CN100496118C - A Encryption and Decryption Method of Residual Coefficient Amplitude in H.264 Standard Video Information - Google Patents

Abstract

The invention relates to a kind of encryption and deciphering method of remained difference coefficient breadth in H. 264 standard video information, it characterized in that: based on self adapting updating rule of remained difference coefficient breadth coding middle variable suffiXLength, adopting difference value absolute value which calculates current remained difference coefficient breadth and suffixLength self adapting updating threshold value, then based on the difference value absolute value, deciding bit number of last bit of the remained difference coefficient breadth which is to be encrypted, video code flow encrypted fulfill the H. 264 standard format. The characteristics of the invention are following: independent on video coder and decoder, video code flow encrypted fulfill the H. 264 standard; complexity of calculation is low, can fulfill the real time requirement of video information process and transmission; not affect the compression ratio; not affect video communication and network fault tolerant of video transmission; secrecy class can be realized; independent on concrete cryptogram arithmetic, it can use different international normal cryptogram arithmetic, the secrecy can achieve level that current technique can not decrypt.

Description

A Encryption and Decryption Method of Residual Coefficient Amplitude in H.264 Standard Video Information

technical field

The invention belongs to the field of video security communication, and specifically relates to an encryption and decryption method for residual coefficient amplitude in H.264 standard video information.

Background technique

With the wide application of various video, the security problem of video information is also becoming more and more prominent. Source encryption, that is, the encryption of the video content itself, is the main and more important means to solve the security of video information. The encryption of the video content itself can ensure that even if the illegal user obtains the video information, he cannot correctly decrypt and watch the original correct video image without the key.

Because video information has massive data, in specific applications, such as in videophone or video conferencing applications, dozens of images must be transmitted per second. If a universal encryption algorithm for general information (binary or text) is used to encrypt all Encrypting video information requires a huge amount of calculation, which will generate a large additional delay, and it is difficult to meet the real-time requirements of video communication.

In addition to confidentiality and real-time performance, maintaining format compatibility is an important requirement for video encryption algorithms. The meaning of format compatibility is that the syntax and semantics of video ciphertext information is still compatible with the standard, and its core performance is that video ciphertext information can still be successfully decoded by standard decoders. The importance of maintaining format compatibility lies primarily in the following points:

1) Keeping the video format compatible with the standard can ensure that channel information such as network synchronization and error tolerance fields in the original video plaintext will not be changed, thereby ensuring the network adaptability and error tolerance of the ciphertext code stream.

2) Keeping format information such as video syntax and semantics compatible with standards also keeps fixed fields such as various initial codeword format information in the video stream unchanged. These fields are widely known because they are disclosed in protocols or standards. If these fields are encrypted, known plaintext attacks will be caused, which is not conducive to the security of video information.

3) Keeping the video format compatible with the standard is conducive to the effective and flexible control of the video information content for commercial applications. In commercial applications such as video-on-demand and pay TV, in order to attract users to purchase, it is often necessary to allow unauthorized users to successfully decode and play video ciphertext code streams to watch certain video content.

4) Keeping the video format unchanged is also conducive to the normal application of standard coded video information for storage, because keeping the format compatible can keep all information or attributes except the video content unchanged, so that the video information can still be retrieved and is conducive to maintaining Interactive functions such as fast forward and rewind of video information.

Since the digital video encryption technology appeared in the mid-1990s, it has been constantly developing along with the development of video compression coding standards. The video encryption technology based on DCT transformation and predictive coding hybrid system and Huffman variable length codec (H.261/H.263 and MPEG-1/MPEG-2/MPEG-4) has produced many research results. The most mainstream method is to select the key information of the video image content in the video information: residual coefficient, motion vector and quantization step size, etc. for selective encryption. With the rapid development of hardware performance, H.264/AVC will replace the previous series of standards due to its good compression efficiency, image quality and network fault tolerance. However, since H.264/AVC uses intra-frame prediction, context-based adaptive variable length coding (CAVLC) and many other codec technologies that are completely different from the past, most of the previous video encryption research results are not applicable to H.264 Encryption of /AVC standard encoded video.

For the new intra-frame prediction mode words and I_PCM mode macroblocks in the H.264 code stream, and the motion vectors that have been converted to Exp-Golomb entropy coding, some basic ideas of existing video encryption achievements that maintain format compatibility can still be used for reference. , such as selectively encrypting it. However, the residual coefficient in H.264 adopts the context-based coding strategy (CAVLC), and the encoded residual coefficient amplitude codeword context is closely related, directly encrypting it or selecting a fixed part of bit encryption will definitely affect the context The smooth decoding of other residual coefficient magnitudes will inevitably destroy its format compatibility. However, the current video encryption research results that maintain format compatibility are usually based on Huffman entropy coding, which is not suitable for H.264/AVC. Therefore, how to encrypt the residual coefficient amplitude in the H.264 bit stream while maintaining format compatibility is a problem. a difficulty.

