CN104320657B - The predicting mode selecting method of HEVC lossless video encodings and corresponding coding method - Google Patents

Abstract

The present invention proposes a prediction mode selection method of HEVC lossless video coding and a corresponding coding method. The selection method first initializes the intra-frame prediction and inter-frame prediction of each LCU of the video frame for the Lagrangian multiplication of RDO. and perform intra-frame prediction and inter-frame prediction, select the best mode for encoding, calculate the bit rate required for encoding the frame, and use it as the initial reference bit rate, then change the multipliers of the K LCUs in the frame and select Find the best mode for encoding, calculate the current bit rate, calculate the difference between the current bit rate and the reference bit rate, accept or reject the changed multiplier according to predetermined rules, and use the current bit rate as a reference when accepting Bit rate, judge whether there is no need to change the multiplier, if not, then end, otherwise return to iterate. The method of the invention makes little changes to the original encoding structure of HEVC, eliminates the adverse effect of QP in the original standard on the encoding efficiency of the lossless mode, and greatly reduces the encoding bit rate.

Description

HEVC lossless video coding prediction mode selection method and corresponding coding method

technical field

The present invention relates to the field of video coding, and more specifically, to a prediction mode selection method of HEVC lossless video coding and a video coding method based on the method.

Background technique

In recent years, with the continuous development of communication technology and multimedia technology, people's demand for multimedia communication such as video is also increasing. However, the amount of video data is huge, and video data that has not been encoded and compressed cannot be transmitted in existing channels. In order to meet the above various requirements, various video coding schemes have been proposed in the world. Since the 1990s, ITU.T and ISO have jointly formulated a series of international standards and recommendations on video compression codecs. Among them, the H.26X series video proposed by ITU Compression standards and the MPEG series of international standards introduced by ISO/IEC JTC have the greatest influence. In January 2013, the video coding standardization organization JCT-VC (Joint Collaborative Team on Video Coding) officially released the latest generation of video coding international standards - High Efficiency Video Standard HEVC (High Efficiency Video Coding), under the same subjective video quality, its The bit rate is about 50% of the previous generation video coding standard H.264/AVC.

The basic framework of HEVC encoding is similar to the previous H.264/AVC standard, and still adopts the hybrid encoding mode. The entire coding process is mainly divided into four steps: prediction, transformation, quantization, and entropy coding. The prediction part is divided into two categories: intra prediction and inter prediction. Intra-frame prediction uses the reconstructed pixels of the current frame as reference pixels for preliminary prediction, and inter-frame prediction uses the reconstructed pixel values of previous or subsequent frames as reference. Then the obtained prediction value is subtracted from the original pixel value of the current block to obtain the residual. The residual is transformed and quantized to obtain the transform coefficient, and finally the transform coefficient is entropy encoded to obtain the final code stream. Whether it is inter-frame prediction or intra-frame prediction, the information of the reconstructed image needs to be used. Therefore, in the encoding process, the residual image needs to be inversely transformed and quantized, and then the residual image is added to the predicted value, and finally The noise in the video image is filtered out through a loop filter, and at the same time, the degradation of the video image such as block effect can be avoided.

In the HEVC standard, there are two types of encoding modes: lossy compression and lossless compression. For most videos transmitted on the Internet, proper lossy compression can reduce the bit rate very well, thereby improving the efficiency of transmission. For medical video, remote sensing video, fingerprint and other fields, lossless compression also has great applications. In the HEVC standard, lossless compression exists as an extension of lossy compression. Since there is distortion in quantization, and transformation and quantization are combined in HEVC, the transformation process also has distortion. Therefore, in HEVC lossless compression coding, the transformation and quantization process is turned off. In addition, because there is no distortion of pixels before and after encoding, loop filtering is not required, and the HEVC lossless compression operation therefore skips the loop filtering process. That is to say, in the current HEVC lossless compression operation, there are only two parts, prediction and entropy coding. It can be seen that the quality of prediction will directly affect the performance of HEVC lossless compression.

