CN103024395A - Device and method for multi-view video coding mode selection - Google Patents

Abstract

The invention provides a device and a method for multi-view video coding mode selection. The method includes the steps: inputting a video image and a corresponding depth map, and dividing each frame of the video image into a plurality of basic coding units, namely macroblocks; judging the position of the current coding macroblock, setting and predicating a rate distortion threshold of a SKIP mode and calculating rate distortion cost of the current macroblock selecting the SKIP mode; calculating motion amplitude of the current macroblock according to the rate distortion threshold and the rate distortion cost; and selecting corresponding candidate modes according to motion amplitude levels of the current macroblock. According to the method, the SKIP mode is predicated through the block based adaptive threshold, and a great quantity of unnecessary calculation is avoided. In addition, the motion amplitude of the current macroblock is predicated according to adjacent macroblocks and motion vectors of adjacent viewpoints corresponding to the macroblock, and coincident candidate modes are checked according to motion amplitude levels, so that searching range can be narrowed, and operation time is shortened.

Description

Multiple view video coding mode selecting method and device

Technical field

The present invention relates to technical field of video coding, particularly a kind of multiple view video coding mode selecting method and device.

Background technology

Along with the continuous growth of people to stereoscopic vision perception demand, the multi-view point video technology has attracted the concern of increasing Chinese scholars and industrial quarters.Compare with the ordinary two dimensional video, multi-view point video has increased the depth information of scene, thereby can provide visual presence and realism to the user, and the video data that has still also increased simultaneously at least one times needs to process.Therefore, for the coding of multi-view point video, the prediction mode of introducing disparity estimation removes the redundancy between the viewpoint.

Existing video encoding standard has obvious advantage although H.264 wait aspect code efficiency; But the high computation complexity that they are introduced by the raising compression efficiency has become unaffordable defective in the practical application on the contrary.For example in JMVM model H.264/MVC, in order to obtain higher Video coding efficient, JMVM provides 7 kinds of block sizes (16 * 16,16 * 8,8 * 16,8 * 8,8 * 4,4 * 8, and 4 * 4) to be used for the model selection of motion prediction and parallax prediction.For the macroblock coding of inter frame, candidate pattern in 11: SKIP is arranged, inter_16 * 16, inter_16 * 8, inter_8 * 16, inter_8 * 8, inter_8 * 4, inter_4 * 8, inter_4 * 4, intra_4 * 4, intra_8 * 8, intra_16 * 16, then need the different patterns of cutting apart is carried out respectively estimation when selecting coding mode, utilance aberration optimizing (RDO) calculation code cost relatively also selects minimum coding mode as the coding mode of optimum.This multi-mode estimation and mode decision have very high computation complexity, and its scramble time has accounted for greatly the over half of total processing time.As seen, reasonable efficiently significant for Video coding of mode decision algorithm.

Nowadays, the multi-viewpoint video image of extensive angle generates in order to satisfy more, usually except the input multi-view point video, also can input corresponding deep video when encoding, thereby output generates the synthetic viewpoint video image of more perspective again.Although this coding mode has satisfied wide-angle even holographic video requirement, huge data volume and operand have had very large restriction for the real-time application of video.

At present, many fast schema selection methods based on the single view Video coding are arranged.But, owing to introduced the Forecasting Methodology of disparity estimation in the multi-vision-point encoding framework and introduced the information of the video degree of depth, thereby cause most of existing single view Video coding fast algorithm all not to be suitable for the coding of multi-view point video.

Summary of the invention

Purpose of the present invention is intended to solve at least one of above-mentioned technological deficiency.

For achieving the above object, the embodiment of one aspect of the present invention proposes a kind of multiple view video coding mode selecting method, may further comprise the steps: S1: inputted video image and corresponding depth map, and each frame of described video image is divided into a plurality of basic coding units, i.e. macro block; S2: judge the position of the described macro block of present encoding, the rate distortion threshold value of prediction SKIP pattern is set and calculates current macro and select rate distortion costs under the SKIP pattern; S3: according to the motion amplitude of described rate distortion threshold value and described rate distortion costs calculating current macro; And S4: the described motion amplitude size according to current macro is chosen corresponding candidate pattern.