For the H.264 standard, when the luma and chrominance block residual data is coded using CAVLC, the residual coefficients are composed of trailing coefficient codewords, non-zero coefficient amplitude prefix level_prefix and suffix level_suffix codewords. The trailing coefficients indicate the signs of the last few trailing ±1 coefficients of the residual block, and the number of trailing coefficients is indicated by the initial code word coeff_token of each residual block, so the trailing coefficients can be directly calculated according to their number Encrypt it. That is, when the number of tailing coefficients of coeff_token is k in decoding analysis, the following k bits can be directly encrypted. The amplitude level encoding and decoding of non-zero coefficients is more complicated. The level_prefix code word is coded and decoded according to the variable-length code table specified in the protocol, and direct encryption of the code word will inevitably destroy the code stream compatibility. suffixLength and levelCode are variables that need to be used when encoding. The suffix level_suffix is an unsigned integer whose length is levelSuffixSize bits. Normally, the value of levelSuffixSize is equal to the value of variable suffixLength, but there are two exceptions: 1. When the prefix is equal to 14 and suffixLength is equal to 0, levelSuffixSize is equal to 4; 2. When prefix is equal to 15, levelSuffixSize is equal to 12.

The variable suffixLength is adaptively updated based on context, and the update of suffixLength is related to the value of the current suffixLength and the value (level) of the decoded non-zero coefficient. After suffixLength is initialized, the adaptive update rule is: if the current decoded non-zero coefficient value is greater than (3<<(suffixLength-1)), and the current suffixLength is less than 6, suffixLength is added by 1. ("<<", ">>" points are expressed as binary left shift and right shift operation symbols)

The variable levelCode is equal to (level_prefix<<suffixLength)+level_suffix. Then the level value is obtained according to the following rules: if the levelCode is an even number, the value of level is (levelCode+2)>>1; otherwise, the value of level is (-level-1)>>1. Therefore, a small change to the levelCode will cause a change in the sign of the level value. Then a small change to the level_suffix may cause a large change in the level value, so that the encryption of the level_suffix can be considered to achieve the purpose of encrypting the residual coefficient.

However, the entire level_suffix cannot be directly encrypted, because the change of level_suffix will cause the change of the adaptive update of suffixLength, resulting in an error in judging the length of the codeword level_suffix when decoding the next residual coefficient amplitude codeword, that is, destroying If the semantic format of the code stream is changed, it will inevitably cause decoding failure. Therefore, the encryption of level_suffix needs to consider the adaptive update process that cannot affect suffixLength.

Contents of the invention

The purpose of the present invention is to directly encrypt the residual coefficient amplitude in the entire H.264 code stream or encrypt its fixed part of the bit will inevitably destroy the problem of format compatibility, by carefully analyzing the residual in the H.264 code stream based on CAVLC entropy Based on the encoding and decoding process and context-related characteristics of coefficients, an encryption and decryption method for the magnitude of residual coefficients in the corresponding H.264 standard video information is proposed, and the encryption is maintained while maintaining format compatibility, so as to overcome the above shortcomings.

In order to ensure that the encrypted residual coefficient amplitude does not destroy the semantic format, the present invention dynamically analyzes the code stream of each coded slice of H.264, and adopts an adaptive encryption method according to the adaptive update rule of suffixLength when encoding the residual coefficient amplitude. The specific adaptive rules are as follows:

(1) When the prefix is equal to 14 and the suffixLength is equal to 0, the 4-bit level_suffix is directly encrypted.

(2) When the prefix is equal to 15, directly encrypt the 12-bit level_suffix.

(3) When suffixLength is greater than or equal to 6, suffixLength is no longer updated, and level_suffix is directly encrypted.

用hread来表示，当前level值和thread差值的绝对值Abs(level-thread)用abs_diff表示)，当abs_diff大于(1<<k)时，可以加密suffixLength的后k个bit.这样可以保证加密不会造成影响level和阈值的关系，从而影响suffixLength自适应更新的判断，从而保证语义格式兼容。(4) When none of the above three special conditions are met, (assuming suffixLength adaptive update threshold 3<<(suffixLength

Expressed by hread, the absolute value of the difference between the current level value and thread Abs (level-thread) is represented by abs_diff), when abs_diff is greater than (1<<k), the last k bits of suffixLength can be encrypted. This can ensure encryption It will not affect the relationship between level and threshold, thereby affecting the judgment of suffixLength adaptive update, so as to ensure that the semantic format is compatible.