As mentioned above, the prediction of HEVC is divided into intra prediction and inter prediction. For intra prediction, there are as many as 35 modes of predicted directions. Each mode corresponds to an intra prediction direction, as shown in FIG. 1 . With such a fine mode, it is convenient to find a mode that best removes intra-frame redundancy, so as to maximize the improvement of intra-frame coding performance. In HEVC, rate distortion optimization (RDO) is used to select the best mode, that is, a mode with the smallest RD cost (rate distortion cost, rate distortion cost) is selected as the prediction mode. The calculation of RDcost is shown in formula (1):

RD cost＝D+λR (1)

Among them, D represents the degree of distortion between the reconstructed image block and the original image block, R represents the bit rate required to encode the prediction block in this mode (including the coding residual and the bit rate required to code the prediction mode), and λ is Rag The Langer multiplier is used to weigh the weight of the bit rate in the RD cost, which can be obtained by calculating formula (2):

λ＝Qpfactor×2 ^(QP-12)/3 (2)

Among them, QP is a quantization parameter (quantization parameter), which is used to indicate different quantization step sizes. The larger the QP, the larger the quantization step size, that is, the larger the quantization distortion generated. Qpfactor is the quantization coefficient. In this formula It is a constant, which is determined by the specific encoding conditions such as the frame type.

It can be seen from formula (1) that for each calculation of an RD cost, two parts D and R need to be calculated, especially when calculating D, it is necessary to transform and quantize the residual image (the difference between the predicted image and the original image), and then reconstruct it to obtain Reconstructing an image block and then calculating the degree of distortion requires a high amount of computation. However, if the rate-distortion of these 35 modes is directly calculated one by one when selecting the best mode, the computation load will be very high. In order to reduce the amount of computation, HEVC first calculates the RD cost in a rough way to select several better candidate modes, and then performs accurate RD cost calculations on these candidate modes to select the best mode. The so-called rough method refers to using the simple Hadamard transformation of the residual image block to replace the calculation of the complex distortion degree D, thereby effectively reducing the complexity.

For inter-frame prediction, the encoder searches for the image block most similar to the prediction block around the position of the prediction block in the reference frame. In order to find the best matching image block, the precision of the motion vector in HEVC reaches 1/4 pixel. Similarly, when comparing which position of the image block in the reference frame is used as the reference block, the position with the smallest RD cost is also searched for. Similar to intra-frame prediction, if the optimal image block is found within the range of 64×64, and the precision is 1/4 pixel, then RDcost will be calculated 16384 times, and the calculation amount is also very large. In order to reduce the amount of computation, during the motion search, HEVC uses the method of calculating the absolute difference and SAD (sum of abstract distortion) instead of calculating the actual distortion to perform a rough search.

In many current applications, there is an increasing demand for lossless video compression. Since HEVC's lossless coding mode can provide a higher compression rate, its application in video lossless compression coding will become more and more extensive. In the mainstream applications of the HEVC standard, such as set-top boxes, video surveillance, etc., high coding efficiency is required, so it has to be at the cost of some distortion, that is to say, the mainstream applications are mostly lossy compression. However, on top of the HEVC lossy compression standard, it is obviously impractical to develop a new set of encoding tools to obtain the best possible lossless encoding efficiency. Therefore, when designing the HEVC lossless coding scheme, we should follow the following design principles: make use of the existing HEVC lossy coding structure as much as possible, and the proposed lossless coding scheme changes the original lossy coding structure as little as possible, while Consider the balance of coding efficiency and complexity.

As can be seen from the above introduction, HEVC lossless compression is relatively lossy compression, only skipping the three parts of transformation, quantization, and loop filtering, while the strategies of prediction and entropy coding remain unchanged. Lossy compression needs to balance reconstruction video quality and encoding bit rate for encoding. Different from lossy compression, in lossless compression, D in formula (1) is finally zero, so there is no problem of degradation of reconstructed video quality, and the only goal of its coding is to achieve as much as possible under the allowable codec complexity. Possibly lower bitrate. In lossless compression, if D is accurately calculated, the final distortion D is zero. At this time, no matter what λ is, minimizing RD cost is equivalent to minimizing bit rate R, so the original lossy compression at this time Strategies still apply to lossless compression. However, whether it is intra prediction or inter prediction, it is impossible to accurately calculate the RD cost for each mode, and the rough selection strategy mentioned above must be used. In lossy compression, the roughly calculated D is not much different from the actual D, but in lossless compression, the roughly calculated D is quite different from the actual zero D, then according to formulas (1), (2 ) may not contain the actual optimal mode, so its final coding performance may not be optimal. Therefore, formulas (1) and (2) must be redefined in order to obtain better lossless compression performance.

