WO2009052743A1

WO2009052743A1 - Method and system for determining corresponding macro block

Info

Publication number: WO2009052743A1
Application number: PCT/CN2008/072691
Authority: WO
Inventors: Sixin Lin; Shan Gao; Yingjia Liu; Jiantong Zhou; Jiali Fu; Qingpeng Xie; Xiaosen Zhang; Lianhuan Xiong
Original assignee: Huawei Technologies Co., Ltd.
Priority date: 2007-10-15
Filing date: 2008-10-15
Publication date: 2009-04-30
Also published as: CN101415116A; CN101415116B

Abstract

A method and system for determining the corresponding macro block are disclosed, wherein, the method includes: determining, in one or more areas into which the image to be encoded is divided according to disparity or depth information equivalent to the disparity of each macro block, disparity or depth information of an area in which the current macro block to be encoded locates, determining the corresponding macro block in reference image according to disparity or depth information of the area in which the macro block to be encoded locates. The present invention divides the image to be encoded into a plurality of areas according to disparityor depth information, and determines the corresponding macro block according to disparity or depth information of the area in which the macro block to be encoded locates, so that the obtained disparity or depth information is more precise, and the corresponding macro block which is found in reference image is more accurate. When the subsequent motion compensation is performed by using motion information of the corresponding macro block as motion information of the current macro block to be encoded, it is possible to obtain less compensation residual error, and thereby it is possible to improve image coding efficiency.

Description

TECHNICAL FIELD The present invention relates to video codec technology, and in particular, to a method and system for determining a corresponding macroblock. BACKGROUND OF THE INVENTION In the currently widely used hybrid coding techniques, techniques such as intra prediction, inter prediction, macroblock division, motion estimation, motion compensation, frequency domain transform, loop filtering, quantization, and entropy coding are widely used. It is used to remove redundant information in the video to achieve video compression. Joint video coding group

(JVT) In the multi-view coding (MVC) standard being developed, the motion information skip mode is one of techniques for reducing the amount of information encoded and improving coding efficiency by using motion information between viewpoints. The motion information skip mode utilizes the high similarity of the motion of the same object in the adjacent viewpoints, and uses the motion information in the adjacent viewpoint as the motion information of the current viewpoint, that is, for the currently coded macroblock of the encoding, Determining, in the encoded reference image, a corresponding macroblock that is most similar to the motion information of the current coded macroblock, and using the motion information of the corresponding macroblock as the motion information of the current coded macroblock, as a subsequent motion compensation, that is, calculating the current Encoding a residual of the macroblock and the motion information, and calculating whether an overhead of using the motion mode is smaller than other macroblock mode overhead, and if yes, indicating that the motion information skip mode is an optimal coding mode of the macroblock, determining motion information The skip mode is the final encoding mode of the macroblock. In this way, when encoding, it is only necessary to encode the mode information and the calculated residuals of each macroblock without encoding the motion information of the macroblock, thereby improving the compression efficiency of the MVC.

It can be seen that in the motion mode skip information, it is an important process to determine the corresponding macroblock. It is determined in the prior art that the corresponding macroblock is performed using Global Vision Difference (GDV). The global viewpoint difference is the global viewpoint difference between the current encoded image and the reference image. FIG. 1 is a schematic diagram of determining a corresponding macroblock in the prior art. In the prior art, determining that a corresponding macroblock is a 16×16 partitioning block is a basic unit, and the specific method includes: calculating GDVs of images A and B. The images A and B are anchor frames, ie, the global view differences GDV _A and GDV _{B of the} anchor frame A and the anchor frame B are calculated; the global view difference of the current coded image Cur is determined using the determined GDV of the anchor frame. The current encoded image Cur is a non-anchor frame, POC _cur - POC _A

GDV _cur = GDV _{A +} x(GDV _B - GDV _A ) , where POC , poc _n

POC _B -POC _A _u r is an image sequence number of all images of the same time coordinate of the anchor frame A, the anchor frame B and the non-anchor frame Cur in a set of multi-view video respectively; according to the obtained, the image sequence number can be determined The corresponding macroblock 当前 of the currently encoded macroblock MB in the non-anchor frame Cur in the reference image, the macroblock in the reference image of the current image Cur pointed by the GDV _cur is the corresponding macroblock ΜΒ. After determining the corresponding macroblock ,, the motion information of the corresponding macroblock 作为 is used as the motion information of the current coded macroblock , for subsequent motion compensation.