In order to achieve the above object, the method of the present invention is: first need to calculate the absolute value of the difference between the current level value and the thread, and then determine how many bits to encrypt according to the size of the absolute value of the difference. Based on the above four situations, the method for encrypting the magnitude of residual coefficients proposed by the present invention is an adaptive process that requires dynamic analysis to determine how to encrypt.

In addition, the present invention proposes to combine the encrypted residual coefficient magnitude, encrypted motion vector, intra-frame prediction mode word, quantization coefficient, residual coefficient trailing coefficient part and I_PCM macroblock pixel information to form a complete scheme to realize video information confidentiality.

Use cryptographic algorithms to directly encrypt the 3-bit prediction mode word of the 4×4 luma block, the trailing part of the coefficient of the residual block and the pixel value of the I_PCM macroblock.

The last bit of the 16x16 luma block and chroma block intra prediction mode word is encrypted.

The suffix part of the Exp-Golomb code word is encrypted for the quantization step size and the difference value of the motion vector. As for the magnitude of the residual coefficient, the bit length of the level_suffix part that can be encrypted is dynamically determined according to the adaptive method proposed above.

The purpose of this partial bit encryption method is to keep the video code stream conforming to the H.264 standard format after encryption. The specific encryption method above is determined after in-depth analysis of the syntax, semantics and decoding process specified in the H.264 standard. Therefore, encryption according to the above method according to the H.264 standard can ensure that the format remains unchanged.

When implementing it, analyze the H.264 video code stream, extract the data bits to be encrypted according to the above rules, and flexibly select various high-strength cryptographic algorithms (such as block cipher algorithm AES/IDEA or sequence cipher algorithm A5/RC4, etc.) Encrypt, and then restore the ciphertext of the encrypted key data to the extraction location and mix the unencrypted plaintext as the encrypted video stream output, and decryption is the opposite process.

Because the present invention is based on the self-adaptive update rule of suffixLength when residual coefficient magnitude is encoded, has adopted corresponding self-adaptive encryption method, makes the present invention have the following advantages: 1, the video information that conforms to the H.264 standard is obtained after being encrypted and still conforms to the standard 2. The amount of encrypted and decrypted data is small, and the calculation complexity is low, which can ensure the real-time transmission of video information; 3. The residual coefficient contains a large amount of video image content information, which is encrypted ( (combined with encrypting other information) can significantly improve the confidentiality of video information; 4. It can provide users with a variety of choices for different cryptographic algorithms, and the cryptographic algorithms are easy to update and upgrade.

Description of drawings

Fig. 1 is a schematic diagram of the encryption residual coefficient amplitude method of the present invention.

Fig. 2 is a functional block diagram of the present invention.

Fig. 3 is an overall flow chart of the application of the present invention in a normal state.

Fig. 4 is an overall flowchart of the application of the present invention in an abnormal state.

Fig. 5 is a flow chart of the encryption implementation method of the present invention.

Fig. 6 is a flow chart of the decryption implementation method of the present invention.

Detailed ways

The present invention will be further described below in conjunction with the accompanying drawings and examples, but the following examples should not be construed as limiting the present invention.

The application process of this method is shown in Figure 3.

The video encryption module in Fig. 1 is the realization part of this method. The compressed original video code stream passes through the video encryption module to obtain a standard video code stream, and the encrypted code stream is restored to the original video code stream through the decryption module, and then the correct video image can be obtained through the decoder. However, if the code stream after encryption does not pass through the decryption module or does not obtain the correct decryption key, a damaged video image will be obtained after passing through the decoder (Figure 4).

The encryption process of the present invention is shown in Fig. 5 . The video encryption module of the method is mainly composed of three parts: a code stream analysis module, a cryptographic algorithm encryption module and a code stream mixing module. The main function of the code stream analysis module is to read and analyze the compressed original video code stream, extract the defined key data and combine them together, and record the position and length information of each key data in the code stream (indicated by bit is the unit). Cryptographic Algorithm The main function of the encryption module is to encrypt the extracted key data using the existing recognized high-strength cryptographic algorithm. The main function of the code stream mixing module is to return the encrypted ciphertext to the original code stream according to the position and length information records of each key data to replace the key data extracted from the code stream before encryption, that is, with other non-key data The data is mixed to obtain the encrypted ciphertext code stream.