In addition, during lossy compression, we can obtain code streams with different qualities and different bit rates by adjusting QP to meet the needs of different application scenarios. This is because, on the one hand, adjusting QP can obtain different λ, thereby adjusting the trade-off relationship between bit rate and reconstructed image quality in RD cost. On the other hand, and the most important aspect, the smaller the QP, the smaller the quantization error, and the better the quality of the decoded and reconstructed video image, but the higher the bit rate of the code stream generated by encoding. Therefore, in lossy compression, if other encoding conditions are the same, the bit rate decreases with the increase of QP, and the quality of video image decoding and reconstruction decreases with the increase of QP. However, in lossless compression, there is no quantization anymore, and QP has no other role except for adjusting λ. Figure 2 is a graph of the change of the bit rate of the sequence Racehorse with QP (HM 10.0) in lossless compression coding. It can be found from Figure 2 that in lossless compression, the bit rate of the coded stream has a non-monotonic relationship with QP, regardless of QP The minimum bitrate cannot be reached neither at minimum (0 for encoded video at 8 bit, -12 for 10 bit) nor at maximum (51 for encoded video at 8 bit, 39 for 10 bit). Therefore, for lossless compression, the function of the coding parameter QP is different, and its corresponding relationship with λ is no longer as shown in formula (2), and needs to be re-determined.

Contents of the invention

In view of this, the object of the present invention is to provide a prediction mode selection method for HEVC lossless video coding, to readjust the rough selection strategy for many modes in the coding prediction process, and reduce the lossless compression bit rate to the greatest extent.

In order to achieve the above object, the present invention provides 1. A prediction mode selection method for HEVC lossless video coding, comprising the following steps in sequence:

S1. For a video frame, use the intra prediction and inter prediction of each LCU of the video frame for the Lagrangian multiplier of RDO with Initialize, perform intra-frame prediction and inter-frame prediction, select the best intra-frame prediction mode and the best inter-frame prediction mode, and select the best mode from the best intra-frame prediction mode and the best inter-frame prediction mode for encoding , calculate the bit rate required for encoding the frame, as the initial reference bit rate R ₀ , i, j represent the position coordinates of the LCU; S2, change the K LCUs of the video frame with Perform intra-frame prediction and inter-frame prediction, select the best intra-frame prediction mode and the best inter-frame prediction mode, select the best mode from the best intra-frame prediction mode and the best inter-frame prediction mode for encoding, and calculate the encoding The required bit rate of the video frame is used as the current bit rate R′, K is a natural number; S3, calculating the difference ΔR between the current bit rate and the reference bit rate, accepting or rejecting the changed according to predetermined rules with And when accepting, use the current bit rate as the reference bit rate; S4, judge whether there is no need to change the with When no further changes are required, the current with as the final with Otherwise return to step S2.

Preferably, in step S1, the Lagrangian multiplier with initialized to with said The value range is [3, 8], the The value range is [0.5, 1.5].

Preferably, in step S1, for intra prediction, use the minimized formula (8) The cost criterion selects the candidate mode from various intra-frame prediction modes, and then selects the best intra-frame prediction mode from the candidate modes according to the formula (1), the formula is as follows:

RD cost＝D+λR (1)

in, Represents the rough rate-distortion cost of intra prediction, D _intra is the prediction residual distortion of intra prediction; R _intra represents the bit rate required to encode other prediction information except residual information in intra prediction; λ is Lagrangian Multiplier; D is the actual residual distortion of intra prediction, R is the actual coding bit rate, and RD cost is the actual rate-distortion cost.