Method according to the embodiment of the invention, predict the SKIP pattern by block-based adaptive threshold, a large amount of unnecessary calculating have been avoided, and predict the motion amplitude of current macro according to the motion vector of adjacent macroblocks and the corresponding macro block of adjacent viewpoint, the candidate pattern that meets according to the large minor inspection of motion amplitude again can reduce thus the hunting zone, reduce operation time.

In one embodiment of the invention, described step S2 specifically comprises: S21: if current macro place frame is the first row or first row that I frame or current macro are in the place frame, then the utilization rate distortion optimizing method checks inter-frame mode in all frames, and the pattern of selection rate distortion cost minimum is the forced coding pattern; S22: if current macro place viewpoint is the absolute coding viewpoint, the place frame is not the first row or the first row that I frame and current macro are not in the place frame simultaneously, then calculates current macro and selects the first rate distortion costs of SKIP pattern and the first rate distortion threshold value that is used for predicting the SKIP pattern is set; And S23: if current macro place viewpoint is the associated prediction coded views, the place frame is not the first row or the first row that I frame and current macro are not in the place frame simultaneously, then arranges to be used for predicting the second rate distortion threshold value of SKIP pattern and to calculate the second rate distortion costs that current macro is selected the SKIP pattern.

In one embodiment of the invention, described step S3 specifically comprises: more described the first rate distortion costs and described the first rate distortion threshold value; When described the first rate distortion costs during less than described the first rate distortion threshold value, be the forced coding pattern of current macro with the SKIP pattern; And when described the first rate distortion costs during greater than described the first rate distortion threshold value, calculate the first motion amplitude of current macro.

In one embodiment of the invention, described step S3 also comprises: more described the second rate distortion costs and described the second rate distortion threshold value; When described the second rate distortion costs during less than described the second rate distortion threshold value, be the forced coding pattern of current macro with the SKIP pattern; And when described the second rate distortion costs during greater than described the second rate distortion threshold value, calculate the first motion amplitude of current macro.

For achieving the above object, the present invention embodiment on the other hand proposes a kind of multiple view video coding mode selector, comprises the division module, is used for inputted video image and corresponding depth map, and each frame of described video image is divided into a plurality of basic coding units, i.e. macro block; Judge module is set, be used for judging the position of the described macro block of present encoding, the rate distortion threshold value of prediction SKIP pattern is set and calculates current macro and select rate distortion costs under the SKIP pattern; Computing module is used for the motion amplitude according to described rate distortion threshold value and described rate distortion costs calculating current macro; And choose module, be used for choosing corresponding candidate pattern according to the described motion amplitude size of current macro.

Device according to the embodiment of the invention, predict the SKIP pattern by block-based adaptive threshold, a large amount of unnecessary calculating have been avoided, and predict the motion amplitude of current macro according to the motion vector of adjacent macroblocks and the corresponding macro block of adjacent viewpoint, the candidate pattern that meets according to the large minor inspection of motion amplitude again can reduce thus the hunting zone, reduce operation time.

In one embodiment of the invention, described judgement arranges module and specifically comprises: selected cell, being used for working as current macro place frame is the first row or the first row that I frame or current macro are in the place frame, then the utilization rate distortion optimizing method checks inter-frame mode in all frames, and the pattern of selection rate distortion cost minimum is the forced coding pattern; The first computing unit, being used for working as current macro place viewpoint is the absolute coding viewpoint, the place frame is not the first row or the first row that I frame and current macro are not in the place frame simultaneously, then calculates current macro and selects the first rate distortion costs of SKIP pattern and the first rate distortion threshold value that is used for predicting the SKIP pattern is set; And second computing unit, being used for working as current macro place viewpoint is the associated prediction coded views, the place frame is not the first row or the first row that I frame and current macro are not in the place frame simultaneously, then arranges to be used for predicting the second rate distortion threshold value of SKIP pattern and to calculate the second rate distortion costs that current macro is selected the SKIP pattern.

