to a fixed step length to obtain a plurality of sampling segments;
2) extracting an audio embedded domain: carrying out DWT conversion on each sampling point segment to generate a low-frequency sub-band L and a high-frequency sub-band H; then, the L subband is subjected to SVD decomposition to obtain a singular value array U ═ (μ)1 μ2 … μn) Obtaining the maximum singular value mu1As an audio embedded domain;
3) quantization index modulation embedding: for maximum singular value mu1Performing QIM modulation, and embedding a string f to be embedded with secret information1;
4) Writing the maximum singular value after QIM modulation back to the array U, recalculating to generate new sampling point segments, combining all modified sampling point segments, writing the sampling point segments back to the sampling point sequence, and generating the WAV format audio after steganography.
6. A robust video steganography method based on an audio side channel comprises a secret information extraction process, and is characterized in that the secret information extraction process extracts secret information from sound-accompanying steganography video generated by the method of any one of claims 1-5; the secret information extraction process comprises the following steps:
shunting the audio steganography video to obtain a carrier video and a carrier audio;
the steganographic protocol is extracted from the carrier audio and then the robust region is selected and the secret information is extracted from the carrier video.
7. The method according to claim 6, wherein the extracting process of the secret information comprises the steps of:
1) pretreatment: shunting the audio and video subjected to steganography to obtain a carrier video and a carrier audio, transcoding the carrier audio into WAV format audio, decoding the carrier video to an airspace, and acquiring a video frame brightness component sequence, namely a Y component sequence;
2) executing the following operations on the WAV format audio obtained in the step 1):
2.1) segmentation of sampling points: segmenting a sampling point sequence in the WAV format audio according to a fixed step length to obtain a plurality of sampling segments;
2.2) extracting an audio embedded domain: carrying out DWT conversion on each sampling point segment to generate a low-frequency subband L and a high-frequency subband H, and then carrying out SVD (singular value decomposition) on the L subband to obtain a singular value array U ═ (mu)1 μ2 … μn) Obtaining the maximum singular value mu1As an audio embedded domain;
2.3) quantization index modulation extraction: using QIM method to measure the maximum singular value mu1Extracting to generate a secret string f'1;
3) Secret information string f 'is matched by using error correcting code BCH code'1Decoding to generate a steganographic protocol P;
4) performing the following operations on the Y component sequence obtained in the step 1):
4.1) selecting robust frames: selecting a robust frame in the Y component sequence by using the steganographic protocol P obtained in the step 3);
4.2) extracting a video embedding domain: discrete cosine transform (DWT) is carried out on each pixel block to generate four sub-bands of LL, LH, HL and HH, and then Singular Value Decomposition (SVD) is carried out on the LL sub-band to obtain a diagonal matrixObtaining the maximum singular value alpha1As a video embedding domain;
4.3) quantization index modulation extraction: using QIM algorithm to measure the maximum singular value alpha1Extracting to obtain secret string m'1;
5) Secret information string m 'is paired by using error correction code BCH code'1Decoding is carried out to generate a secret message, and the secret message is decrypted to generate a message plaintext.
8. A robust video steganography device based on an audio side channel is characterized by comprising a secret information embedding unit and/or a secret information extracting unit; the secret information embedding unit embeds the secret information by adopting the method of any one of claims 1 to 5 to generate the sound steganography video; the secret information extraction unit adopts the method of claim 6 or 7 to extract secret information from the audio steganographic video.
9. An electronic apparatus, comprising a memory and a processor, the memory storing a computer program configured to be executed by the processor, the computer program comprising instructions for performing the method of any of claims 1 to 7.
