CN1067204C - Global decision method for video frequency coding - Google Patents

Abstract

The present invention belongs to the field of the coding technique of a moving image, which comprises motion estimation, frame prediction, the generation of prediction error, conversion, quantization, image reconstruction and entropy coding. The present invention is characterized in that a global judging device for judging the result of each stage in the process of the block matching of a graded search is arranged in the motion estimation, and whether the image reconstruction error in a current frame surpasses the image reconstruction error in a previous frame is judged; the global judging device adopts the criterion of the judgment to the reconstruction error. With the adoption of the present invention, coding speed can be greatly improved.

Description

The global decision method of video coding

The invention belongs to the motion image coding technical field.

In the current information age, the storage and the transmission of image become more and more important.Because visual primary data information (pdi) amount is very huge, store with the information channel of limited capacity at the storage medium of limited capacity and transmit image, the compression of pictorial data, just motion video being encoded is absolutely necessary.Motion image coding be by the comprehensive utilization picture intelligence in the redundancy aspect time, space, the statistics three, and the knowledge of scene, human vision property realize.Comparatively ripe at present coding method is the hybrid coding method that combines predictive coding, transition coding, several coding methods of entropy coding and motion compensation technique, and block diagram as shown in Figure 1.Its coding is realized the implementation method following steps: (1) makes estimation ME to the current image and a last image reconstruction of input, and estimation obtains motion vector MV; (2) according to motion vector a last reconstruction frames is made prediction P, obtain predictive image current image; (3) predictive image that current image is deducted current image obtains predicated error PE; (4) predicated error is made discrete cosine transform DCT and quantized Q; (5) result of (4) is made variable-length encoding VLC and obtain the present encoding image; (6) result to (4) makes the predicated error that re-quantization IQ, inverse discrete cosine transform IDCT obtain rebuilding, and obtains current image reconstruction with current predictive image addition again, becomes a last image reconstruction through frame memory FM.

The function of each step is as follows among Fig. 1: the calculating of estimation, prediction and predicated error (subtracter) constitutes predictive coding, is in order to eliminate picture intelligence correlation in time.Certain part that in time correlation of image shows as the present frame image is obtained through motion by certain part of previous frame image.And motion is described by motion vector, and estimation will be tried to achieve motion vector exactly, and prediction then is the variation that compensates, offsets the signal that produces because of motion between present frame image and the previous frame image area according to motion vector.Discrete cosine transform DCT constitutes transition coding, and purpose is in order to eliminate picture intelligence correlation spatially.Quantizing Q is the needs of back entropy coding, is again to have utilized human vision property to improve encoding quality.Variable-length encoding VLC constitutes entropy coding, has further eliminated the correlation of the statistics of picture intelligence.Re-quantization IQ, inverse discrete cosine transform IDCT, adder realize image reconstruction, for prediction provides object of reference.

The encoder of the above-mentioned hybrid coding method of existing realization as shown in Figure 2.Piece DCT, piece Q, piece IQ, piece IDCT, piece VLC are meant that one (is 8 * 8 pixels by international standard) advanced discrete cosine transform DCT, quantized Q, re-quantization IQ, inverse discrete cosine transform IDCT, variable-length encoding VLC for unit in a frame (width of cloth) image among the figure.Macro block P is meant that with a macro block (is 6 pieces by international standard) in the frame image be that unit predicts P.Macro block MV refers to the motion vector of macro block.0 is meant that 8 * 8 elements all are 0 in the piece.

The course of work that this encoder is realized is: at first, a frame (width of cloth) image is that unit encodes with one (8 * 8 pixel) or a macro block (6 pieces).Secondly, whole cataloged procedure is divided into two parts: estimation and coding core are shown in the empty frame of Fig. 2.The present frame macro block at first enters estimation and partly makes estimation, and estimation obtains motion vector MV, is input to the coding core present frame macro block is encoded, and specifically is divided into following steps:

At first, a present frame macro block and a last reconstruction frames macro block are made estimation.Estimation is divided into whole pel search and two steps of half pel search.Whole pel search adopts the graded movement searching method, is about to the region of search classification: rest point, little region of search, large search territory.This is typical 3 grades, and specific implementation can change, and as 2 grades: rest point, little region of search are about to little region of search and large search territory and merge; Or 4 grades: rest point, little region of search, large search territory, large search territory more are about to the large search territory and segment two-stage again.After finishing, the search of preceding two-stage region of search to judge that judge whether to satisfy criterion and can stop the coding core that whole pel search enters into half pel search and back, concrete decision rule may be diversified.

Secondly, enter after the coding core, the motion vector that obtains macro block according to estimation is predicted a last reconstruction frames macro block and is obtained current predictive frame macro block, with subtracter the present frame macro block is deducted current predictive frame macro block then and calculate predicated error, then do DCT, Q, entropy coding (VLC) and image reconstruction.

