CN102769748B - Motion vector prediction method, device and system - Google Patents

Abstract

The invention provides a motion vector prediction method, device and system. The method includes: detecting whether the motion vector of the reference block at the first adjacent position of the current block is available, and if available, recording it as a candidate motion vector in space; detecting whether the motion vector of the reference block at the second adjacent position of the current block is available , if available, record it as a candidate motion vector in the spatial domain; if the motion vector of the reference block at the second adjacent position of the current block is not available, the reference block at the first adjacent position and/or the second adjacent position Check whether the scaled motion vectors of reference blocks in two adjacent positions are available, and if available, record them as candidate motion vectors in space; obtain and record candidate motion vectors in time domain; A candidate motion vector is used to predict the motion vector of the current block.

Description

Motion vector prediction method, device and system

technical field

The present invention relates to video encoding and decoding technology, in particular to a motion vector prediction method, device and system, and belongs to the field of communication technology.

Background technique

The basic principle of video coding and compression is to use the correlation between space domain, time domain and code words to remove redundancy as much as possible. Therefore, how to use adjacent coding blocks to predict the motion vector (MotionVector, MV) of the current block is an important factor affecting the execution efficiency and gain of coding and decoding.

The JCTVC working group is developing the next-generation video coding standard, the High Efficiency Video Coding (HEVC) standard. The HEVC standard uses the method of multi-motion vector competition for motion vector prediction, that is, the space-domain candidate motion vector is derived from the spatial-domain adjacent coding block, and the time-domain candidate motion vector is derived from the temporal-domain adjacent coding block, and the best candidate motion vector is selected from these candidate motion vectors. The best motion vector is used as the predictor of the motion vector of the current block.

FIG. 1 is a schematic diagram of spatial adjacent coding blocks used for deriving spatial candidate motion vectors in the draft HEVC standard. As shown in Figure 1, A0 is the lower left block corresponding to the lower left corner of the current block, A1 is the left block corresponding to the lower left corner of the current block, B0 is the upper right block corresponding to the upper right corner of the current block, and B1 is the upper right block corresponding to the current block B2 is the upper left block corresponding to the upper left corner of the current block. With reference to Figure 1, the derivation of spatial candidate motion vectors in the draft HEVC standard includes the following steps performed in sequence:

Step a, sequentially detect whether the MV of A0 and A1 is available, and if the MV is available, use the MV as the candidate motion vector A in the space;

Step b, if the MV in step a is not available, then sequentially detect whether the scaled MV (scaled MV) of A0 and A1 is available, and if the scaled MV is available, use the scaled MV as the spatial candidate motion vector A;

Step c, sequentially detect whether the MVs of B0, B1, and B2 are available, and if the MV is available, use the MV as the spatial candidate motion vector B;

Step d, if the MV of step c is not available, and the MV of step a and the scaled MV of step b are not available, then check whether the scaled MV of B0, B1, B2 is available in turn, and if the scaled MV is available, the scaled MV As the spatial domain candidate motion vector B.

Through the above steps, at most two spatial candidate motion vectors can be obtained, including the spatial candidate motion vector A and the spatial candidate motion vector B.

It can be seen that the above-mentioned process of deriving the motion vector candidate in the space domain is relatively complicated, thereby reducing the execution efficiency of encoding and decoding.

Contents of the invention

Aiming at the defects existing in the prior art, the embodiments of the present invention provide a motion vector prediction method, device and system, so as to realize relatively simple motion vector prediction and improve the execution efficiency of encoding and decoding.

According to a first aspect of an embodiment of the present invention, a motion vector prediction method is provided, including:

Detect whether the motion vector of the reference block at the first adjacent position of the current block is available, and if available, record it as a spatial candidate motion vector;

Detecting whether the motion vector of the reference block at the second adjacent position of the current block is available, and if available, recording it as a spatial candidate motion vector;

If the motion vector of the reference block of the second neighboring position of the current block is not available, the scaled adjustment of the reference block of the first neighboring position and/or the reference block of the second neighboring position Whether the motion vector of is available for detection, if available, then recorded as the space domain candidate motion vector;

Acquiring and recording candidate motion vectors in time domain;

Perform motion vector prediction on the current block according to the recorded space-domain candidate motion vectors and time-domain candidate motion vectors.

According to the second aspect of the embodiments of the present invention, there is provided a motion vector prediction device, including a spatial domain candidate motion vector acquisition module, a time domain candidate motion vector acquisition module, and the spatial domain candidate motion vector acquisition module and the time domain candidate motion vector acquisition module respectively The motion vector prediction module connected to the candidate motion vector acquisition module, wherein:

The airspace candidate motion vector acquisition module includes:

The first detection unit is used to detect whether the motion vector of the reference block at the first adjacent position of the current block is available, and if available, record it as a candidate motion vector in the space;

The second detection unit is used to detect whether the motion vector of the reference block at the second adjacent position of the current block is available, and if available, record it as a spatial candidate motion vector; and

The third detection unit is configured to, if the motion vector of the reference block at the second adjacent position of the current block is not available, perform the motion vector detection on the reference block at the first adjacent position and/or the second adjacent position detecting whether the scaled motion vector of the reference block is available, and recording it as a spatial candidate motion vector if available;

The time-domain candidate motion vector acquisition module is used to acquire and record the time-domain candidate motion vector;

The motion vector prediction module is configured to perform motion vector prediction on the current block according to the recorded space-domain candidate motion vectors and time-domain candidate motion vectors.

According to a third aspect of the embodiments of the present invention, an encoder is provided, including the motion vector prediction apparatus of the embodiments of the present invention.

According to a fourth aspect of the embodiments of the present invention, a decoder is provided, including the motion vector prediction apparatus of the embodiments of the present invention.

According to a fifth aspect of the embodiments of the present invention, a codec system is provided, including the encoder and the decoder of the embodiments of the present invention.