Preferably, in step S1, for inter-frame prediction, use the minimized formula (9) The cost criterion selects the optimal MV and the corresponding reference frame, and then selects the best inter-frame prediction mode from each inter-frame mode according to the formula (1), the formula is as follows:

RD cost＝D+λR (1)

in, Represents the rough rate-distortion cost of inter-frame prediction, D _inter is the prediction residual distortion of inter-frame prediction, R _inter represents the bit rate required to encode other prediction information except residual information in inter-frame prediction, λ is Lagrangian Multiplier, D is the actual residual distortion of intra prediction; R is the actual coding bit rate, and RD cost is the actual rate-distortion cost.

Preferably, in step S2, K is the number of LCUs contained in each column of the video frame.

Preferably, in step S3, if ΔR≤0, then accept the changed with Otherwise _accept the with

Preferably, step S3 is: recording the number of iterations Ti and continuously rejecting the with Tr, if the number of iterations Ti reaches an iteration threshold THi, or the number of consecutive rejections Tr reaches a threshold THr of consecutive rejections, the current with as the final with

Preferably, step S1 further includes initializing the number of iterations Ti and the number of consecutive rejections Tr to 0.

Preferably, in step S1, the value range of THi is [1, 1000], and the value range of THr is [1, 50].

The present invention also proposes a HEVC lossless video coding method, including the aforementioned HEVC lossless video coding prediction mode selection method.

From the above technical solutions, it can be seen that the prediction mode selection method for HEVC lossless video coding of the present invention has the following advantages: the original coding structure of HEVC is changed very little, and the existing coding tools of HEVC are fully utilized; Transformation and quantization are skipped. Therefore, the coding tool QP has no practical significance in lossless compression. This method eliminates the adverse effect of QP on coding efficiency in the original standard; compared with the lossless compression method in HM 10.0, this method The coding bit rate of the invention has been greatly reduced. Under three kinds of coding environments of RA-Main, LDB-Main and LDP-Main, the bit rate of the inventive method has reduced by 1.2-1.4% on average, especially the F class test sequence, It achieves bitrate savings of up to 2.5-2.9% on average.

Description of drawings

FIG. 1 is a schematic diagram of directions corresponding to various intra prediction modes in the prior art;

Fig. 2 is a graph of the sequence Racehorse bit rate varying with QP (HM10.0) in lossless compression coding;

Fig. 3 is a schematic diagram of the Lagrangian multipliers (each block in the figure represents an LCU) of intra-frame prediction and inter-frame prediction for RDO of each LCU;

FIG. 4 is a specific flow chart of the prediction mode selection method for HEVC lossless video coding according to the present invention.

detailed description

In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with specific embodiments and with reference to the accompanying drawings.

In the HEVC standard, there are two types of encoding modes: lossy compression and lossless compression. In lossless compression, there is no reconstruction distortion, so its only purpose is to minimize the bit rate, ie, minimize the bit rate R. Since lossless compression only exists as an extension of HEVC, in principle, it is not expected to make too many changes to the original HEVC. When performing coarse mode selection, the variables in Equation (1) are different from those for precise selection, where the degree of distortion D between the reconstructed image patch and the original image patch essentially represents the residual-related distortion function, namely

D=f(residue) (3)

where residue represents the residual.

Likewise, R _predInfo indicates the bitrate required to encode this prediction mode. When intra-frame prediction is performed, R _predInfo indicates the bit rate required for encoding the direction corresponding to the mode, and when inter-frame prediction is performed, R _predInfo indicates the bit rate required for encoding the motion vector (MV) information corresponding to the mode. Since the entropy encoding of the residual information is performed subsequently, it can be considered that the larger the residual, the greater the bit rate required for subsequent encoding of the residual, namely

R _residue = αf(residue), α > 0 (4)

Among them, R _residue represents the bit rate actually required for encoding the residual information, and α represents the proportional coefficient.