However, since the corresponding macroblock determined in the prior art is determined by the GDV, the GDV can only approximate the viewpoint difference of the dominant region in the current coded image, and cannot accurately reflect the viewpoint of each macroblock in the current coded image. Poor, that is to say, the difference between the viewpoint difference of some macroblocks and GDV is very large, which inevitably makes the corresponding macroblock pointed to by GDV not the optimal matching macroblock, and the motion information of the corresponding macroblock is taken as the current When encoding the motion information of a macroblock, the coding efficiency is greatly reduced. Summary of the invention

Embodiments of the present invention provide a method and system for determining a corresponding macroblock in order to improve image coding efficiency.

A method of determining a corresponding macroblock, the method comprising:

Determining, according to the viewpoint difference of each macroblock or the depth information equivalent to the viewpoint difference, the viewpoint difference or depth information of the region where the current macroblock is located;

And determining corresponding macroblocks in the reference image according to the view difference or depth information of the region where the currently coded macroblock is located.

A system for determining a corresponding macroblock, the system comprising: a regional view difference or depth information determining unit and a corresponding macroblock determining unit;

a regional view difference or depth information determining unit, configured to determine a region in which the currently coded macroblock is located in one or more regions in which the coded image is divided according to a view difference of each macroblock or depth information equivalent to the view difference Field of view difference or depth information;

And a corresponding macroblock determining unit, configured to determine a corresponding macroblock in the reference image according to the view difference or depth information of the region where the current coded macroblock is determined by the region view difference or depth information determining unit.

It can be seen from the above technical solution that the method and system provided by the embodiments of the present invention determine the viewpoint difference of the area where the current coded macroblock is located, by determining the coded image according to the view difference or depth information of each macroblock. The depth information determines a corresponding macroblock in the reference image according to the view difference or depth information of the region where the macroblock is encoded. That is, the coded image is divided into a plurality of regions according to the view difference or the depth information, and the corresponding macroblock is determined by the view difference or depth information of the region where the current coded macroblock is located, instead of determining the global view difference of the reference image. Corresponding to the macroblock, the obtained visual difference or depth information is more accurate, and the corresponding macroblock found in the reference view is also more accurate, and the motion information of the corresponding macroblock is used as the motion compensation of the motion information of the current coded macroblock. At the same time, it is possible to obtain a smaller compensation residual, thereby improving image coding efficiency. DRAWINGS

1 is a schematic diagram of determining a corresponding macroblock in the prior art;

2 is a main flowchart of a method according to an embodiment of the present invention;

FIG. 3 is a schematic diagram of a method for dividing an area according to an embodiment of the present invention;

4 is a flow chart of a method for determining a viewpoint difference of a region where a current coded macroblock is located according to an embodiment of the present invention;

FIG. 5 is a structural diagram of a system according to an embodiment of the present invention;

FIG. 6 is a structural diagram of another system according to an embodiment of the present invention. detailed description

The present invention will be described in detail below with reference to the drawings and specific embodiments.

Since the depth information and the viewpoint difference can be converted to each other, the depth information and the viewpoint difference are equivalent techniques. The embodiment of the present invention will be specifically described below by taking the case of using the viewpoint difference as an example, and the related description is equally applicable to the case of using the depth information.

Preferably, according to the principle of aperture imaging, the embodiment of the present invention provides the following method for converting depth information and viewpoint difference:

Where / is the camera pitch, / is the camera focal length, and for the viewpoint difference, Ζ is the depth information equivalent to the viewpoint difference. However, the depth information and the method of converting the viewpoint difference included in the embodiment of the present invention are not limited thereto, and any related or similar depth information and a method of converting the viewpoint difference are also included.