According to the motion vector prediction method, device, and system provided by the embodiments of the present invention, in the process of obtaining the candidate spatial motion vector, firstly, it is detected whether the motion vector of the reference block at the first adjacent position and the second adjacent position of the current block is available, and when one of the two positions is not available, then check whether the scaled motion vector is available, so that in the worst case, three steps are required to obtain the spatial candidate motion vector. However, in the prior art, in the worst case, four steps need to be performed to obtain the spatial candidate motion vector. Therefore, it can be seen that, compared with the prior art, the motion vector prediction method of this embodiment reduces the complexity and reduces the flow of performing motion vector prediction in the prior art, thereby improving the execution efficiency of encoding and decoding.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only These are some embodiments of the present invention. For those skilled in the art, other drawings can also be obtained according to these drawings without any creative effort.

FIG. 1 is a schematic diagram of spatial adjacent coding blocks used to derive spatial candidate motion vectors in the draft HEVC standard;

FIG. 2 is a schematic flowchart of a motion vector prediction method according to an embodiment of the present invention;

FIG. 3 is a schematic diagram of reference blocks used to derive spatial candidate motion vectors in an embodiment of the present invention;

FIG. 4 is a schematic structural diagram of a motion vector prediction device according to an embodiment of the present invention;

FIG. 5 is a system architecture diagram of a codec system provided by an embodiment of the present invention.

Detailed ways

The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some, not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

Fig. 2 is a schematic flowchart of a motion vector prediction method according to an embodiment of the present invention. As shown in Figure 2, the motion vector prediction method includes:

101. Detect whether the motion vector of the reference block at the first adjacent position of the current block is available, and if available, record it as a spatial candidate motion vector;

102. Detect whether the motion vector of the reference block at the second adjacent position of the current block is available, and if available, record it as a spatial candidate motion vector;

103. If the motion vector of the reference block at the second adjacent position of the current block is unavailable, perform the motion vector of the reference block at the first adjacent position and/or the reference block at the second adjacent position, Whether the scaled motion vector is available is detected, and if available, it is recorded as a spatial candidate motion vector;

104. Acquire and record the time-domain candidate motion vector;

105. Perform motion vector prediction on the current block according to the recorded candidate motion vectors in the space domain and the candidate motion vectors in the time domain.

In the above 101-105, the execution order of 101-103 and 104 is not limited, that is, 104 can be executed after 101-103 is executed; 104 can be executed first, and then 101-103 can be executed; 101-103 and 101 can also be executed 104 are executed in parallel. Similarly, the execution order of 101 and 102 is not limited.

In the motion vector prediction method of this embodiment, the first adjacent reference block of the current block and the second adjacent reference block of the current block are used as spatial adjacent reference blocks for deriving at most two spatial candidate motions vector. Wherein, the first adjacent position of the current block may be the left position of the current block, the second adjacent position of the current block may be the upper position of the current block; the reference block of the first adjacent position of the current block and the first adjacent position of the current block The reference blocks at two adjacent positions may be inter-frame (Inter) coding blocks or intra-frame (Intra) coding blocks, which are not limited here.

Specifically, in the process of obtaining the spatial candidate motion vector, firstly, the reference block at the first adjacent position of the current block and the reference block at the second adjacent position of the current block are detected, and the first adjacent position of the current block is judged Whether the MV of the reference block and the second adjacent reference block of the current block are available, where "available" means that the MV exists and the reference image corresponding to the MV is the same as the reference image of the current block. More specifically, the step of detecting whether the reference block is available includes: detecting whether the reference block has an MV. When the reference block is an inter-coded block, it has an MV. When the reference block is an intra-frame coded block, it does not have an MV. If there is no MV, it is directly determined that it is not available; if there is an MV, then compare whether the reference image index corresponding to the MV of the reference block is the same as the reference image index of the current block, if they are the same, then determine that the MV is available, and record the corresponding MV as Spatial candidate motion vector; if different, the decision is not available.

After the above process is completed, if the MV of the reference block at the first adjacent position of the current block is available (that is, a spatial candidate motion vector is derived, for example, recorded as candidate A), and the MV of the reference block at the second adjacent position of the current block If the MV is available (that is, one spatial candidate motion vector is derived, for example, it is recorded as candidate B), then the requirement of deriving at most two spatial candidate motion vectors required by the HEVC standard has been met, and there is no need to perform the operation of 103 at this time.

If it is known that the MV of the reference block at the second adjacent position of the current block is not available (that is, the above process derives a spatial candidate motion vector, or does not derive a spatial candidate motion vector), then detect the reference block at the first adjacent position and/or Whether the scaled MV (scaledMV) of the reference block at the second adjacent position is available. More specifically, detecting whether the scaled MV is available includes: detecting whether there is an MV in the reference block, if it does not exist, then judging that the scaled MV is unavailable; if it exists, then judging that the scaled MV is available; Adjustment, that is, scaling processing (scaling), so that the reference image corresponding to the scaled MV is the same as the reference image of the current block, then the scaled MV is recorded as a spatial candidate motion vector.

The detection order of the scaled MV of the reference block at the first adjacent position and the reference block at the second adjacent position can be: first, the scaled MV of the reference block at the first adjacent position is detected, and if the detection result is scaled MV available , then stop executing 103; if the detection result is that the scaled MV is unavailable, continue to detect the scaled MV of the reference block at the second adjacent position; another example: if the detection result of 101 is available, the detection result of 102 is unavailable , then in 103, first detect the scaled MV of the reference block at the second adjacent position, and when the detection result is unavailable, then detect the scaled MV of the reference block at the first adjacent position.

Through the above steps 101-103, the spatial candidate motion vectors are obtained. In 104, adopt the existing time-domain candidate motion vector derivation method to obtain the time-domain candidate motion vector, specifically, for example, according to the time-domain reference block information, such as whether the mode information of the reference block is inter or intra, or the MV motion vector information etc., to obtain the candidate motion vectors in time domain. These spatial candidate motion vectors, together with temporal candidate motion vectors, constitute a motion vector list.

When encoding at the encoding end, select the best motion vector from the motion vector list as the motion vector predictor of the current block, and the best one can be the best in the sense of rate-distortion It can also be selected based on other performance indicators as required, which is not limited herein. The encoder writes the position index number of the selected motion vector in the list into the code stream.