The deficiency of the implementation method of this encoder is that coding rate is not high enough.To QCIF form simple motion image sequence (as the Claire sequence), utilize Pentium-133PC still can't accomplish software real-time coding (25 frame/second).

The objective of the invention is to overcome the weak point of prior art, on the basis of original coding method, set up to image in the judgement of a large amount of unnecessary computings, improve encoder speed.

The present invention proposes a kind of global decision method of video coding, may further comprise the steps:

(1) at first, a present frame macro block and a last reconstruction frames macro block are made estimation, estimation comprises whole pel search and half pel search;

(2) after the search of each grade region of search finishes, all enter global decision device G, whether judgement current input macro block under current this one-level search satisfies the reconstruction error criterion, if satisfy the then not only search procedure of current macro end, and the whole cataloged procedure of this macro block also finishes, and coding forwards next macro block to; If do not satisfy the search that then continues next stage, half pel search obtains motion vector after finishing, and enters the coding core;

(3) enter after the coding core, the motion vector that obtains macro block according to estimation is predicted a last reconstruction frames macro block and is obtained current predictive frame macro block, deduct current predictive frame macro block with the present frame macro block then and calculate predicated error, then do DCT, Q, entropy coding (VLC) and image reconstruction.

Said whole pel search can adopt the graded movement searching method, with the region of search classification: rest point, little region of search, large search territory.

The said global decision device of the present invention adopts reconstruction error decision rule.

The said reconstruction error decision rule of the present invention is: to present frame (k frame) macro block 1, if satisfy in estimation

SAD ₁ ^(k)(i, j)≤E _Reconl ^(k), (then current estimation can finish for i, j) ∈ S, and (i j) is the motion vector of the macro block 1 that will look for, and other coding computing of macro block 1, and predicated error calculating and DCT, Q, IQ, IDCT can save; If inequality above not satisfying is then proceeded estimation.

The principle of the reconstruction error decision rule that the present invention adopts is summarized as follows: in estimation, adopts BMA usually, SAD (the Sum of Absolute Difference) value of computing macro block, ${\mathrm{SAD}}_1^{\left(k\right)}(i,j)=\underset{m=1}{\overset{N}{\mathrm{\Σ}}}\underset{n=1}{\overset{N}{\mathrm{\Σ}}}\left|I_o^{\left(k\right)}\right(m,n)-I_o^{(k-1)}(m+i,n+j\left)\right|,(i,j)\∈S---\left(1\right)$ SAD wherein ₁ ^(k)Be the sad value of k frame macro block 1, (i j) is candidate's motion vector, and S is a region of search, I ₀ ^(k)Be k primitive frame gray value, I ₀ ^(k-1)Be k-1 primitive frame gray value, the size of macro block is N * N.The reconstruction error E of former frame (k-1 frame) macro block 1 _Reconl ^(k-1)Can calculate by following formula, $E_{\mathrm{recon}1}^{(k-1)}=\underset{m=1}{\overset{N}{\mathrm{\Σ}}}\underset{n=1}{\overset{N}{\mathrm{\Σ}}}\left|I_0^{(k-1)}\right(m,n)-I_0^{(k-1)}(m,n\left)\right|-----\left(2\right)$

I wherein _r ^(k-1)It is k-1 frame reconstruction frames.Then reconstruction error decision rule is: to present frame (k frame) macro block 1, if satisfy in estimation

SAD ₁ ^(k)(i, j)≤E _Reconl ^(k), ((3) then current estimation can finish for i, j) ∈ S, and (i j) is the motion vector of the macro block 1 that will look for, and other coding computing of macro block 1, and predicated error calculating and DCT, Q, IQ, IDCT can save; If do not satisfy inequality (3), then proceed estimation.

Novel Video Coder Implementation block diagram of the present invention as shown in Figure 3, its course of work is: at first, a present frame macro block and a last reconstruction frames macro block are made estimation.Estimation comprises whole pel search and half pel search.Whole pel search adopts the graded movement searching method, with the region of search classification: rest point, little region of search, large search territory.(this is typical 3 grades, and specific implementation can change, and is the same).After the search of each grade region of search finishes, all enter global decision device G, whether judgement current input macro block under current this one-level search satisfies the reconstruction error criterion, if satisfy the then not only search procedure of current macro end, and the whole cataloged procedure of this macro block also finishes, and coding forwards next macro block to; If do not satisfy the search that then continues next stage.Half pel search obtains motion vector after finishing, and enters the coding core.

Secondly, enter after the coding core, the motion vector that obtains macro block according to estimation is predicted a last reconstruction frames macro block and is obtained current predictive frame macro block, deduct current predictive frame macro block with the present frame macro block then and calculate predicated error, then do DCT, Q, entropy coding (VLC) and image reconstruction.