In one embodiment of the invention, described computing module specifically comprises: the first comparing unit is used for more described the first rate distortion costs and described the first rate distortion threshold value; The first processing unit is used for when described the first rate distortion costs during less than described the first rate distortion threshold value, is the forced coding pattern of current macro with the SKIP pattern; And second processing unit, be used for when described the first rate distortion costs during greater than described the first rate distortion threshold value the first motion amplitude of calculating current macro.

In one embodiment of the invention, described computing module specifically also comprises: the second comparing unit is used for more described the second rate distortion costs and described the second rate distortion threshold value; The 3rd processing unit is used for when described the second rate distortion costs during less than described the second rate distortion threshold value, is the forced coding pattern of current macro with the SKIP pattern; And the manage the unit everywhere, is used for when described the second rate distortion costs during greater than described the second rate distortion threshold value the first motion amplitude of calculating current macro.

The aspect that the present invention adds and advantage in the following description part provide, and part will become obviously from the following description, or recognize by practice of the present invention.

Description of drawings

Above-mentioned and/or the additional aspect of the present invention and advantage are from obviously and easily understanding becoming the description of embodiment below in conjunction with accompanying drawing, wherein:

Fig. 1 is the flow chart of multiple view video coding mode selecting method according to an embodiment of the invention;

Fig. 2 is the particular flow sheet of multiple view video coding mode selecting method according to an embodiment of the invention;

Fig. 3 is in the absolute coding viewpoint according to an embodiment of the invention, the location diagram of macro block and its time, space adjacent macroblocks;

Fig. 4 is in the associated prediction coded views according to an embodiment of the invention, the location diagram of adjacent macroblocks between macro block and its time, space, viewpoint; And

Fig. 5 is the frame diagram of multiple view video coding mode selector according to an embodiment of the invention.

Embodiment

The below describes embodiments of the invention in detail, and the example of embodiment is shown in the drawings, and wherein identical or similar label represents identical or similar element or the element with identical or similar functions from start to finish.Be exemplary below by the embodiment that is described with reference to the drawings, only be used for explaining the present invention, and can not be interpreted as limitation of the present invention.

Fig. 1 is the flow chart of the multiple view video coding mode selecting method of the embodiment of the invention.Fig. 2 is the particular flow sheet of multiple view video coding mode selecting method according to an embodiment of the invention.As depicted in figs. 1 and 2, the multiple view video coding mode selecting method according to the embodiment of the invention may further comprise the steps:

S101, inputted video image and corresponding depth map, and each frame of video image is divided into a plurality of basic coding units, i.e. macro block.

S102 judges the position of the macro block of present encoding, the rate distortion threshold value of prediction SKIP pattern is set and calculates current macro and selects rate distortion costs under the SKIP pattern.

If current macro place frame is the first row or first row that I frame or current macro are in the place frame, then the utilization rate distortion optimizing method checks inter-frame mode in all frames, and the pattern of selection rate distortion cost minimum is the forced coding pattern.

If current macro place viewpoint is the absolute coding viewpoint, the place frame is not the first row or the first row that I frame and current macro are not in the place frame simultaneously, then calculates the first rate distortion costs RD that current macro is selected the SKIP pattern _Skip1And setting is used for predicting the first rate distortion threshold value of SKIP pattern

Calculate by following formula:

$T_{{\mathrm{skip}}_1}=\frac{\underset{i=1}{\overset{4}{\mathrm{\Σ}}}{W}_i\·I_i\·\mathrm{RD}\cos t{\left(\mathrm{SKIP}\right)}_i}{\underset{i=1}{\overset{4}{\mathrm{\Σ}}}{W}_i\·I_i},$

Wherein, RDcost (SKIP) _iBe respectively the left macro block (macro block 1 among Fig. 3), the top macro block (macro block 2 among Fig. 3) that calculate current macro (macro block 0 among Fig. 3), macro block (in Fig. 3 macro block 4) the rate distortion costs under SKIP pattern identical with the current macro position in upper right side macro block (macro block 3 among Fig. 3) and the former frame.W _iBe the respective weights coefficient, drawn by training under the line.