Correspondingly, when decoding at the decoding end, obtain the motion vector index number through decoding, and according to the motion vector information of the decoded reference block of the current block, use 101-104 in this embodiment to construct a motion vector list, and pass from In the motion vector list, the motion vector corresponding to the motion vector index number obtained through decoding is selected as the motion vector predictor of the current block.

According to the motion vector prediction method of this embodiment, in the process of obtaining the candidate motion vector in the spatial domain, it is firstly detected whether the motion vectors of the reference blocks at the first adjacent position and the second adjacent position of the current block are available, and when the second When the motion vectors of the adjacent reference blocks are not available, it is checked whether the scaled motion vectors are available, so that in the worst case, three steps are required to obtain the spatial candidate motion vectors. However, in the prior art, in the worst case, four steps need to be performed to obtain the spatial candidate motion vector. Therefore, it can be seen that, compared with the prior art, the motion vector prediction method of this embodiment reduces the complexity and reduces the flow of performing motion vector prediction in the prior art, thereby improving the execution efficiency of encoding and decoding.

Embodiment two

On the basis of the first embodiment, in this embodiment, an extended description is made on the process of obtaining the candidate motion vector in the spatial domain.

Fig. 3 is a schematic diagram of reference blocks used for deriving spatial candidate motion vectors in an embodiment of the present invention. As shown in Figure 3, the reference block at the first adjacent position includes the lower left block A0 and the left block A1 corresponding to the lower left corner; the reference block at the second adjacent position includes the upper right block B0 and the upper block B1 corresponding to the upper right corner.

Specifically, the spatial candidate motion vectors in this embodiment include:

201. Detect whether the MV of the lower left block A0 and/or the left block A1 corresponding to the lower left corner of the current block is available, and if available, record it as candidateA;

Specifically, for example: first detect whether the MV of the lower left block A0 is available, if available, record the MV of the lower left block A0 as a candidate motion vector A (candidate A) in the spatial domain, and directly execute 202; if not, detect the left block Whether the MV of A1 is available, if the MV of the left block A1 is available, record the MV of the left block A1 as candidate A, and execute 202; if the MV of the left block A1 is not available, execute 202.

202. Detect whether the MV of the upper right block B0 and/or the upper block B1 corresponding to the upper right corner of the current block is available, and if available, record it as candidate B;

Specifically, for example: first detect whether the MV of the upper right block B0 is available, if available, record the MV of the upper right block B0 as a candidate motion vector B in the space, and directly execute 203; if not, detect the upper block Whether the MV of B1 is available, if the MV of the upper block B1 is available, record the MV of the upper block B1 as candidate B, and execute 203; if the MV of the upper block B1 is not available, execute 203.

203, whether there is no candidate B recorded; if so, execute 204, if not, end the process;

204. Detect whether the scaledMV of the lower left block A0 or the left block A1 corresponding to the lower left corner of the current block is available; if available, record it as a candidate motion vector in the space, for example, as candidate B, and end the process; if not available, execute 205;

Specifically, select one of the lower left block A0 and the left block A1, detect whether there is an MV in it, and if it exists, adjust the proportion of its MV to obtain the scaled MV, so that the reference image corresponding to the scaled MV is the same as the reference image of the current block The same, and record the scaled MV as candidate B. Wherein, for the selection method of the lower left block A0 and the left block A1, the lower left block A0 may be fixedly selected according to the previous configuration, or the left block A1 may be fixedly selected according to the previous configuration.

205. Detect whether the scaledMV of the upper right block B0 or the upper block B1 corresponding to the upper right corner of the current block is available; if available, record it as a candidate motion vector in the space, for example record as candidate B, and end the process; if not available, directly End the process.

Specifically, select one of the upper right block B0 and the upper block B1, detect whether there is an MV in it, and if it exists, adjust the proportion of its MV to obtain the scaled MV, so that the reference image corresponding to the scaled MV is the same as the reference image of the current block The same, and record the scaled MV as candidate B. Wherein, for the selection method of the upper right block B0 and the upper block B1, either the upper right block B0 can be fixedly selected according to the previous configuration, or the upper block B1 can be fixedly selected according to the previous configuration.

According to the motion vector prediction method of this embodiment, since the lower left block A0 and the left block A1 are used as the reference blocks at the lower left corner at the same time, and the upper right block B0 and the upper block B1 are used as the reference blocks at the upper right corner at the same time, when the When detecting whether the scaled MV of the reference block at the lower left corner position and/or the reference block at the upper right corner position is available, select one of the lower left block A0 and the left block A1, and one of the upper right block B0 and the upper block B1 as The object to be detected does not need to perform scaled MV detection on the lower left block A0 and the left block A1, as well as the upper right block B0 and the upper block B1. Therefore, compared with step b and step d in the prior art, the complexity is further reduced, the flow of executing motion vector prediction is reduced, and the execution efficiency of encoding and decoding is improved.

Embodiment three

On the basis of the foregoing embodiments, the motion vector prediction method of this embodiment includes:

301. Detect whether the MV of the lower left block A0 and/or the left block A1 corresponding to the lower left corner of the current block is available, and if available, record it as a candidate spatial motion vector;

This step is the same as step 201 in the above embodiment, so it will not be repeated here.

302. Detect whether the MV of the upper right block B0 and/or the upper block B1 corresponding to the upper right corner of the current block is available, and if available, record it as a candidate spatial motion vector;

This step is the same as step 201 in the above embodiment, so it will not be repeated here.

303, if the detection result of 302 is unavailable, then detect whether the scaled MV of the left block A1 corresponding to the lower left corner of the current block is available; if available, record it as a candidate motion vector in the space, and execute 305; if unavailable, then Execute 305;

304. Detect whether the scaled MV of the upper block B1 corresponding to the upper right corner of the current block is available; if available, record it as a spatial candidate motion vector, and execute 305; if not, directly execute 305;

In the above 303 and 304, the order of the detected reference blocks can be interchanged, that is: in 303, it can be detected whether the scaled MV of the upper block B1 corresponding to the upper right corner of the current block is available, and in 304, the lower left of the current block can be detected Whether the scaled MV of the left block A1 corresponding to the corner position is available.

305. Acquire and record a time-domain candidate motion vector;

306. Perform motion vector prediction on the current block according to the recorded candidate motion vectors in the space domain and the candidate motion vectors in the time domain.