The decryption process of the present invention is shown in Fig. 6 . The internal structure of the video decryption module of the method is mainly composed of three parts: a code stream analysis module, a cryptographic algorithm decryption module and a code stream mixing module. The main function of the code stream analysis module is to read and analyze the encrypted video code stream, extract the ciphertext of the encrypted key data and combine them together, and record the ciphertext of each key data in the code stream The position and length information (in bits). The main function of the cryptographic algorithm decryption module is to decrypt the ciphertext of the extracted key data using the same cryptographic algorithm as that used for encryption. The main function of the code stream mixing module is to return the key data plaintext restored after decryption to the code stream according to the position and length information records of each key data ciphertext to replace the ciphertext data extracted before decryption, that is, to replace the ciphertext data extracted before decryption. The ciphertext data is mixed to obtain the decrypted plaintext code stream.

The cipher algorithm of the present invention can choose block cipher algorithm (3DES, AES, IDEA etc.) or serial cipher algorithm (A5, RC4, SEAL etc.), but not limited to above-mentioned algorithm, also can select some kind of user-designated according to the needs of user. special algorithm. The cryptographic algorithm can be designed as an independent software or hardware interface, which can be flexibly selected and exchanged in the encryption and decryption modules.

The present invention obtains encryption methods of different confidentiality levels by encrypting different combinations of key information in H.264, so as to adapt to application requirements of different levels. These key information include: residual coefficient magnitude, motion vector, intra-frame prediction mode word, quantization coefficient, residual coefficient trailing coefficient part and I_PCM macroblock pixel information.

The normal application of video information needs to focus on the following five aspects. The video encryption and decryption method proposed by the present invention is devoted to satisfying the actual needs as far as possible without affecting the application, so the impact on these five aspects is very small or even basically has no impact.

Impact on video interactivity and direct operability: the encryption method proposed by the present invention basically does not affect H.264 interaction because it does not change data such as flag information about objects and flag information about fast forward and rewind in the video information. and direct operability (e.g. video transcoding);

Impact on video encoding and decoding and video quality: The encryption and decryption method proposed by the present invention has an encryption and decryption process independent of the encoding and decoding process, and since encryption and decryption only operate on the most basic video information, it will not limit various encoding and decoding methods. characteristics, so there will be no loss of video quality;

The impact on the size of the code stream, that is, the bandwidth requirement: the encryption method proposed by the present invention does not affect the video compression ratio because it does not use a cryptographic algorithm that increases the size of the plaintext, so the size of the code stream does not increase substantially after encryption;

Impact on video transmission delay: the encryption and decryption method proposed in the present invention only encrypts a small amount of data, and the encryption and decryption calculation complexity is low, so the additional delay of the encryption method for video transmission is very small, and basically does not affect real-time performance;

Impact on network fault tolerance: the encryption and decryption method proposed by the present invention takes H.264 coded slices as a unit, when packet loss occurs in the network, the loss of one or several coded slices will not affect the correct decryption of subsequent coded slices, so this The encryption method has little effect on the fault tolerance of the network.

The content not described in detail in this specification belongs to the prior art known to those skilled in the art.

Claims (6)

1. An encryption and decryption method of residual coefficient magnitude in H.264 standard video information, characterized in that: according to the adaptive update rule of intermediate variable suffixLength when encoding residual coefficient magnitude, adopting to calculate current residual coefficient magnitude first value and suffixLength adaptive update threshold difference absolute value, and then according to the size of the absolute value of the difference, to determine the number of bits of the last few bits of the residual coefficient magnitude to be encrypted, the residual coefficient magnitude after encryption The change of does not affect the association relationship of the context, so that the encrypted video code stream still complies with the H.264 standard format. 2. The method for encrypting and decrypting the magnitude of the residual coefficient in the H.264 standard video information as claimed in claim 1, characterized in that: when encrypting, the magnitude of the encrypted residual coefficient is combined with encrypting other information containing video content Together they form a complete encryption scheme to ensure the confidentiality of H.264 standard video information. 3. The method for encrypting and decrypting residual coefficient magnitudes in H.264 standard video information as claimed in claim 2, wherein the information containing video content includes motion vectors, intra-frame prediction mode words, quantization step size, residual Difference coefficient trailing coefficient part and I_PCM macroblock pixel information. 4. The method for encrypting and decrypting the magnitude of the residual coefficient in the H.264 standard video information according to claim 2, characterized in that: for the 3-bit prediction mode word of the 4×4 luminance block, the trailing part of the coefficient of the residual block and the pixel values of the I_PCM macroblock are directly encrypted using a cryptographic algorithm. 5. The method for encrypting and decrypting residual coefficient magnitudes in H.264 standard video information according to claim 2, characterized in that the last bit of the intra-frame prediction mode words for 16x16 luma blocks and chrominance blocks is encrypted. 6. The method for encrypting and decrypting the magnitude of residual coefficients in H.264 standard video information according to claim 2, characterized in that the suffix part of the Exp-Golomb codeword is encrypted for the quantization step size and the difference value of the motion vector.

Images