FIG. 2 is a main flowchart of a method according to an embodiment of the present invention. As shown in FIG. 2, the method mainly includes:

Step 201: Determine a viewpoint difference of a region where the current coded macroblock is located in one or more regions in which the coded image is divided according to the difference of the viewpoint of each macroblock.

Step 202: Determine a corresponding macroblock in the reference image according to a view difference of a region where the currently coded macroblock is located.

In the above process, the corresponding macroblock in the reference image is determined according to the view difference of the area where each macroblock in the encoded image is located, so that only the area difference information of the previously decoded macroblock needs to be used to derive the area information of each macroblock, instead of It is necessary to separately transmit the viewpoint difference information of each macroblock in the code stream, thereby saving the code stream.

Wherein, in step 201, there may be different ways of dividing the area, and the coded image may be divided into different areas according to the viewpoint difference of each macro block in advance. The method for determining the corresponding macroblock at this time is specifically as follows: the image may be pre-coded first, and the difference of the viewpoints of each macroblock is obtained; according to the obtained viewpoint difference of each macroblock, the encoded image is divided into different regions, and each macro in each region The absolute value of the difference between the viewpoint difference of the block and the viewpoint difference of the region is within a set range, and the area identifier and the regional viewpoint difference where each macroblock is located are recorded; when each macroblock is encoded, according to the area where the macroblock is located Identification, find the corresponding macroblock in the reference image with the difference of the viewpoint of the area in which it is located.

Step 201 may also determine, according to the coding order, the range of regions to which the currently coded macroblock belongs, one by one. The method can be as shown in FIG.

FIG. 3 is a schematic diagram of a method for dividing an area according to an embodiment of the present invention. As shown in FIG. 3, the method may include the following steps:

Step 301: Set the current area viewpoint difference.

Before performing encoding of the first macroblock of the encoded image, it is necessary to initialize the current regional viewpoint difference. This step may be initialized before a coded image begins to be encoded. It may also be used during the process of encoding the image. The initialization area is set by the viewpoint difference. When the current coded macroblock is not the first macroblock of the coded picture, the current area view difference is set to the view difference of the area in which the previous inter-view predictive coded macroblock is located.

The area difference of the initial setting can be set according to the actual performance of the image collection device, usually an empirical value, or it can be initially set to 0. The image collection device may be a camera or the like.

In this step, the current regional view difference statistics count value may be further set initially, and is usually initially set to 0, used to count the sequence numbers of the macroblocks in the area where they are located, and may be used in the area where the current coded macro block is located in the subsequent step. The calculation of the viewpoint difference.

Step 302: Determine a view difference of a macroblock of the previous coded macroblock that utilizes inter-view predictive coding.

For convenience of description, the former macroblock using inter-view predictive coding is simply referred to as the previous macroblock.

If the previous macroblock is encoded using the inter-view prediction technique, the view difference obtained by the inter-view predictive coding is sufficiently accurate, and the next coded macroblock also has a similar viewpoint difference, therefore, the two Macroblocks are likely to belong to the same area.

The viewpoint difference of the previous macroblock using inter-view predictive coding may be determined according to different coding division modes of the macroblock, for example, when the previous macroblock is divided by 16×16 partitioned blocks, the previous macro The view difference of the block may be the view difference of the 16×16 partition block; when the previous macro block is divided by the 16×8 partition block, the view difference of the previous macro block may be the two views belonging to the same view. The viewpoint difference of the 16×8 partitioning block is averaged to generate a viewpoint difference of the previous macroblock, where, The averaging process may obtain the viewpoint difference of the previous macroblock by using the average or the median of the viewpoint differences of the two blocks; when the previous macroblock is 8×16, 8×8, When the mode division is performed such as 8 x 4, 4 x 8, and 4 4 division blocks, the viewpoint difference of the previous macro block may be determined by synthesizing the viewpoint differences of the plurality of blocks belonging to the same view.