I _iShown in being defined as follows:

If current macro place viewpoint is the associated prediction coded views, the place frame is not the first row or the first row that I frame and current macro are not in the place frame simultaneously, and the second rate distortion costs RD that is used for predicting the SKIP pattern then is set _Skip2And the calculating current macro is selected the second rate distortion threshold value T of SKIP pattern _Skip2

$T_{{\mathrm{skip}}_2}=\frac{\underset{i=1}{\overset{13}{\mathrm{\Σ}}}{W}_i\·I_i\·\mathrm{RD}\cos t{\left(\mathrm{SKIP}\right)}_i}{\underset{i=1}{\overset{13}{\mathrm{\Σ}}}{W}_i\·I_i},$

Wherein, RDcost (SKIP) _iBe respectively among Fig. 4 macro block 1 to macro block 13 rate distortion costs under the SKIP pattern respectively.These macro blocks are respectively left macro block (macro block 1), top macro block (macro block 2), the upper right side macro block (macro block 3) of current macro 0 in the current view point video, and the macro block (macro block 4) identical with the current macro position and adjacent coded views are by macro block (macro block 5) and adjacent 8 macro blocks (macro block 6 ~ macro block 13) thereof of parallactic displacement in the former frame.The parallax that macro block 0 and macro block are 5 is derived by the degree of depth of current view point current macro and is gone out.

W _iBe the respective weights coefficient, drawn by training under the line.I _iShown in being defined as follows:

Step S103 is according to the motion amplitude of rate distortion threshold value and rate distortion costs calculating current macro.

Compare the first rate distortion costs RD _Skip1With the first rate distortion threshold value

As the first rate distortion costs RD _Skip1Less than the first rate distortion threshold value

The time, be the forced coding pattern of current macro with the SKIP pattern.

As the first rate distortion costs RD _Skip1Greater than the first rate distortion threshold value

The time, the first motion amplitude A of calculating current macro _MV1, A _MV1Computing formula as follows:

At first get among Fig. 3 macro block 1 to the motion vector MV of macro block 4 _i, (i=1,2,3,4), and note MV _i=(x _i, y _i).Then, $A_{\mathrm{MV}1}=\left|\frac{\underset{i=1}{\overset{4}{\mathrm{\Σ}}}{W}_i\·K_i\·x_i}{\underset{i=1}{\overset{4}{\mathrm{\Σ}}}{W}_i}\right|+\left|\frac{\underset{i=1}{\overset{4}{\mathrm{\Σ}}}{W}_i\·K_i\·y_i}{\underset{i=1}{\overset{4}{\mathrm{\Σ}}}{W}_i}\right|,$ Wherein, W _iBe the respective weights coefficient, draw K by training under the line _iShown in being defined as follows:

Compare the second rate distortion costs RD _Skip2With the second rate distortion threshold value T _Skip2

As the second rate distortion costs RD _Skip2Less than the second rate distortion threshold value T _Skip2The time, be the forced coding pattern of current macro with the SKIP pattern.

When the second rate distortion costs during greater than the second rate distortion threshold value, calculate the second motion amplitude A of current macro _MV2, A _MV2Computing formula as follows:

At first get among Fig. 4 macro block 1 to the motion vector MV of macro block 13 _i, (i=1,2 ..., 13), and note MV _i=(x _i, y _i).Then, $A_{\mathrm{<figure-callout\; id="2"\; label="MV"\; filenames="HDA00002693098900021.png"\; state="\{\{state\}\}">MV</figure-callout>}2}=\left|\frac{\underset{i=1}{\overset{13}{\mathrm{\&Sigma;}}}{W}_i\&CenterDot;K_i\&CenterDot;x_i}{\underset{i=1}{\overset{13}{\mathrm{\&Sigma;}}}{W}_i}\right|+\left|\frac{\underset{i=1}{\overset{13}{\mathrm{\&Sigma;}}}{W}_i\&CenterDot;K_i\&CenterDot;y_i}{\underset{i=1}{\overset{13}{\mathrm{\&Sigma;}}}{W}_i}\right|,$ Wherein, W _iBe the respective weights coefficient, draw K by training under the line _iShown in being defined as follows:

S104 chooses corresponding candidate pattern according to the motion amplitude size of current macro.

Motion amplitude A according to current macro _MV1Or A _MV2Size choose corresponding candidate pattern.

(1) works as A _MV1≤ 1 o'clock is the forced coding pattern of current macro with SKIP and ME16 * 16 as candidate pattern and the pattern of choosing the rate distortion minimum.

(2) for 1＜A _MV1≤ 2 o'clock, with ME16 * 8, the pattern that rate-distortion optimization calculating and selection rate distortion cost minimum are carried out as candidate pattern in ME8 * 16 was the forced coding pattern.

(3) for 2＜A _MV1≤ 5 o'clock, selecting MV_P8 * 8 patterns was the situation of cutting apart that current macro coding and utilization rate distortion optimizing method are chosen 8 * 8 inferior macro blocks of rate distortion costs minimum.

(4) for A _MV15 o'clock, select DE (16 * 16,16 * 8,8 * 16,8 * 8), intra_4 * 4, intra_8 * 8, intra_16 * 16 are the forced coding pattern of current macro for candidate pattern and the pattern of choosing the rate distortion minimum.

(5) work as A _MV2≤ 1 o'clock is the forced coding pattern of current macro with SKIP and ME16 * 16 as candidate pattern and the pattern of choosing the rate distortion minimum.

(6) for 1＜A _MV2≤ 2 o'clock, with ME16 * 8, the pattern that rate-distortion optimization calculating and selection rate distortion cost minimum are carried out as candidate pattern in ME8 * 16 was the forced coding pattern.

(7) for 2＜A _MV2≤ 5 o'clock, selecting MV_P8 * 8 patterns was the situation of cutting apart that current macro coding and utilization rate distortion optimizing method are chosen 8 * 8 inferior macro blocks of rate distortion costs minimum.

(8) for A _MV25 o'clock, select DE (16 * 16,16 * 8,8 * 16,8 * 8), intra_4 * 4, intra_8 * 8, intra_16 * 16 are the forced coding pattern of current macro for candidate pattern and the pattern of choosing the rate distortion minimum.

Finish after the coding of current macro, continue next macro block of coding until finish a plurality of macro blocks of each frame of video image.

Method according to the embodiment of the invention, predict the SKIP pattern by block-based adaptive threshold, a large amount of unnecessary calculating have been avoided, and predict the motion amplitude of current macro according to the motion vector of adjacent macroblocks and the corresponding macro block of adjacent viewpoint, the candidate pattern that meets according to the large minor inspection of motion amplitude again can reduce thus the hunting zone, reduce operation time.

Fig. 5 is the frame diagram of multiple view video coding mode selector according to an embodiment of the invention.As shown in Figure 5, according to the multiple view video coding mode selector of the embodiment of the invention, comprise that division module 100, judgement arrange module 200, computing module 300 and choose module 400.

Divide module 100 and be used for inputted video image and corresponding depth map, and each frame of video image is divided into a plurality of basic coding units, i.e. macro block.

Judge the position that module 200 is used for judging the macro block of present encoding is set, the rate distortion threshold value of prediction SKIP pattern is set and calculates current macro and select rate distortion costs under the SKIP pattern.

In one embodiment of the invention, judge that module 200 is set comprises selected cell 210, the first computing unit 220 and the second computing unit 230.

It is the first row or the first row that I frame or current macro are in the place frame that selected cell 210 is used for working as current macro place frame, then the utilization rate distortion optimizing method checks inter-frame mode in all frames, and the pattern of selection rate distortion cost minimum is the forced coding pattern.