The above steps 305 and 306 are the same as the steps 104 and 105 in the first embodiment, so they are not repeated here.

On the HEVC reference software platform HM7.0, the motion vector prediction method of the embodiment of the present invention is compared with the prior art in various configurations. The comparison results are shown in Table 1-3 below. Specifically, Table 1 shows the comparison results in the random access scenario (including the Main configuration with lower complexity and the HE10 configuration with higher complexity in this scenario), and Table 2 shows the comparison results in the low-latency B scenario ( Including the lower-complexity Main configuration and the more-complex HE10 configuration in this scenario), Table 3 shows the comparison results in the low-latency P scenario (including the lower-complexity Main configuration and the higher-complexity HE10 configuration).

Table 1

Table 2

table 3

As shown in Tables 1-3, each percentage value is the difference between the motion vector prediction method of the embodiment of the present invention and the prior art. For example, under the Main configuration of the random access scene in Table 1, the video luminance component value (Y) corresponding to Class A (ClassA) is 0.0%, which means that for this parameter, the motion vector prediction method of the embodiment of the present invention is better than the existing The technology consumes the same bit; the chrominance component value (U) corresponding to class A is -0.1%, which means that the motion vector prediction method of the embodiment of the present invention saves 0.1% of bits compared with the prior art for this parameter. Based on the overall performance difference in Tables 1-3, it can be seen that compared with the prior art, the motion vector prediction method of the embodiment of the present invention reduces the complexity and process, and does not cause video The performance is reduced, and in some scenarios, the video performance can be improved to a certain extent.

Embodiment four

On the basis of the first embodiment, in this embodiment, an extended description is made on the process of obtaining the candidate motion vector in the spatial domain.

In this embodiment, FIG. 1 is used as a schematic diagram of a reference block for deriving a candidate spatial motion vector. That is: the reference block at the first adjacent position includes the lower left block and the left block corresponding to the lower left corner position; the reference block at the second adjacent position includes the upper right block and upper side block corresponding to the upper right corner position, and the upper left block corresponding to the upper left corner position Upper left block.

Specifically, the spatial candidate motion vectors in this embodiment include:

401. Detect whether the MV of the lower left block A0 and/or the left block A1 corresponding to the lower left corner of the current block is available, and if available, record it as candidate A;

Specifically, for example: first detect whether the MV of the lower left block A0 is available, if available, record the MV of the lower left block A0 as a candidate motion vector A (candidate A) in the spatial domain, and directly execute 202; if not, detect the left block Whether the MV of A1 is available, if the MV of the left block A1 is available, record the MV of the left block A1 as candidate A, and execute 402; if the MV of the left block A1 is not available, execute 402.

402. Detect whether the MV of the upper right block B0 corresponding to the upper right corner of the current block, the upper block B1 and/or the upper left block B2 corresponding to the upper left corner of the current block is available, and if available, record it as candidate B;

Specifically, for example: first check whether the MV of the upper right block B0 is available, if available, record the MV of the upper right block B0 as a candidate motion vector B (candidate B) in the space domain, and end the process; if not, detect the upper block B1 Whether the MV of the upper block B1 is available, if the MV of the upper block B1 is available, record the MV of the upper block B1 as candidate B, and end the process; if the MV of the upper block B1 is not available, check whether the MV of the upper left block B2 is available. If the MV of block B2 is available, record the MV of the upper left block B2 as candidate B, and end the process; if the MV of the upper left block B2 is not available, execute 403.

403. Detect whether the scaledMV of the lower left block A0 or the left block A1 corresponding to the lower left corner of the current block is available; if available, record it as a candidate motion vector in the space, for example, as candidate B, and end the process; if not available, execute 404;

Specifically, select one of the lower left block A0 and the left block A1, detect whether there is an MV in it, and if it exists, adjust the proportion of its MV to obtain the scaled MV, so that the reference image corresponding to the scaled MV is the same as the reference image of the current block The same, and record the scaled MV as candidate B. Wherein, for the selection method of the lower left block A0 and the left block A1, the lower left block A0 may be fixedly selected according to the previous configuration, or the left block A1 may be fixedly selected according to the previous configuration.

404. Detect whether the scaledMV of the upper right block B0 or the upper block B1 corresponding to the upper right corner of the current block is available; if available, record it as a candidate motion vector in the space, for example record as candidate B, and end the process; if not available, directly End the process.

Specifically, select one of the upper right block B0 and the upper block B1, detect whether there is an MV in it, and if it exists, adjust the proportion of its MV to obtain the scaled MV, so that the reference image corresponding to the scaled MV is the same as the reference image of the current block The same, and record the scaled MV as candidate B. Wherein, for the selection method of the upper right block B0 and the upper block B1, either the upper right block B0 can be fixedly selected according to the previous configuration, or the upper block B1 can be fixedly selected according to the previous configuration.

According to the motion vector prediction method of this embodiment, since the lower left block A0 and the left block A1 are used as the reference blocks at the lower left corner at the same time, and the upper right block B0 and the upper block B1 are used as the reference blocks at the upper right corner at the same time, when the When detecting whether the scaled MV of the reference block at the lower left corner position and/or the reference block at the upper right corner position is available, select one of the lower left block A0 and the left block A1, and one of the upper right block B0 and the upper block B1 as The object to be detected does not need to perform scaled MV detection on the lower left block A0 and the left block A1, as well as the upper right block B0 and the upper block B1. Therefore, compared with step b and step d in the prior art, the complexity is further reduced, the flow of executing motion vector prediction is reduced, and the execution efficiency of encoding and decoding is improved.

Embodiment five

On the basis of the fourth embodiment above, in this embodiment, an extended description is given to the process of obtaining a candidate motion vector in the spatial domain.