Step 303: Determine the regional point difference of the current coded macroblock by using the determined difference between the view difference of the previous macroblock and the current regional view difference.

Step 304: Determine, by using the determined regional view difference of the currently coded macroblock, the macroblock pointed by the determined regional view difference as a corresponding macroblock in the reference image.

In this step, the determined corresponding macroblock may be either a decoded macroblock in the reference image or a synthesized macroblock of the decoded macroblock in the reference image.

In addition, in this step, after determining the macroblock pointed to by the regional view difference, the search may be performed in a certain area centered on the pointed macroblock, and the macroblock which makes the coding efficiency optimal is searched as the corresponding macroblock. The coding efficiency is optimal, and the rate distortion is optimal (RD0). It is a prior art to search for macroblocks with the most coding efficiency in a certain area, and details are not described herein.

The method for determining the spot difference of the region in which the current coded macroblock is located in the above step 303 may be as shown in FIG. 4, and may include the following steps:

Step 401: Determine whether the absolute value of the difference between the viewpoint difference of the previous macroblock and the current regional viewpoint difference is greater than a set threshold. If yes, go to step 402; if no, go to step 403.

Step 402: Update the current regional view difference to the view difference of the previous macro block, and perform the steps.

404.

When the absolute value of the difference between the viewpoint difference of the current macroblock and the current regional viewpoint difference is greater than the set threshold, the difference between the viewpoint difference of the previous macroblock and the current region is larger, and the current region is larger. The viewpoint difference does not accurately reflect the current regional viewpoint difference. Therefore, it is necessary to update the current regional viewpoint difference and re-count the count value of the regional viewpoint difference statistics.

The step may further include: zeroing the current regional view difference statistics count value.

Step 403: Average the current regional view difference and the previous macroblock's viewpoint difference, and obtain The average value is used as the updated current regional viewpoint difference, and step 404 is performed.

The method for performing the averaging process in this step may be taking the median value of the current regional view difference and the previous macroblock, or using the current regional view difference statistical count value, the current regional view difference and the previous macroblock. The weighted average of the viewpoint differences is also used, and it is also possible to use an algebraic averaging method such as averaging. The weighted average algorithm may be: multiplying the current regional viewpoint difference by the current regional viewpoint difference statistical count value, plus the sum of the viewpoint differences of the previous macroblock, and dividing by the current regional viewpoint difference count value and 1 with. which is:

RDV = ( DV dvNum + mbDV ) / ( dvNum + 1 )

Where RDV is the updated current regional view difference, DV is the current regional view difference, dvN is the current regional view difference count value, and mbDV is the previous macroblock's view difference. In the specific calculation, the calculation of the horizontal component and the vertical component can be performed separately for each viewpoint difference. It is also possible to round the calculation results according to different calculation accuracy.

The step further includes continuously counting the current regional view difference statistical count value.

Step 404: The updated current regional view difference is used as the regional view difference of the current coded macroblock.

In addition, before step 404, it may be determined whether the updated region view difference is 0. If 0, the obtained updated region view difference may be added to the set offset. This offset can be set according to the performance of the video capture device.

The flow shown in Figure 4 above can be expressed in the form of pseudo code as follows:

If (the absolute value of the previous macroblock's viewpoint difference mbDV and the current area viewpoint difference DV is greater than the threshold DV-THRESH) Update the current area viewpoint difference (DV): DV.set( mbDV.getHor(), mbDV.getVer( )); dvNum = 0; else Update the horizontal component of the current regional view difference (DV):

Let temp Value be the sum of the product of the current regional view difference DV horizontal component and the regional view difference count value dvNum and the current macroblock visual difference mbDV horizontal component:

Tem Value= (DV.getHor() * dvNum + mbDV.getHorQ);

Set rem to the absolute value of temp Value divided by the remainder of (dvNum + 1);

If temp Value is non-negative, then sgn takes 1 or takes -1;