It is the absolute coding viewpoint that the first computing unit 220 is used for working as current macro place viewpoint, the place frame is not the first row or the first row that I frame and current macro are not in the place frame simultaneously, then calculates current macro and selects the first rate distortion costs of SKIP pattern and the first rate distortion threshold value that is used for predicting the SKIP pattern is set.

The first rate distortion threshold value Calculate by following formula: $T_{{\mathrm{skip}}_1}=\frac{\underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}{W}_i\&CenterDot;I_i\&CenterDot;\mathrm{RD}\cos t{\left(\mathrm{SKIP}\right)}_i}{\underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}{W}_i\&CenterDot;I_i}.$

Wherein, RDcost (SKIP) _iBe respectively the left macro block (macro block 1 among Fig. 3), the top macro block (macro block 2 among Fig. 3) that calculate current macro (macro block 0 among Fig. 3), macro block (in Fig. 3 macro block 4) the rate distortion costs under SKIP pattern identical with the current macro position in upper right side macro block (macro block 3 among Fig. 3) and the former frame.W _iBe the respective weights coefficient, drawn by training under the line.

I _iShown in being defined as follows:

It is the associated prediction coded views that the second computing unit 230 is used for working as current macro place viewpoint, the place frame is not the first row or the first row that I frame and current macro are not in the place frame simultaneously, then arranges to be used for predicting the second rate distortion threshold value of SKIP pattern and to calculate the second rate distortion costs that current macro is selected the SKIP pattern.

The second rate distortion threshold value T _Skip2, calculate by following formula: $T_{{\mathrm{skip}}_2}=\frac{\underset{i=1}{\overset{13}{\mathrm{\&Sigma;}}}{W}_i\&CenterDot;I_i\&CenterDot;\mathrm{RD}\cos t{\left(\mathrm{SKIP}\right)}_i}{\underset{i=1}{\overset{13}{\mathrm{\&Sigma;}}}{W}_i\&CenterDot;I_i},$

Wherein, RDcost (SKIP) _iBe respectively among Fig. 4 macro block 1 to macro block 13 rate distortion costs under the SKIP pattern respectively.These macro blocks are respectively left macro block (macro block 1), top macro block (macro block 2), the upper right side macro block (macro block 3) of current macro 0 in the current view point video, and the macro block (macro block 4) identical with the current macro position and adjacent coded views are by macro block (macro block 5) and adjacent 8 macro blocks (macro block 6 ~ macro block 13) thereof of parallactic displacement in the former frame.The parallax that macro block 0 and macro block are 5 is derived by the degree of depth of current view point current macro and is gone out.

W _iBe the respective weights coefficient, drawn by training under the line.I _iShown in being defined as follows:

Computing module 300 is used for the motion amplitude according to rate distortion threshold value and rate distortion costs calculating current macro.

In one embodiment of the invention, computing module 300 comprises that the first comparing unit 310, the first processing unit 320, the second processing unit 330, the second comparing unit 340, the 3rd processing unit 350 and manage unit 360 everywhere.

The first comparing unit 310 is used for relatively the first rate distortion costs and the first rate distortion threshold value.

The first processing unit 320 is used for when the first rate distortion costs during less than the first rate distortion threshold value, is the forced coding pattern of current macro with the SKIP pattern.

The second processing unit 330 is used for when the first rate distortion costs during greater than the first rate distortion threshold value, the first motion amplitude of calculating current macro.

The second comparing unit 340 is used for relatively the second rate distortion costs and the second rate distortion threshold value.

The 3rd processing unit 350 is used for when the second rate distortion costs during less than the second rate distortion threshold value, is the forced coding pattern of current macro with the SKIP pattern.

The manages unit 360 everywhere is used for when the second rate distortion costs during greater than the second rate distortion threshold value the first motion amplitude of calculating current macro.