In this embodiment, Fig. 1 is still used as a schematic diagram of a reference block used for deriving a candidate spatial motion vector. Specifically, the spatial candidate motion vectors in this embodiment include:

501. Detect whether the MV of the lower left block A0 and/or the left block A1 corresponding to the lower left corner of the current block is available, and if available, record it as candidate A;

Specifically, for example: first detect whether the MV of the lower left block A0 is available, if available, record the MV of the lower left block A0 as a candidate motion vector A (candidate A) in the spatial domain, and directly execute 202; if not, detect the left block Whether the MV of A1 is available, if the MV of the left block A1 is available, record the MV of the left block A1 as candidate A, and execute 502; if the MV of the left block A1 is not available, execute 502.

502. Detect whether the MV of the upper right block B0 corresponding to the upper right corner of the current block, the upper block B1 and/or the upper left block B2 corresponding to the upper left corner of the current block is available, and if available, record it as candidate B;

Specifically, for example: first check whether the MV of the upper right block B0 is available, if available, record the MV of the upper right block B0 as a candidate motion vector B (candidate B) in the space domain, and end the process; if not, detect the upper block B1 Whether the MV of the upper block B1 is available, if the MV of the upper block B1 is available, record the MV of the upper block B1 as candidate B, and end the process; if the MV of the upper block B1 is not available, check whether the MV of the upper left block B2 is available. If the MV of block B2 is available, record the MV of the upper left block B2 as candidate B, and end the process; if the MV of the upper left block B2 is not available, execute 503.

503, if the detection result of 502 is unavailable, then detect whether the scaled MV of the lower left block A0 or the left block A1 corresponding to the lower left corner of the current block is available; if available, record it as a candidate motion vector in the space, for example, as candidate B , and end the process; if not available, execute 504;

Specifically, select one of the lower left block A0 and the left block A1, detect whether there is an MV in it, and if it exists, adjust the proportion of its MV to obtain the scaled MV, so that the reference image corresponding to the scaled MV is the same as the reference image of the current block The same, and record the scaled MV as candidate B. Wherein, for the selection method of the lower left block A0 and the left block A1, the lower left block A0 may be fixedly selected according to the previous configuration, or the left block A1 may be fixedly selected according to the previous configuration.

504. Detect whether the scaledMV of the upper right block B0 or the upper block B1 corresponding to the upper right corner of the current block is available; if available, record it as a candidate motion vector in the space, for example record as candidate B, and end the process; if not available, execute 505;

Specifically, select one of the upper right block B0 and the upper block B1, detect whether there is an MV in it, and if it exists, adjust the proportion of its MV to obtain the scaled MV, so that the reference image corresponding to the scaled MV is the same as the reference image of the current block The same, and record the scaled MV as candidate B. Wherein, for the selection method of the upper right block B0 and the upper block B1, either the upper right block B0 can be fixedly selected according to the previous configuration, or the upper block B1 can be fixedly selected according to the previous configuration.

505. Detect whether the scaled MV of the upper left block B2 corresponding to the upper left corner of the current block is available; if available, record it as a candidate motion vector in space, for example, as candidate B, and end the process; if not available, end the process.

Specifically, detect whether there is an MV in the upper left block B2, and if so, adjust the proportion of the MV to obtain the scaled MV, so that the reference image corresponding to the scaled MV is the same as the reference image of the current block, and record the scaled MV as a candidate b.

Embodiment six

Fig. 4 is a schematic structural diagram of a motion vector prediction device according to an embodiment of the present invention. As shown in Figure 4, the motion vector prediction device includes a spatial candidate motion vector acquisition module 41, a time domain candidate motion vector acquisition module 42, and is connected with the spatial domain candidate motion vector acquisition module 41 and the temporal domain candidate motion vector acquisition module respectively. Module 42 is connected to a motion vector prediction module 43, wherein:

The spatial candidate motion vector acquisition module 41 includes:

The first detection unit 41a is configured to detect whether the motion vector of the reference block at the first adjacent position of the current block is available, and if available, record it as a spatial candidate motion vector;

The second detection unit 41b is configured to detect whether the motion vector of the reference block at the second adjacent position of the current block is available, and if available, record it as a spatial candidate motion vector; and

The third detection unit 41c is configured to, if the motion vector of the reference block at the second adjacent position of the current block is unavailable, perform a check on the reference block at the first adjacent position and/or at the second adjacent position Check whether the scaled motion vector of the reference block is available, and if available, record it as a spatial candidate motion vector;

The time-domain candidate motion vector acquisition module 42 is configured to acquire and record the time-domain candidate motion vector;

The motion vector prediction module 43 is configured to perform motion vector prediction on the current block according to the recorded space-domain candidate motion vectors and time-domain candidate motion vectors.

The motion vector prediction apparatus of this embodiment executes the motion vector prediction process is the same as the motion vector prediction method of the above embodiment, so it will not be repeated here. The motion vector prediction apparatus in this embodiment is, for example, an encoder or a decoder, or is integrated in an encoder or a decoder. Wherein, the spatial domain candidate motion vector acquisition module 41, the time domain candidate motion vector acquisition module 42 and the motion vector prediction module 43 can all be processors, and the first detection unit 41a, the second detection unit 41b and the third detection unit 41c can be composed of implemented on the same processor or different processors.

According to the motion vector prediction device of this embodiment, in the process of obtaining the candidate spatial motion vector, it is firstly detected whether the motion vectors of the reference blocks at the first adjacent position and the second adjacent position of the current block are available, and when the second When the motion vectors of the adjacent reference blocks are not available, it is checked whether the scaled motion vectors are available, so that in the worst case, three steps are required to obtain the spatial candidate motion vectors. However, in the prior art, in the worst case, four steps need to be performed to obtain the spatial candidate motion vector. Therefore, it can be seen that compared with the prior art, the complexity is reduced, and the flow of performing motion vector prediction in the prior art is reduced, thereby improving the execution efficiency of encoding and decoding.

Further, in the motion vector prediction device of the above embodiment, the reference block at the first adjacent position includes the lower left block and the left block corresponding to the lower left corner position; the reference block at the second adjacent position includes the upper right corner position Corresponding upper right block and upper block.

Further, in the motion vector prediction device of the above embodiment, the first detection unit is configured to detect whether the motion vector of the lower left block corresponding to the lower left corner position is available, and detect the lower left corner position only when it is not available Whether the motion vector of the corresponding left block is available;

The second detection unit is configured to detect whether the motion vector of the upper right block corresponding to the upper right corner position is available, and only if it is unavailable, detect whether the motion vector of the upper right block corresponding to the upper right corner position is available.