If(rem is greater than half of (dvNum+ 1 )) Let tempO be tempValue/(dvNum + 1) + sgn; else tem O = temp Value/ (dvNum + 1); Let the horizontal component of D V be tempO ;

Similarly, the vertical component of the current regional viewpoint difference (DV) can be obtained;

If (the horizontal and vertical components of the current regional view difference (DV) are equal to 0) Let D V be mbDV plus a certain offset

dvNum counting force port 1

In the above pseudo code, mbDV.ge tHor() represents the horizontal difference component of the previous macroblock, mbDV.ge tVer 0 represents the vertical component of the viewpoint difference of the previous macroblock, and DV. ge tHor () represents the current regional viewpoint difference level. Component.

Regardless of which of the above-mentioned area division methods is used, after determining the area where the current coded macroblock is located, Marking the region where the current coded macroblock is located, transmitting the regional view difference of each region on the slice header, and using the view difference of the region where the current coded macroblock is located, finding the corresponding macroblock in the reference image, and using the motion information of the corresponding macroblock as the current The motion information of the encoded macroblock is used for subsequent motion compensation.

According to the above method, it is no longer necessary to transmit the original GDV in the code stream. The original stripe level syntax is as shown in Table 1. After the above method for determining the corresponding macroblock is used, the existing stripe level syntax can be as follows. 2 is shown.

Table 1

The following is a detailed description of a system for determining a corresponding macroblock provided by an embodiment of the present invention, and FIG. 5 is a The system structure diagram provided by the embodiment, as shown in FIG. 5, may include: a region view difference or depth information determining unit 500 and a corresponding macro block determining unit 510.

The area view difference or depth information determining unit 500 is configured to determine the view difference or depth information of the area where the current coded macro block is located in one or more areas in which the coded picture is divided according to the view difference or depth information of each macro block.

The corresponding macroblock determining unit 510 is configured to determine a corresponding macroblock in the reference image according to the view difference or depth information of the region where the currently coded macroblock is determined by the regional view difference or depth information determining unit 500.

The system may further include: a reference image storage unit 520 for storing the reference image.

The corresponding macroblock determining unit 510 is configured to determine a corresponding macroblock of the currently coded macroblock in the reference image stored in the reference image storage unit 520 according to the view difference or depth information of the region where the currently coded macroblock is located.

The regional view difference or depth information determining unit 500 may include: a precoding unit 501, a region dividing unit 502, a recording unit 503, and a first determining unit 504.

The precoding unit 501 is configured to precode the coded image to obtain view difference or depth information of each macroblock in the coded image.

The area dividing unit 502 is configured to divide each macroblock in the coded image into different regions according to the view difference or depth information of each macroblock acquired by the precoding unit 501, so that the view difference or depth information of each macroblock is located The absolute value of the difference between the viewpoint or depth information of the area is within the set range.

The recording unit 503 is configured to record the division result of the area dividing unit 502, and record the area identifier and the area viewpoint difference or depth information where each macro block is located.

The first determining unit 504 is configured to search, in the recording unit 503, the view difference or depth information of the area where the current coded macro block is located, according to the area identifier where the currently coded macro block is located.

The structure of the regional view difference or depth information determining unit 500 can be used to pre-divide the regions of the coded image. When the method of determining the range of the region to which the currently coded macroblock belongs is determined one by one according to the coding order, In the structure shown in FIG. 6, the area view difference or depth information determining unit 500 may include: a front macroblock view difference or depth information determining unit 601, a difference obtaining unit 602, and a second determining unit 603. The pre-macroblock view difference or depth information determining unit 601 is configured to determine a view difference or depth information of a macroblock of the current coded macroblock using the inter-view predictive coding.

The difference obtaining unit 602 is configured to determine an absolute value of the difference between the view difference or depth information determined by the previous macroblock view difference or depth information determining unit 601 and the set current region view difference or depth information.

The second determining unit 603 is configured to determine, according to the absolute value of the difference determined by the difference acquiring unit 602, the area difference or depth information of the area where the currently coded macro block is located.