Then the final bit rate R _total required for a certain mode is the sum of the bit rate required for encoding the residual and the bit rate required for encoding the prediction mode information except for the residual information, that is

R _total = R _residue + R _predInfo (5)

From formulas (4) and (5), we can get:

R _total ＝αf(residue)+R _predInfo ＝αD+R _predInfo (6)

Since α>0, then minimizing R _total is equivalent to minimizing R _total /α, that is, minimizing

D+R _predInfo /α (7)

Comparing formulas (1) and (7), it can be found that when λ=1/α, minimizing the RD cost in formula (1) is consistent with minimizing the final bit rate R _total . Therefore, as long as the proportional relationship α between R _residue and f(residue) is found, λ=1/α can be obtained, and the optimization of lossless mode selection can be realized on the basis of the original optimization strategy of lossy mode selection. That is, the lossless coding bit rate minimization.

How to obtain the proportional relationship α will be analyzed below.

As mentioned above, there is no transformation and quantization of the residual in the HEVC lossless compression standard, so after the residual is obtained, entropy coding is performed directly, so entropy coding is an important factor affecting the bit rate required for subsequent coding residuals . In the residual entropy coding of the HEVC standard, entropy coding is performed by using Context-based Adaptive Binary Arithmetic Coding (CABAC for short). In the CABAC entropy coding of the HEVC standard, except that the initialization of the context probability model of each frame is related to QP, other parts do not involve QP. In CABAC, the context model will be adaptively adjusted during the encoding process to optimize the encoding efficiency. Therefore, the initial context probability model only affects the encoding of the first few residual bits, and has very little impact on the entire residual encoding. Therefore, it can be considered that in the process of entropy encoding, QP has very little impact on the final encoding efficiency. Therefore, it can be considered that there is no significant relationship between α and QP. However, it can be seen from formula (2) that in lossy compression, λ=1/α is closely related to QP, which is the difference between lossless compression and lossy compression in HEVC standard prediction mode selection.

Therefore, the present invention proposes a λ parameter design method that is completely independent of QP.

Considering the differences in the residual characteristics of each block, the following method tries to find the optimal combination of Lagrangian multipliers for each largest coding unit (LCU) in a certain frame.

FIG. 3 is a schematic diagram of the Lagrangian multipliers (each block in the figure represents an LCU) used for RDO by intra-frame prediction and inter-frame prediction of each LCU. As shown in Figure 3, the present invention first uses the intra prediction and inter prediction of each LCU for the Lagrangian multiplier of RDO with are initialized to with The value range is [3, 8], The value range is [0.5, 1.5], which is based on the test provided in the literature (F.Bossen, "Common Test Conditions and SoftwareReference Configurations", JCT-VC document, JCTVC-L 1100, Geneva, Jan.2013) Sequence and test conditions are obtained, which can be adjusted according to actual coding conditions. Set the number of consecutive rejections Tr and the number of iterations Ti to 0, and obtain the encoding bit rate R ₀ of the current parameter after encoding the frame, and randomly change the number of K LCUs with Then encode, K is a natural number, calculate the bit rate R' required to encode the frame under the new Lagrangian multiplier, and calculate the bit rate difference ΔR=R'-R ₀ , if ΔR≤0, then accept the pull The Lagrangian multiplier is used as the new Lagrangian multiplier, otherwise, the Lagrangian multiplier setting is accepted as the new Lagrange multiplier with the probability of exp(-ΔR/R ₀ ). If the new Lagrangian multiplier is accepted, the number of consecutive rejections Tr is set to 0, and R ₀ =R′, otherwise, the number of consecutive rejections Tr is increased by 1. Next, the number of iterations Ti is increased by 1.

After updating the number of consecutive rejections Tr and the number of iterations Ti, judge the following two conditions:

(1) Whether the number of consecutive rejections Tr is less than the threshold THr;

(2) Whether the iteration number Ti is smaller than the threshold THi.