Further, in the motion vector prediction apparatus of the above embodiment, the reference block at the second adjacent position further includes the upper left block corresponding to the upper left corner position.

Further, in the motion vector prediction device of the above embodiment, the first detection unit is configured to detect whether the motion vector of the lower left block corresponding to the lower left corner position is available, and detect the lower left corner position only when it is not available Whether the motion vector of the corresponding left block is available;

The second detection unit is used to detect whether the motion vector of the upper right block corresponding to the upper right corner position is available, and only if it is unavailable, detect whether the motion vector of the upper right block corresponding to the upper right corner position is available, and only if it is unavailable When used, it is detected whether the motion vector of the upper left block corresponding to the upper left corner position is available.

Further, in the motion vector prediction device of the above embodiment, the third detection unit is used to select at least one of the reference blocks at the first adjacent position, and/or select at least one of the reference blocks at the second adjacent position With reference to at least one of the blocks, it is detected whether scaled motion vectors are available.

Further, in the motion vector prediction device of the above-mentioned embodiment, the third detection unit is used to determine whether the scale-adjusted motion vector of the left block corresponding to the lower left corner position and/or the upper block corresponding to the upper right corner position is available for testing.

Further, in the motion vector prediction device of the above embodiment, the third detection unit is configured to detect whether the scaled motion vector of the left block corresponding to the lower left corner position is available, and only when it is unavailable, detect whether the scaled motion vector of the upper block corresponding to the upper right corner position is available; or

Detecting whether the scaled motion vector of the upper block corresponding to the upper right corner position is available, and only if not available, detecting whether the scaled motion vector of the left block corresponding to the lower left corner position is available.

Further, in the motion vector prediction device of the above-mentioned embodiment, the third detection unit is used to detect the left block corresponding to the lower left corner position, and/or the upper block corresponding to the upper right corner position, and/or the upper left block corresponding to the upper left corner position Whether the scaled motion vector of the upper left block is available is detected.

Further, in the motion vector prediction device of the above embodiment, the third detection unit is configured to detect whether the scaled motion vector of the left block corresponding to the lower left corner position is available, and only when it is unavailable, detect whether the scaled motion vector of the upper block corresponding to the upper right corner position is available, and only if not available, detecting whether the scaled motion vector of the upper left block corresponding to the upper left corner position is available; or

Detecting whether the scaled motion vector of the upper block corresponding to the upper right corner position is available, and only if not available, detecting whether the scaled motion vector of the left block corresponding to the lower left corner position is available, and only When not available, it is detected whether the scaled motion vector of the upper left block corresponding to the upper left corner position is available.

Embodiment seven

An embodiment of the present invention provides an encoder, including the apparatus for predicting a motion vector in the foregoing embodiment.

The flow of motion vector prediction performed by the encoder in this embodiment is the same as that of the motion vector prediction apparatus in the above embodiment, so it will not be repeated here.

In the encoder of this embodiment, when the encoding end of the motion vector prediction module of the motion vector prediction device performs encoding, it selects an optimal motion vector from the motion vector list according to preset rules as the motion vector of the current block The predicted value, where the best one is selected, may be the best in the sense of rate-distortion (Rate-distortion), or may be selected based on other performance indicators as required, which is not limited herein. Wherein, the motion vector list includes the recorded spatial candidate motion vectors and temporal candidate motion vectors, and each spatial candidate motion vector and temporal candidate motion vector has a position index number in the motion vector list.

The encoder may also include an encoding and sending device for writing the position index number of the selected motion vector in the list into the code stream and sending the code stream.

According to the encoder of this embodiment, in the process of obtaining the motion vector candidates in the spatial domain, it is firstly detected whether the motion vectors of the reference blocks at the first adjacent position and the second adjacent position of the current block are available, and when the second adjacent When the motion vector of the reference block at the location is not available, it is checked whether the scaled motion vector is available, so that in the worst case, three steps are required to obtain the spatial candidate motion vector. However, in the prior art, in the worst case, four steps need to be performed to obtain the spatial candidate motion vector. Therefore, it can be seen that compared with the prior art, the complexity is reduced, and the flow of performing motion vector prediction in the prior art is reduced, thereby improving the execution efficiency of encoding.

Embodiment eight

An embodiment of the present invention provides an encoder, including the apparatus for predicting a motion vector in the foregoing embodiment.

The flow of motion vector prediction performed by the encoder in this embodiment is the same as that of the motion vector prediction apparatus in the above embodiment, so it will not be repeated here.

The decoder in this embodiment may further include a receiving and decoding device, which receives the code stream sent by the encoder, decodes the code stream, obtains the motion vector index number of the current block, and provides it to the motion vector prediction device. The motion vector prediction device obtains the space candidate motion vector and the time domain candidate motion vector of the current block, generates a motion vector list including the space domain candidate motion vector and the time domain candidate motion vector, selects and receives the motion vector provided by the decoding device from the motion vector list The candidate motion vector corresponding to the motion vector index number is used as the motion vector predictor of the current block.

According to the decoder of this embodiment, in the process of obtaining the candidate motion vector in the spatial domain, it is firstly detected whether the motion vectors of the reference blocks at the first adjacent position and the second adjacent position of the current block are available, and when the second adjacent When the motion vector of the reference block at the location is not available, it is checked whether the scaled motion vector is available, so that in the worst case, three steps are required to obtain the spatial candidate motion vector. However, in the prior art, in the worst case, four steps need to be performed to obtain the spatial candidate motion vector. Therefore, it can be seen that compared with the prior art, the complexity is reduced, and the flow of performing motion vector prediction in the prior art is reduced, thereby improving the execution efficiency of decoding.

Embodiment nine

FIG. 5 is a system architecture diagram of a codec system provided by an embodiment of the present invention. As shown in FIG. 5 , the codec system includes the encoder 51 of the above embodiment and the decoder 52 of the above embodiment.

The encoding and decoding process of the encoding and decoding system in this embodiment is the same as that of the encoder and decoder in the above embodiments, so it will not be repeated here.