The second determining unit 603 may include: a determining unit 6031, an average processing unit 6032, and a third determining unit 6033.

The determining unit 6031 is configured to determine whether the absolute value of the difference determined by the difference obtaining unit 602 is greater than a set threshold, and if yes, send a determination notification to the third determining unit 6033, and if not, send the processing to the average processing unit 6032. Notice.

The average value processing unit 6032 is configured to, when receiving the processing notification of the determining unit 6031, average the current regional view difference or depth information and the previous view difference or depth information of the macroblock using the inter-view predictive coding, The average value obtained is sent to the third determining unit 6033.

The averaging process performed by the average processing unit 6032 may be: obtaining an algebraic mean, finding a median, finding a weighted average, or other means of averaging.

The third determining unit 6033 is configured to, when receiving the determining notification of the determining unit 6031, set the current regional view difference or depth information update to the previous use of the inter-view predictive coding determined by the previous macroblock view difference or depth information determining unit 601. The viewpoint difference or depth information of the macroblock; when the average value provided by the average value processing unit 6032 is received, the current region view difference or depth information update is set to the average value.

The corresponding macroblock determining unit 510 may include: a pointing macroblock determining unit 511 and a searching unit.

512.

The pointing macroblock determining unit 511 is configured to determine a macroblock pointed by the view difference or depth information of the region where the currently coded macroblock is located in the reference image.

The search unit 512 is configured to search for a macroblock that makes encoding efficiency optimal as the corresponding macroblock in a set region centered on the macroblock determined by the macroblock determining unit 511. Due to the inter-convertibility of the viewpoint difference and the depth, the depth difference information can be replaced by the depth information correspondingly.

It can be seen from the above description that the method and system provided by the embodiments of the present invention determine the viewpoint difference or depth of the region where the current coded macroblock is located in one or more regions in which the coded image is divided according to the view difference or depth information of each macroblock. The information determines a corresponding macroblock in the reference image according to the view difference or depth information of the region where the macroblock is encoded. That is to say, the coded image is divided into multiple regions according to the viewpoint difference, and the corresponding macroblock is determined by the view difference or depth information of the region where the currently coded macroblock is located, instead of using the global view difference of the reference image to determine the corresponding macroblock. , so that the obtained viewpoint difference or depth information is more accurate, and the corresponding macroblock found in the reference viewpoint is also more accurate, and the motion information of the corresponding macroblock is used as the motion compensation of the current encoded macroblock for subsequent motion compensation. Get smaller compensation residuals to improve image coding efficiency.

In addition, when determining the view difference of the region where the macroblock is encoded in the embodiment of the present invention, it may be determined according to the difference of the viewpoint of the previous macroblock in the coded image, without using the coded image as in the prior art. By viewing the difference information, it is not necessary to transmit the viewpoint difference information between the images in the code stream, so that the error resistance performance is also greatly improved.

Due to the inter-convertibility of the viewpoint difference and the depth, the depth difference information can be used instead of the viewpoint information, and the related description is also applicable to the case where the depth information is used.

The above is only the preferred embodiment of the present invention, and is not intended to limit the present invention. Any modifications, equivalents, improvements, etc., which are made within the spirit and principles of the present invention, should be included in the present invention. Within the scope of protection.

Claims

Claim

A method for determining a corresponding macroblock, the method comprising:

2. The method according to claim 1, wherein the obtaining the depth information equivalent to the viewpoint difference comprises at least:

Where / is the camera pitch, / is the camera focal length, and for the viewpoint difference, Ζ is the depth information equivalent to the viewpoint difference.

The method according to claim 2, wherein the determining the view difference or depth information of the region where the current coded macroblock is located in one or more regions in which the coded image is divided according to the view difference or depth information of each macroblock Includes:

Encoding the encoded image to obtain visual difference or depth information of each macro block;

Pre-decoding each macroblock in the coded image into different regions according to the view difference or depth information of each macroblock, so that the absolute value of the difference between the viewpoint difference or depth information of each macroblock and the viewpoint difference or depth information of the region in which the macroblock is located Within the set range, the area identifier and the area view difference or depth information of each macro block are recorded; and the view difference or depth information of the area where the current coded macro block is located is determined according to the area identifier of the current coded macro block.