Wherein, the value range of THi is [1, 1000], and the value range of THr is [1, 50]. This value range is obtained based on the test sequence and test conditions provided in the literature (F.Bossen, "Common Test Conditions and Software Reference Configurations", JCT-VC document, JCTVC-L 1100, Geneva, Jan.2013), and can be obtained according to the actual Coding condition adjustment. If both conditions are met, randomly change the LCUs of K LCUs on the basis of the current Lagrangian multiplier with Carry out the next encoding, then calculate the encoding bit rate R′ under the new Lagrangian multiplier, calculate ΔR again, and use the same strategy to decide whether to accept the new Lagrangian multiplier according to ΔR, and update the number of consecutive rejections Tr and The number of iterations Ti. If only one of the above two conditions is not satisfied, the latest Lagrangian multiplier is used as the optimal setting of the Lagrangian multiplier for each LCU intra prediction and inter prediction of the current frame, so far the end of the The encoding of the frame.

Repeat the above steps until one of the conditions (1) and (2) is not satisfied, then use the latest Lagrangian multiplier as the Lagrangian multiplier of each LCU intra prediction and inter prediction of the current frame Optimal settings for numbers. So far, the encoding of this frame ends.

Fig. 4 is the flow chart of the prediction mode selection method of HEVC lossless video coding of the present invention, as shown in the figure, it comprises the following steps:

S1. For a video frame, initialize the intra prediction and inter prediction of each LCU of the video frame for the Lagrangian multiplier of RDO with And perform intra-frame prediction and inter-frame prediction, select the best intra-frame prediction mode and the best inter-frame prediction mode, select the best mode from the best intra-frame prediction mode and the best inter-frame prediction mode for encoding, and calculate and encode the frame The required bit rate is used as an initial reference bit rate R ₀ , and i and j represent the location coordinates of the LCU.

And initialize the number of consecutive rejections Tr and the number of iterations Ti.

For example, in a specific embodiment, the intra prediction and inter prediction of each LCU are used for the Lagrangian multiplier of RDO with are set to with with Set to 5 and 1 respectively. Both the number of consecutive rejections Tr and the number of iterations Ti are initialized to 0.

Next, in the specific implementation process, the present invention proposes that, for intra prediction, the minimized formula (8) can be used The cost criterion selects the candidate mode from various intra-frame prediction modes, and then selects the best intra-frame prediction mode from the candidate modes according to the formula (1); for the inter-frame prediction, use the minimized formula (9) The cost criterion selects the optimal MV, and then selects the best inter prediction mode from each inter mode according to formula (1).

in, Represents the rough rate-distortion cost of inter prediction and intra prediction, respectively, D _intra and D _inter are the prediction residual distortion of intra prediction and inter prediction, respectively, and can be expressed by the absolute difference between the predicted value and the original value. R _intra and D _inter respectively represent the bit rates required for coding other prediction information except residual information in intra prediction and inter prediction.

After selecting the best mode from the best intra-frame prediction mode and the best intra-frame prediction mode for prediction and entropy encoding, the current bits required for encoding the frame are obtained as the reference bit rate R ₀ .

Step S2, changing the K LCUs of the video frame with Perform intra-frame prediction and inter-frame prediction, select the best intra-frame prediction mode and the best inter-frame prediction mode, select the best mode from the best intra-frame prediction mode and the best inter-frame prediction mode for encoding, and calculate The bit rate required to encode the frame is taken as the current bit rate R', and K is a natural number.

In a specific embodiment, the value of K cannot be too small, otherwise the optimal parameter combination cannot be obtained in the end; nor can it be too large, otherwise the computational complexity will be very high. K may be the number of LCUs contained in each column of the video frame.

Step S3, calculating the difference ΔR between the current bit rate and the reference bit rate, accepting or rejecting the changed bit rate according to predetermined rules with And take the current bit rate as the reference bit rate when accepting.

Specifically, the difference ΔR is calculated using ΔR=R′−R ₀ . If ΔR≤0, accept the changed with Otherwise _accept the with accepting the said with , let R ₀ =R'.

Step S4, judging whether there is no need to change the with When there is no need to change, the current with as the final with Otherwise return to step S2.