Choosing module 400 is used for choosing corresponding candidate pattern according to the motion amplitude size of current macro.

Motion amplitude A according to current macro _MV1Or A _MV2Size choose corresponding candidate pattern.

(1) works as A _MV1≤ 1 o'clock is the forced coding pattern of current macro with SKIP and ME16 * 16 as candidate pattern and the pattern of choosing the rate distortion minimum.

(2) for 1＜A _MV1≤ 2 o'clock, with ME16 * 8, the pattern that rate-distortion optimization calculating and selection rate distortion cost minimum are carried out as candidate pattern in ME8 * 16 was the forced coding pattern.

(3) for 2＜A _MV1≤ 5 o'clock, selecting MV_P8 * 8 patterns was the situation of cutting apart that current macro coding and utilization rate distortion optimizing method are chosen 8 * 8 inferior macro blocks of rate distortion costs minimum.

(4) for A _MV15 o'clock, select DE (16 * 16,16 * 8,8 * 16,8 * 8), intra_4 * 4, intra_8 * 8, intra_16 * 16 are the forced coding pattern of current macro for candidate pattern and the pattern of choosing the rate distortion minimum.

(5) work as A _MV2≤ 1 o'clock is the forced coding pattern of current macro with SKIP and ME16 * 16 as candidate pattern and the pattern of choosing the rate distortion minimum.

(6) for 1＜A _MV2≤ 2 o'clock, with ME16 * 8, the pattern that rate-distortion optimization calculating and selection rate distortion cost minimum are carried out as candidate pattern in ME8 * 16 was the forced coding pattern.

(7) for 2＜A _MV2≤ 5 o'clock, selecting MV_P8 * 8 patterns was the situation of cutting apart that current macro coding and utilization rate distortion optimizing method are chosen 8 * 8 inferior macro blocks of rate distortion costs minimum.

(8) for A _MV25 o'clock, select DE (16 * 16,16 * 8,8 * 16,8 * 8), intra_4 * 4, intra_8 * 8, intra_16 * 16 are the forced coding pattern of current macro for candidate pattern and the pattern of choosing the rate distortion minimum.

Finish after the coding of current macro, continue next macro block of coding until finish a plurality of macro blocks of each frame of video image.

Device according to the embodiment of the invention, predict the SKIP pattern by block-based adaptive threshold, a large amount of unnecessary calculating have been avoided, and predict the motion amplitude of current macro according to the motion vector of adjacent macroblocks and the corresponding macro block of adjacent viewpoint, the candidate pattern that meets according to the large minor inspection of motion amplitude again can reduce thus the hunting zone, reduce operation time.

The specific operation process that should be appreciated that modules among the device embodiment of the present invention and unit can be identical with the description in the embodiment of the method, is not described in detail herein.

Although the above has illustrated and has described embodiments of the invention, be understandable that, above-described embodiment is exemplary, can not be interpreted as limitation of the present invention, those of ordinary skill in the art can change above-described embodiment in the situation that does not break away from principle of the present invention and aim within the scope of the invention, modification, replacement and modification.

Claims (8)

1. a multiple view video coding mode selecting method is characterized in that, comprises step:

S1: inputted video image and corresponding depth map, and each frame of described video image is divided into a plurality of basic coding units, i.e. macro block;

S2: judge the position of the described macro block of present encoding, the rate distortion threshold value of prediction SKIP pattern is set and calculates current macro and select rate distortion costs under the SKIP pattern;

S3: according to the motion amplitude of described rate distortion threshold value and described rate distortion costs calculating current macro; And

S4: the described motion amplitude size according to current macro is chosen corresponding candidate pattern.