The method according to claim 2, wherein the determining the coded image is divided into one or more regions according to a view difference of each macroblock or depth information equivalent to the view difference, the current coded macroblock The viewpoint or depth information of the area includes:

Determining a view difference or depth information of a macroblock of the current coded macroblock that utilizes inter-view predictive coding; Determining the area disparity or depth information of the region where the currently coded macroblock is located by using the determined difference between the view point difference or depth information of the macroblock using the inter-view predictive coding and the set current region disparity or depth information.

The method according to claim 4, wherein when the previous macroblock using inter-view prediction coding uses other coding division modes other than the 16×16 partitioning block, the previous utilization viewpoint is used. The view difference or depth information obtained by performing averaging processing on the view difference or depth information of one or more divided blocks of the inter-predictive coded macroblock is determined as the view point of the macroblock of the current coded macroblock that is previously encoded using inter-view prediction. Poor or depth information.

The method according to claim 4, wherein the set current region view difference or depth information is set according to the image collection device performance of the encoded image; or, is the last use The inter-view predicts the viewpoint difference or depth information of the area where the macroblock is encoded.

The method according to claim 4, wherein the determining, by using the determined visual difference or depth information of the macroblock using the inter-view predictive coding, and the set current region disparity or depth information The difference between determining the view point difference or depth information of the current coded macroblock includes: determining the absolute difference between the view point difference or depth information of the previous macroblock using the inter-view predictive coding and the current regional view difference or depth information Whether the value is greater than a set threshold, and if so, setting the current region view difference or depth information update to the view point difference or depth information of the previous macroblock using inter-view predictive coding; if not, the current region The view difference or depth information is averaged with the view difference or depth information of the previous macroblock using the inter-view predictive coding, and the current regional view difference or depth information update is set to the average value obtained after the averaging process. .

8. The method according to claim 7, wherein the method further comprises: initializing a current regional view difference statistical count value;

Setting the current region view difference statistical count value to 0 when the absolute value of the difference between the viewpoint difference of the macroblock using the inter-view predictive coding and the current region view difference is greater than the set threshold value; Determining the current regional view difference and the current coded macroblock when determining that the absolute value of the difference between the view difference of the macroblock using the inter-view predictive coding and the current region view difference is less than the set threshold The averaging of the viewpoint differences of the previous macroblock using the inter-view predictive coding includes: multiplying the current regional view difference by the current regional view difference statistical count value, plus the previous view using the inter-view predictive coding macroblock The sum of the differences is divided by the sum of the current regional view difference count value and 1; the method further includes: accumulating the current region view difference count value by 1; or

Initially setting a current area depth information statistical count value;

Setting the current region depth information statistical count value when the absolute value of the difference between the depth information of the macroblock using the inter-view predictive coding and the current region depth information is greater than the set threshold

0;

And determining, when the absolute value of the difference between the depth information of the macroblock using the inter-view predictive coding and the current region depth information is less than the set threshold, the current region depth information and the current coded macroblock The averaging of the depth information of the previous macroblock using inter-view predictive coding includes: multiplying the current region depth information by the current region depth information statistical count value, and adding the depth information of the previous inter-view predictive coding macroblock And , divided by the current region depth information to count the sum of the count value and 1; the method further includes: accumulating the current region depth information statistics count value by one.

The method according to claim 2, wherein the determining, according to the view difference or depth information of the region where the currently coded macroblock is located, determining the corresponding macroblock in the reference image comprises: using the current coded macroblock The macroblock pointed to by the region or the depth information in the reference image is determined to be a corresponding macroblock; or

Determining a macroblock pointed by the view difference or depth information of the region where the current coded macroblock is located in the reference image, searching for a macroblock that makes the coding efficiency optimal in a set region centered on the pointed macroblock The corresponding macro block.