One implementation is to record the number of iterations Ti and continuously reject the with The number of times Tr, if the number of iterations Ti reaches a threshold of iterations, or the number of consecutive rejections Tr reaches a threshold of consecutive rejections, the current with as the final with In a specific embodiment, Ti and Tr may be initialized to 0 in step S1. When performing iterations, the number of iterations Ti is increased by 1 for each iteration. If said with If it is accepted, set the number of consecutive rejections Tr to 0, otherwise add 1 to the number of consecutive rejections Tr. The value range of THi may be [1, 1000], and the value range of THr may be [1, 50].

For the HEVC lossless video coding method of the present invention, the selected method according to the above prediction mode selection method with Perform lossless video encoding.

Example

In order to verify the beneficial effects of the present invention, the method is implemented on the latest HEVC reference code HM10.0, and compared with the original lossless coding method in the reference code that skips transform, quantization, and filtering. In the simulation comparison, the test sequence and test conditions provided in the literature (F.Bossen, "Common Test Conditions and Software Reference Configurations," JCT-VC document, JCTVC-L1100, Geneva, Jan.2013) were used as the simulation comparison environment. Among them, we adopted three encoding environments, namely Random Access Main Profile encoding (RA-Main), Low Delay BMain Profile encoding (LDB-Main) and Low Delay BMain Profile encoding (LDB-Main). P main file encoding (Lowdelay P Main Profile encoding, LDP-Main). We tested all video sequences from category A to category F, where category F is a screen recording video sequence, and the resolutions of the contained sequences are different, there are multiple resolutions of 1280×720, 1024×768 and 832×480. The resolutions of the video sequences of the categories are shown in Table 1. Since the default encoding method in the reference software HM10.0 is lossy encoding, we need to set the flag indicating whether to perform lossless encoding in the configuration file to 1, that is, perform lossless encoding.

Table 1 Various test sequences

sequence type resolution Class A 2560×1600 Class B 1920×1080 Class C 832×480 Class D 416×240 Class E 1280×720

Table 2-4 shows the bit rates of the original lossless coding method in HM 10.0 and the method proposed by the present invention under the three coding environments of RA-Main, LDB-Main and LDP-Main, respectively, and the bit rate saving rate is obtained through the following Obtained by:

Bit rate change rate = 100 x (bit rate of the present invention - bit rate of the original method) / bit rate of the original method %.

Table 2 Experimental results in RA-Main configuration environment

Table 3 Experimental results in LDB-Main configuration environment

Table 4 Experimental results in LDP-Main configuration environment

From Table 2-4, we can see that compared with the lossless compression method in HM10.0, the encoding bit rate of the present invention has been greatly reduced. Under the three encoding environments, the bit rate of the method of the present invention is reduced by 1.2-1.4% on average, especially for the class F test sequence, which obtains an average bit rate saving of 2.5-2.9%.

The specific embodiments described above have further described the purpose, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above descriptions are only specific embodiments of the present invention, and are not intended to limit the present invention. Within the spirit and principles of the present invention, any modifications, equivalent replacements, improvements, etc., shall be included in the protection scope of the present invention.

Claims (10)