According to the encoding and decoding system of this embodiment, in the process of obtaining the motion vector candidates in the spatial domain, it is firstly detected whether the motion vectors of the reference blocks at the first adjacent position and the second adjacent position of the current block are available, and when the second phase When the motion vector of the adjacent reference block is not available, it is checked whether the scaled motion vector is available, so that in the worst case, three steps are required to obtain the spatial candidate motion vector. However, in the prior art, in the worst case, four steps need to be performed to obtain the spatial candidate motion vector. Therefore, it can be seen that compared with the prior art, the complexity is reduced, and the flow of performing motion vector prediction in the prior art is reduced, thereby improving the execution efficiency of encoding and decoding.

Those of ordinary skill in the art can understand that all or part of the steps for realizing the above-mentioned method embodiments can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, the It includes the steps of the above method embodiments; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk and other various media that can store program codes.

Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: it can still be Modifications are made to the technical solutions described in the foregoing embodiments, or equivalent replacements are made to some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims (23)

1. a method of motion vector prediction, is characterized in that, comprising:

Whether the motion vector detecting the reference block of the first adjacent position of current block can be used, if available, is then recorded as spatial domain candidate motion vector;

Whether the motion vector detecting the reference block of the second adjacent position of current block can be used, if available, is then recorded as spatial domain candidate motion vector;

If the motion vector of the reference block of the second adjacent position of described current block is unavailable, then to the reference block of described first adjacent position and/or the reference block of described second adjacent position, whether can with detecting through the motion vector of proportion adjustment, if available, be then recorded as spatial domain candidate motion vector;

Obtain and record time domain candidate motion vector;

According to recorded spatial domain candidate motion vector and time domain candidate motion vector, motion-vector prediction is carried out to described current block;

Wherein, the first-phase ortho position of described current block is set to the position, the lower left corner of described current block, the second-phase ortho position of described current block is set to the position, the upper right corner of described current block, or the second-phase ortho position of described current block is set to position, the upper right corner and the upper left position of described current block;

Described motion vector can with referring to that described motion vector exists, and reference picture corresponding to described motion vector is identical with the reference picture of described current block.

2. method of motion vector prediction according to claim 1, is characterized in that, the reference block of described first adjacent position comprises lower-left block corresponding to position, the lower left corner and left side block; The reference block of described second adjacent position comprises upper right block corresponding to position, the upper right corner and top block.

3. method of motion vector prediction according to claim 2, is characterized in that, whether the motion vector of the reference block of the first adjacent position of described detection current block can with comprising:

Whether the motion vector detecting lower-left block corresponding to position, the described lower left corner can be used, and only when unavailable, whether the motion vector detecting left side block corresponding to position, the described lower left corner can be used;

Correspondingly, whether the motion vector of the reference block of the second adjacent position of described detection current block can with comprising:

Whether the motion vector detecting upper right block corresponding to position, the described upper right corner can be used, and only when unavailable, whether the motion vector detecting top block corresponding to position, the described upper right corner can be used.

4. method of motion vector prediction according to claim 2, is characterized in that, the reference block of described second adjacent position also comprises upper left block corresponding to upper left position.

5. method of motion vector prediction according to claim 4, is characterized in that, whether the motion vector of the reference block of the first adjacent position of described detection current block can with comprising:

Whether the motion vector detecting lower-left block corresponding to position, the described lower left corner can be used, and only when unavailable, whether the motion vector detecting left side block corresponding to position, the described lower left corner can be used;

Correspondingly, whether the motion vector of the reference block of the second adjacent position of described detection current block can with comprising:

Whether the motion vector detecting upper right block corresponding to position, the described upper right corner can be used, and only when unavailable, whether the motion vector detecting top block corresponding to position, the described upper right corner can be used, and only when unavailable, whether the motion vector detecting upper left block corresponding to described upper left position can be used.

6. according to described method of motion vector prediction arbitrary in claim 1-5, it is characterized in that, the reference block of the described reference block to described first adjacent position and/or described second adjacent position, whether can comprise with carrying out detection through the motion vector of proportion adjustment:

Select at least one in the reference block of described first adjacent position, and/or select at least one in the reference block of described second adjacent position, whether the motion vector detected through proportion adjustment can be used.

7. method of motion vector prediction according to claim 6, it is characterized in that, at least one in the reference block of described first adjacent position of described selection, and/or select at least one in the reference block of described second adjacent position, whether the motion vector detected through proportion adjustment can with comprising:

Select the left side block that position, the lower left corner is corresponding, and/or the top block that position, the upper right corner is corresponding, whether the motion vector detected through proportion adjustment can be used.

8. method of motion vector prediction according to claim 7, is characterized in that, the left side block that position, the described selection lower left corner is corresponding, and/or the top block that position, the upper right corner is corresponding, and whether the motion vector detected through proportion adjustment can with comprising:

Whether the motion vector after proportion adjustment detecting left side block corresponding to position, the described lower left corner can be used, and only when unavailable, whether the motion vector after proportion adjustment detecting top block corresponding to position, the described upper right corner can be used; Or

Whether the motion vector after proportion adjustment detecting top block corresponding to position, the described upper right corner can be used, and only when unavailable, whether the motion vector after proportion adjustment detecting left side block corresponding to position, the described lower left corner can be used.

9. method of motion vector prediction according to claim 6, it is characterized in that, at least one in the reference block of described first adjacent position of described selection, and/or select at least one in the reference block of described second adjacent position, whether the motion vector detected through proportion adjustment can with comprising:

Select the left side block that position, the lower left corner is corresponding, and/or the top block that position, the upper right corner is corresponding, and/or the upper left block that upper left position is corresponding, whether the motion vector detected through proportion adjustment can be used.