In conjunction with Fig. 3 and Fig. 2, the present invention compared with prior art has following characteristics as can be seen: first, whole pixel parts in estimation of the present invention, after each grade search of whole pel search finishes, global decision device G is set, in case criterion is to satisfy among the decision device G, then the whole cataloged procedure to this macro block just is through with, and just not only finishes whole pel search process, and half pel search, whole coding core all needn't have been done, thereby improve speed of coding greatly.The second, to decision device G, we have adopted reconstruction error criterion efficiently.Whether this criterion not only can be controlled search and stop, and carries out conversion and processing but also can the control frame difference signal whether send into DCT.

Brief description of drawings

Fig. 1 is existing hybrid coding method block diagram

Fig. 2 is existing video encoder structured flowchart

Fig. 3 is a video encoder structured flowchart of the present invention

Adopt a kind of example structure that coding method of the present invention realizes video encoder as shown in Figure 3.Its course of work may further comprise the steps:

1. as estimation the search of motion vector to a present frame macro block and a last reconstruction frames macro block.The search of motion vector comprises whole pel search and half pel search.At first put in order pel search.Whole pixel adopts the graded movement searching method, and step is as follows:

A. rest point search judges whether to satisfy the reconstruction error criterion of global decision device definition, if satisfy then carry out 2, otherwise continues;

The search of b. little region of search judges whether to satisfy the reconstruction error criterion of global decision device definition, if satisfy then carry out 2, otherwise continues;

C. large search domain search judges whether to satisfy the reconstruction error criterion of global decision device definition, if satisfy then carry out 2, otherwise carries out 3.

2, by current search to a motion vector and a last reconstruction frames obtain current predicted macroblock, current predicted macroblock is current reconstruction macro block, carries out 4;

3, carry out half pel search, obtain motion vector, obtain current predicted macroblock by a motion vector and a last reconstruction frames, deduct current predicted macroblock with the present frame macro block then and calculate predicated error, then make dct transform, quantize Q, inverse quantization IQ, anti-dct transform, result who obtains and current predicted macroblock addition obtain current reconstruction macro block;

4, carry out entropy coding, the current macro end-of-encode enters next macroblock encoding.

Present embodiment is given on the Pentium-200 PC the coding parameter and the coding rate of the simple sequence Claire sequence of QCIF form and complex sequence Foreman sequence as follows: the Claire sequence: I frame quantization parameter is 10, and P frame quantization parameter is 10.Estimation

In whole pel search adopt the two-stage search: rest point, little region of search,

Zone length is 5, and each grade searching algorithm is the sampling method three-step approach, piece

The sub-sampling method is adopted in matching operation.

The result: coding frame per second (speed) reaches average 70 frame/seconds.The Foreman sequence: I frame quantization parameter is 20, and P frame quantization parameter is 20.Motion is estimated

Whole pel search in the meter adopts three grades of search: rest point, little fortune

Moving zone (zone length is 4), large search territory (zone length is 10)

Each grade searching algorithm is a three-step approach, and the piece matching operation is adopted inferior

The method of sampling.

The result: coding frame per second (speed) reaches average 30 frame/seconds.

Claims (3)

1, a kind of global decision method of video coding may further comprise the steps:

(1) at first, a present frame macro block and a last reconstruction frames macro block are made estimation, estimation comprises whole pel search and half pel search;

(2) after the search of each grade region of search finishes, all enter global decision device G, whether judgement current input macro block under current this one-level search satisfies the reconstruction error criterion, if satisfy the then not only search procedure of current macro end, and the whole cataloged procedure of this macro block also finishes, and coding forwards next macro block to; If do not satisfy the search that then continues next stage, half pel search obtains motion vector after finishing, and enters the coding core;

(3) enter after the coding core, the motion vector that obtains macro block according to estimation is predicted a last reconstruction frames macro block and is obtained current predictive frame macro block, deduct current predictive frame macro block with the present frame macro block then and calculate predicated error, then do DCT, Q, entropy coding (VLC) and image reconstruction.

2, the global decision method of video coding as claimed in claim 1 is characterized in that, said whole pel search adopts the graded movement searching method, with the region of search classification: rest point, little region of search, large search territory.

3, the global decision method of video coding as claimed in claim 1 is characterized in that, said reconstruction error decision rule is: to present frame (k frame) macro block 1, if satisfy SAD in estimation ₁ ^(k)(i, j)≤E _Reconl ^(k), (then current estimation can finish for i, j) ∈ S, and (i j) is the motion vector of the macro block 1 that will look for, and other coding computing of macro block 1, and predicated error calculating and DCT, Q, IQ, IDCT can save.

Images