1. A prediction mode selection method for HEVC lossless video coding, comprising the following steps carried out in sequence: S1. For a video frame, use the intra prediction and inter prediction of each LCU of the video frame for the Lagrangian multiplier of RDO with Initialize, perform intra-frame prediction and inter-frame prediction, select the best intra-frame prediction mode and the best inter-frame prediction mode, and select the best mode from the best intra-frame prediction mode and the best inter-frame prediction mode for encoding , calculate the bit rate required for encoding the frame, as the initial reference bit rate R ₀ , i, j represent the position coordinates of the LCU; S2. Change the K LCUs of the video frame with Perform intra-frame prediction and inter-frame prediction, select the best intra-frame prediction mode and the best inter-frame prediction mode, select the best mode from the best intra-frame prediction mode and the best inter-frame prediction mode for encoding, and calculate the encoding The required bit rate of this video frame, as current bit rate R ', K is a natural number; S3. Calculate the difference ΔR between the current bit rate and the reference bit rate, and accept or reject the changed bit rate according to predetermined rules with And take the current bit rate as the reference bit rate when accepting; S4, judging whether there is no need to change the with When no further changes are required, the current with as the final with Otherwise return to step S2. 2. The prediction mode selection method of HEVC lossless video coding according to claim 1, characterized in that, in step S1, the Lagrangian multiplier with initialized to with said The value range is [3, 8], the The value range is [0.5, 1.5]. 3. The prediction mode selection method of HEVC lossless video coding according to claim 1, characterized in that, in step S1, for intra-frame prediction, using the minimized formula (8) The cost criterion selects the candidate mode from various intra-frame prediction modes, and then selects the best intra-frame prediction mode from the candidate modes according to the formula (1), the formula is as follows: <mrow><msubsup><mi>J</mi><mrow><mi>i</mi><mo>,</mo><mi>j</mi></mrow><mi>intra</mi></msubsup><mo>=</mo><msub><mi>D</mi><mi>intra</mi></msub><mo>+</mo><msubsup><mi>&amp;lambda;</mi><mrow><mi>i</mi><mo>,</mo><mi>j</mi></mrow><mi>intra</mi></msubsup><msub><mi>R</mi><mi>intra</mi></msub><mo>-</mo><mo>-</mo><mo>-</mo><mrow><mo>(</mo><mn>8</mn><mo>)</mo></mrow></mrow> RD cost＝D+λR (1) in, Represents the rough rate-distortion cost of intra prediction, D _intra is the prediction residual distortion of intra prediction; R _intra represents the bit rate required to encode other prediction information except residual information in intra prediction; λ is Lagrangian Multiplier; D is the actual residual distortion of intra prediction, R is the actual coding bit rate, and RD cost is the actual rate-distortion cost. 4. The prediction mode selection method of HEVC lossless video coding according to claim 1, characterized in that, in step S1, for inter-frame prediction, using the minimized formula (9) The cost criterion selects the optimal MV and the corresponding reference frame, and then selects the best inter-frame prediction mode from each inter-frame mode according to the formula (1), the formula is as follows: <mrow><msubsup><mi>J</mi><mrow><mi>i</mi><mo>,</mo><mi>j</mi></mrow><mi>inter</mi></msubsup><mo>=</mo><msub><mi>D</mi><mi>inter</mi></msub><mo>+</mo><msubsup><mi>&amp;lambda;</mi><mrow><mi>i</mi><mo>,</mo><mi>j</mi></mrow><mi>inter</mi></msubsup><msub><mi>R</mi><mi>inter</mi></msub><mo>-</mo><mo>-</mo><mo>-</mo><mrow><mo>(</mo><mn>9</mn><mo>)</mo></mrow></mrow> RD cost＝D+λR (1) in, Represents the rough rate-distortion cost of inter-frame prediction, D _inter is the prediction residual distortion of inter-frame prediction, R _inter represents the bit rate required to encode other prediction information except residual information in inter-frame prediction, λ is Lagrangian Multiplier, D is the actual residual distortion of intra prediction; R is the actual coding bit rate, and RD cost is the actual rate-distortion cost. 5. The prediction mode selection method of HEVC lossless video coding according to claim 1, wherein in step S2, K is the number of LCUs contained in each column of the video frame. 6. The prediction mode selection method of HEVC lossless video coding according to claim 1, characterized in that, in step S3, if ΔR≤0, then accept the changed with Otherwise _accept the with 7. The prediction mode selection method of HEVC lossless video coding according to claim 1, characterized in that, step S3 is: recording the number of iterations Ti and continuously rejecting the with Tr, if the number of iterations Ti reaches an iteration threshold THi, or the number of consecutive rejections Tr reaches a threshold THr of consecutive rejections, the current with as the final with 8. The prediction mode selection method for HEVC lossless video coding according to claim 7, wherein the value range of THi is [1, 1000], and the value range of THr is [1, 50]. 9. The prediction mode selection method of HEVC lossless video coding according to claim 7, characterized in that the step S1 further comprises initializing the number of iterations Ti and the number of consecutive rejections Tr to 0. 10. A method for HEVC lossless video coding, comprising the prediction mode selection method for HEVC lossless video coding according to any one of claims 1-9.