10. method of motion vector prediction according to claim 9, is characterized in that, the left side block that position, the described selection lower left corner is corresponding, and/or the top block that position, the upper right corner is corresponding, and/or the upper left block that upper left position is corresponding, whether the motion vector detected through proportion adjustment can be used, and comprising:

Whether the motion vector after proportion adjustment detecting left side block corresponding to position, the described lower left corner can be used, and only when unavailable, whether the motion vector after proportion adjustment detecting top block corresponding to position, the described upper right corner can be used, and only when unavailable, whether the motion vector after proportion adjustment detecting upper left block corresponding to described upper left position can be used; Or

Whether the motion vector after proportion adjustment detecting top block corresponding to position, the described upper right corner can be used, and only when unavailable, whether the motion vector after proportion adjustment detecting left side block corresponding to position, the described lower left corner can be used, and only when unavailable, whether the motion vector after proportion adjustment detecting upper left block corresponding to described upper left position can be used.

11. 1 kinds of motion-vector prediction devices, it is characterized in that, comprise spatial domain candidate motion vector acquisition module, time domain candidate motion vector acquisition module, the motion-vector prediction module be connected with described spatial domain candidate motion vector acquisition module and described time domain candidate motion vector acquisition module respectively, wherein:

Described spatial domain candidate motion vector acquisition module comprises:

Whether the first detecting unit, can use for the motion vector detecting the reference block of the first adjacent position of current block, if available, then be recorded as spatial domain candidate motion vector;

Whether the second detecting unit, can use for the motion vector detecting the reference block of the second adjacent position of current block, if available, then be recorded as spatial domain candidate motion vector; And

3rd detecting unit, if unavailable for the motion vector of the reference block of the second adjacent position of described current block, then to the reference block of described first adjacent position and/or the reference block of described second adjacent position, whether can with detecting through the motion vector of proportion adjustment, if available, then be recorded as spatial domain candidate motion vector, wherein, described 3rd detecting unit performs detection when described first detecting unit and described second detecting unit have performed and detected and there is no two spatial domain candidate motion vectors;

Described time domain candidate motion vector acquisition module, for obtaining and recording time domain candidate motion vector;

Described motion-vector prediction module, for according to recorded spatial domain candidate motion vector and time domain candidate motion vector, carries out motion-vector prediction to described current block;

Wherein, the first-phase ortho position of described current block is set to the position, the lower left corner of described current block, the second-phase ortho position of described current block is set to the position, the upper right corner of described current block, or the second-phase ortho position of described current block is set to position, the upper right corner and the upper left position of described current block;

Described motion vector can with referring to that described motion vector exists, and reference picture corresponding to described motion vector is identical with the reference picture of described current block.

12. motion-vector prediction devices according to claim 11, is characterized in that, the reference block of described first adjacent position comprises lower-left block corresponding to position, the lower left corner and left side block; The reference block of described second adjacent position comprises upper right block corresponding to position, the upper right corner and top block.

13. motion-vector prediction devices according to claim 12, it is characterized in that, whether described first detecting unit can be used for the motion vector detecting lower-left block corresponding to position, the described lower left corner, and only when unavailable, whether the motion vector detecting left side block corresponding to position, the described lower left corner can be used;

Whether described second detecting unit can be used for the motion vector detecting upper right block corresponding to position, the described upper right corner, and only when unavailable, whether the motion vector detecting top block corresponding to position, the described upper right corner can be used.

14. motion-vector prediction devices according to claim 12, is characterized in that, the reference block of described second adjacent position also comprises upper left block corresponding to upper left position.

15. motion-vector prediction devices according to claim 14, it is characterized in that, whether described first detecting unit can be used for the motion vector detecting lower-left block corresponding to position, the described lower left corner, and only when unavailable, whether the motion vector detecting left side block corresponding to position, the described lower left corner can be used;

Whether described second detecting unit can be used for the motion vector detecting upper right block corresponding to position, the described upper right corner, and only when unavailable, whether the motion vector detecting top block corresponding to position, the described upper right corner can be used, and only when unavailable, whether the motion vector detecting upper left block corresponding to described upper left position can be used.

16. according to the arbitrary described motion-vector prediction device of claim 11-15, it is characterized in that, described 3rd detecting unit is for selecting at least one in the reference block of described first adjacent position, and/or at least one selecting in the reference block of described second adjacent position, whether the motion vector detected through proportion adjustment can be used.

17. motion-vector prediction devices according to claim 16, it is characterized in that, described 3rd detecting unit is used for the left side block corresponding to position, the lower left corner, and/or top block corresponding to position, the upper right corner, whether can with detecting through the motion vector of proportion adjustment.

18. motion-vector prediction devices according to claim 17, it is characterized in that, whether described 3rd detecting unit can be used for the motion vector after proportion adjustment detecting left side block corresponding to position, the described lower left corner, and only when unavailable, whether the motion vector after proportion adjustment detecting top block corresponding to position, the described upper right corner can be used; Or

Whether the motion vector after proportion adjustment detecting top block corresponding to position, the described upper right corner can be used, and only when unavailable, whether the motion vector after proportion adjustment detecting left side block corresponding to position, the described lower left corner can be used.

19. motion-vector prediction devices according to claim 16, it is characterized in that, described 3rd detecting unit is used for the left side block corresponding to position, the lower left corner, and/or the top block that position, the upper right corner is corresponding, and/or upper left block corresponding to upper left position, whether can with detecting through the motion vector of proportion adjustment.

20. motion-vector prediction devices according to claim 19, it is characterized in that, whether described 3rd detecting unit can be used for the motion vector after proportion adjustment detecting left side block corresponding to position, the described lower left corner, and only when unavailable, whether the motion vector after proportion adjustment detecting top block corresponding to position, the described upper right corner can be used, and only when unavailable, whether the motion vector after proportion adjustment detecting upper left block corresponding to described upper left position can be used; Or

Whether the motion vector after proportion adjustment detecting top block corresponding to position, the described upper right corner can be used, and only when unavailable, whether the motion vector after proportion adjustment detecting left side block corresponding to position, the described lower left corner can be used, and only when unavailable, whether the motion vector after proportion adjustment detecting upper left block corresponding to described upper left position can be used.

21. 1 kinds of encoders, is characterized in that, comprise the arbitrary described motion-vector prediction device of claim 11-20.

22. 1 kinds of decoders, is characterized in that, comprise the arbitrary described motion-vector prediction device of claim 11-20.

23. 1 kinds of coding/decoding systems, is characterized in that, comprise encoder according to claim 21, and decoder according to claim 22.