10. A computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program which, when executed by a computer, implements the method of any one of claims 1 to 7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011388479.2A CN112714230A (en) | 2020-12-01 | 2020-12-01 | Robust video steganography method and device based on audio side channel |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011388479.2A CN112714230A (en) | 2020-12-01 | 2020-12-01 | Robust video steganography method and device based on audio side channel |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112714230A true CN112714230A (en) | 2021-04-27 |
Family
ID=75542465
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202011388479.2A Pending CN112714230A (en) | 2020-12-01 | 2020-12-01 | Robust video steganography method and device based on audio side channel |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112714230A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115457985A (en) * | 2022-09-15 | 2022-12-09 | 北京邮电大学 | Visual audio steganography method based on convolutional neural network |
CN116506232A (en) * | 2023-06-28 | 2023-07-28 | 南京畅洋科技有限公司 | High-capacity Internet of things hidden channel construction method based on channel coding |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104010193A (en) * | 2014-05-29 | 2014-08-27 | 中国科学院信息工程研究所 | Video steganographic method based on macro block partition mode disturbance |
CN104125467A (en) * | 2014-08-01 | 2014-10-29 | 郑州师范学院 | Embedding and extracting methods for video steganography information |
-
2020
- 2020-12-01 CN CN202011388479.2A patent/CN112714230A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104010193A (en) * | 2014-05-29 | 2014-08-27 | 中国科学院信息工程研究所 | Video steganographic method based on macro block partition mode disturbance |
CN104125467A (en) * | 2014-08-01 | 2014-10-29 | 郑州师范学院 | Embedding and extracting methods for video steganography information |
Non-Patent Citations (3)
Title |
---|
A.R.ELSHAZLY: "Synhronized double watermark audio watermarking scheme based on a transform domain for stereo signals", 《2016 FOURTH INTERNATIONAL JAPAN-EGYPT CONFERENCE ON ELECTRONICS, COMMUNICATIONS AND COMPUTERS (JEC-ECC)》 * |
PINGAN FAN,ET AL.: "A robust video steganographic method against social networking transcoding based on steganographic side channel", 《PROCEEDINGS OF THE 2020 ACM WORKSHOP ON INFORMATION HIDING AND MULTIMEDIA SECURITY》 * |
Y. KAKDE, ET AL.: "Audio-Video steganography", 《2015 INTERNATIONAL CONFERENCE ON INNOVATIONS IN INFORMATION, EMBEDDED AND COMMUNICATION SYSTEMS (ICIIECS)》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115457985A (en) * | 2022-09-15 | 2022-12-09 | 北京邮电大学 | Visual audio steganography method based on convolutional neural network |
CN115457985B (en) * | 2022-09-15 | 2023-04-07 | 北京邮电大学 | Visual audio steganography method based on convolutional neural network |
CN116506232A (en) * | 2023-06-28 | 2023-07-28 | 南京畅洋科技有限公司 | High-capacity Internet of things hidden channel construction method based on channel coding |
CN116506232B (en) * | 2023-06-28 | 2023-10-10 | 南京畅洋科技有限公司 | High-capacity Internet of things hidden channel construction method based on channel coding |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Tao et al. | Towards robust image steganography | |
Su et al. | Steganography in JPEG2000 compressed images | |
US8032758B2 (en) | Content authentication and recovery using digital watermarks | |
Sun et al. | High performance reversible data hiding for block truncation coding compressed images | |
CN104581176B (en) | The insertion of H.264/AVC compression domain robust video watermark and extracting method without frame in error drift | |
Alavianmehr et al. | A lossless data hiding scheme on video raw data robust against H. 264/AVC compression | |
JP2004163855A (en) | Electronic watermark embedding method, and encoder and decoder capable of utilizing such method | |
CN111640444A (en) | CNN-based self-adaptive audio steganography method and secret information extraction method | |
CN108965887B (en) | Video information hiding method and device based on inter-block decoupling | |
Hong et al. | Lossless steganography for AMBTC-compressed images | |
CN112714230A (en) | Robust video steganography method and device based on audio side channel | |
CN109819260B (en) | Video steganography method and device based on multi-embedded domain fusion | |
CN1886991A (en) | Method and apparatus for encoding or decoding a bitstream | |
Ramkumar et al. | Uncompressed digital video watermarking using stationary wavelet transform | |
CN112714231A (en) | Robust steganography method based on DCT (discrete cosine transformation) symbol replacement | |
CN105915912A (en) | Fragile watermark embedding and detection method of MPEG-2 compressed domain video | |
Kumar et al. | An improved data-hiding approach using skin-tone detection for video steganography | |
Mandal et al. | A genetic algorithm based steganography in frequency domain (GASFD) | |
Tran et al. | Lsb data hiding in digital media: a survey | |
Xu et al. | JPEG compression immune steganography using wavelet transform | |
CN114598887A (en) | Anti-recompression video watermarking method for controlling bit rate increase | |
CN115767105A (en) | Robust video steganography method and device based on adaptive quantization | |
Ma et al. | Large capacity and high quality reversible data hiding method based on enhanced side match vector quantization | |
Izadinia et al. | A new steganographic method using quantization index modulation | |
Kumar et al. | A Resilient Approach to Robust JPEG Steganography: Ensuring Hidden Data Integrity |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210427 |