The method according to claim 9, wherein the determined corresponding macroblock is a decoded macroblock decoded in a reference image, or a synthesized macroblock of a decoded macroblock in a reference image.

11. A system for determining a corresponding macroblock, the system comprising: a regional view difference or depth information determining unit (500) and a corresponding macroblock determining unit (510);

a regional view difference or depth information determining unit (500) for determining that the encoded image is in accordance with each macroblock a view difference or depth information of a region in which the current macroblock is located, in one or more regions of the view difference or the depth information equivalent to the view difference;

The corresponding macroblock determining unit (510) is configured to determine a corresponding macroblock in the reference image according to the view difference or depth information of the region where the currently coded macroblock is determined by the region view difference or depth information determining unit (500).

The system according to claim 11, wherein the obtaining the depth information equivalent to the viewpoint difference comprises at least:

The system according to claim 12, characterized in that the system further comprises: a reference image storage unit (520) for storing the reference image.

The system according to claim 12, wherein the area disparity or depth information determining unit (500) comprises: a precoding unit (501), a region dividing unit (502), and a recording unit.

(503) and a first determining unit (504);

a precoding unit (501), configured to precode the encoded image to obtain view difference or depth information of each macroblock in the encoded image;

a region dividing unit (502), configured to divide each macroblock in the coded image into different regions according to the view difference or depth information of each macroblock acquired by the precoding unit (501), so as to obtain a viewpoint of each macroblock The absolute value of the difference between the difference or depth information of the difference or depth information and the depth information of the region in which it is located is within the set range;

a recording unit (503), configured to record a division result of the area dividing unit (502), and record an area identifier and an area viewpoint difference or depth information where each macro block is located;

The first determining unit (504) is configured to search, in the recording unit (503), the view difference or depth information of the area where the currently coded macro block is located, according to the area identifier where the currently coded macro block is located. The system according to claim 12, wherein the area view difference or depth information determining unit (500) comprises: a front macroblock view difference or depth information determining unit (601), a difference obtaining unit (602) And a second determining unit (603);

a pre-macroblock view difference or depth information determining unit (601) for determining a view difference or depth information of a previous macroblock of the currently coded macroblock using inter-view predictive coding;

a difference obtaining unit (602), configured to determine an absolute value of a difference between the view difference or depth information determined by the previous macroblock view difference or depth information determining unit (601) and the set current region view difference or depth information;

The second determining unit (603) is configured to determine, according to the absolute value of the difference determined by the difference acquiring unit (602), the area difference or depth information of the area where the currently coded macro block is located.

The system according to claim 15, wherein the second determining unit (603) comprises: a determining unit (6031), an average value processing unit (6032), and a third determining unit (6033); (6031), configured to determine whether the absolute value of the difference determined by the difference obtaining unit (602) is greater than a set threshold, and if yes, send a determination notification to the third determining unit (6033), if no, then The average value processing unit (6032) sends a processing notification;

The average value processing unit (6032) is configured to, when receiving the processing notification of the determining unit (6031), compare the current regional view difference or depth information with the view point difference of the previous macroblock using the inter-view predictive coding Or the depth information is averaged, and the obtained average value is sent to the third determining unit (6033);

a third determining unit (6033), configured to: when receiving the determining notification of the determining unit (6031), set the current regional view difference or depth information update to the previous macroblock view difference or depth information determining unit (601) Determining the previous view difference or depth information of the macroblock using the inter-view predictive coding; when receiving the average value provided by the average value processing unit (6032), setting the current regional view difference or depth information update to The average value.

The system according to claim 12, wherein the corresponding macroblock determining unit (510) comprises: a pointing macroblock determining unit (511) and a searching unit (512); a macroblock determining unit (511) for determining a macroblock pointed by the view difference or depth information of the region where the currently coded macroblock is located in the reference image;

The search unit (512) is configured to search for a macroblock that makes encoding efficiency optimal as the corresponding macroblock in a set region in which the macroblock determined by the pointing macroblock determining unit (511) is a center.