2. multiple view video coding mode selecting method as claimed in claim 1 is characterized in that, described step S2 specifically comprises:

S21: if current macro place frame is the first row or first row that I frame or current macro are in the place frame, then the utilization rate distortion optimizing method checks inter-frame mode in all frames, and the pattern of selection rate distortion cost minimum is the forced coding pattern;

S22: if current macro place viewpoint is the absolute coding viewpoint, the place frame is not the first row or the first row that I frame and current macro are not in the place frame simultaneously, then calculates current macro and selects the first rate distortion costs of SKIP pattern and the first rate distortion threshold value that is used for predicting the SKIP pattern is set; And

S23: if current macro place viewpoint is the associated prediction coded views, the place frame is not the first row or the first row that I frame and current macro are not in the place frame simultaneously, then arranges to be used for predicting the second rate distortion threshold value of SKIP pattern and to calculate the second rate distortion costs that current macro is selected the SKIP pattern.

3. multiple view video coding mode selecting method as claimed in claim 1 is characterized in that, described step S3 specifically comprises:

More described the first rate distortion costs and described the first rate distortion threshold value;

When described the first rate distortion costs during less than described the first rate distortion threshold value, be the forced coding pattern of current macro with the SKIP pattern; And

When described the first rate distortion costs during greater than described the first rate distortion threshold value, calculate the first motion amplitude of current macro.

4. multiple view video coding mode selecting method as claimed in claim 1 is characterized in that, described step S3 also comprises:

More described the second rate distortion costs and described the second rate distortion threshold value;

When described the second rate distortion costs during less than described the second rate distortion threshold value, be the forced coding pattern of current macro with the SKIP pattern; And

When described the second rate distortion costs during greater than described the second rate distortion threshold value, calculate the first motion amplitude of current macro.

5. multiple view video coding mode selector comprises:

Divide module, be used for inputted video image and corresponding depth map, and each frame of described video image is divided into a plurality of basic coding units, i.e. macro block;

Judge module is set, be used for judging the position of the described macro block of present encoding, the rate distortion threshold value of prediction SKIP pattern is set and calculates current macro and select rate distortion costs under the SKIP pattern;

Computing module is used for the motion amplitude according to described rate distortion threshold value and described rate distortion costs calculating current macro; And

Choose module, be used for choosing corresponding candidate pattern according to the described motion amplitude size of current macro.

6. multiple view video coding mode selector as claimed in claim 5 is characterized in that, described judgement arranges module and specifically comprises:

Selected cell, being used for working as current macro place frame is the first row or the first row that I frame or current macro are in the place frame, and then the utilization rate distortion optimizing method checks inter-frame mode in all frames, and the pattern of selection rate distortion cost minimum is the forced coding pattern;

The first computing unit, being used for working as current macro place viewpoint is the absolute coding viewpoint, the place frame is not the first row or the first row that I frame and current macro are not in the place frame simultaneously, then calculates current macro and selects the first rate distortion costs of SKIP pattern and the first rate distortion threshold value that is used for predicting the SKIP pattern is set; And

The second computing unit, being used for working as current macro place viewpoint is the associated prediction coded views, the place frame is not the first row or the first row that I frame and current macro are not in the place frame simultaneously, then arranges to be used for predicting the second rate distortion threshold value of SKIP pattern and to calculate the second rate distortion costs that current macro is selected the SKIP pattern.

7. multiple view video coding mode selector as claimed in claim 5 is characterized in that, described computing module specifically comprises:

The first comparing unit is used for more described the first rate distortion costs and described the first rate distortion threshold value;

The first processing unit is used for when described the first rate distortion costs during less than described the first rate distortion threshold value, is the forced coding pattern of current macro with the SKIP pattern; And

The second processing unit is used for when described the first rate distortion costs during greater than described the first rate distortion threshold value the first motion amplitude of calculating current macro.

8. multiple view video coding mode selector as claimed in claim 5 is characterized in that, described computing module specifically also comprises:

The second comparing unit is used for more described the second rate distortion costs and described the second rate distortion threshold value;

The 3rd processing unit is used for when described the second rate distortion costs during less than described the second rate distortion threshold value, is the forced coding pattern of current macro with the SKIP pattern; And

The manages the unit everywhere, is used for when described the second rate distortion costs during greater than described the second rate distortion threshold value the first motion amplitude of calculating current macro.

Images