CN1665301A

CN1665301A - A self-adaptive scanning method in video coding/decoding

Info

Publication number: CN1665301A
Application number: CN 200510051311
Authority: CN
Inventors: 何芸; 陈衢清
Original assignee: Tsinghua University
Current assignee: Tsinghua University
Priority date: 2005-03-07
Filing date: 2005-03-07
Publication date: 2005-09-07

Abstract

The invention relates to a self-adaptive scanning method in the video coding/decoding course, belonging to the video coding/decoding field during the signal processing. The invention sets many scanning lists in the video coding/decoding standard according to various picture forecasting methods, where each coding method of a block corresponds to a scanning list; as coding a block, according to the coding method of the present block, selecting a corresponding scanning list for scanning; as decoding a block, according to the decoded coding method of the block, selecting a corresponding scanning list for scanning. The characters of the invention: according to the coding method of the present block, selecting a scanning mode. The invention can further raise the efficiency of compression coding of the video signal.

Description

Adaptive scan method in a kind of coding and decoding video

Technical field

The invention belongs to the coding and decoding video field in the signal processing, particularly in the video signal coding to the scan method of the coefficient behind the change quantization.

Background technology is in existing video coding technique, one images of video sequence generally is divided into several pieces and encodes, said be in the image as the pixel collection of coding base unit, be the pixel number more than or equal to the pixel collection on 1 square, rectangle or arbitrary shape border.Comprise prediction, conversion, quantification, scanning and entropy coding again as the step 1 of coding.The scan mode of a superior performance can make on statistics that the coefficient that need encode is more concentrated, and arrange from big to small according to absolute value, can more help entropy coding like this,, thereby make the compressed encoding performance further improve especially based on contextual entropy coding.

In existing international video encoding and decoding standard, the scan mode of employing has two kinds of frame pattern and field modes, for example in MPEG2, scan table is divided into two kinds, as depicted in figs. 1 and 2, Arabic numerals are represented the scanning sequency of pixel in the piece among the figure, are respectively applied for frame pattern and field mode.Use two kinds of scan tables to be like this because video sequence is carried out encoding and being furnished with certain difference by the encode statistical of the coefficient that produced of field mode by frame pattern, adopt different scan modes more to help the raising of coding efficiency.But, existing international video encoding and decoding standard, no matter be classical MPEG2 or at present H.264/AVC up-to-date, scanning in them is except frame pattern and field mode self adaptation, do not utilize other characteristics to come adaptive selection scan mode again, therefore limited the further raising of compression performance.

Summary of the invention

The objective of the invention is in order to overcome the weak point of prior art, adaptive scan method in a kind of coding and decoding video has been proposed, encoder is according to the characteristics of the coding method of present encoding piece, and its corresponding scan table of selection carries out the scanning in the cataloged procedure in a plurality of scan tables; Decoder selects corresponding scan table to scan from the characteristics of the coding method of this piece of having decoded, to improve the compression performance of existing video coding.

Adaptive scan method in a kind of coding and decoding video that the present invention proposes is characterized in that, specifically comprises:

In video encoding and decoding standard, set up a plurality of scan tables, the corresponding corresponding scan table of every kind of coding method of piece according to various visual Forecasting Methodology statisticses;

Select corresponding scan table to scan according to the coding method of current block when encoding a piece;

The coding method that has decoded according to this piece when decoding a piece selects corresponding scan table to scan.

Characteristics of the present invention and good result:

Still may there be certain difference in the optimum scanning that the present invention is based under different coding methods, and has the characteristics of certain correlation on statistics with the coding method of present encoding piece.Utilize more meticulous characteristic to realize more flexible accurate adaptive scanning, more flexible more accurate than the scanning technique in the existing international video encoding and decoding standard, and improved the performance of video compression.

Description of drawings

Fig. 1 is the scan table schematic diagram of frame pattern among the existing MPEG2.

Fig. 2 is the scan table schematic diagram of field mode among the existing MPEG2.

Scan table 1 schematic diagram of Fig. 3 for setting up in the embodiment of the invention.

Scan table 2 schematic diagrames of Fig. 4 for setting up in the embodiment of the invention.

Scan table 3 schematic diagrames of Fig. 5 for setting up in the embodiment of the invention.

Scan table 4 schematic diagrames of Fig. 6 for setting up in the embodiment of the invention.

Scan table 5 schematic diagrames of Fig. 7 for setting up in the embodiment of the invention.

Scan table 6 schematic diagrames of Fig. 8 for setting up in the embodiment of the invention.

Embodiment

Adaptive scan method embodiment in the coding and decoding video of the present invention is described in detail as follows:

In video encoding and decoding standard, be that piece is encoded with image division, piece is shaped as square in the present embodiment, size can be { 16 * 16,8 * 8,4 * 4}, also can adopt other shape and big or small piece all not to influence the enforcement of present embodiment, this piece is divided between image interior prediction piece and image predicts piece.The prediction piece utilizes the adjacent reference data of this interior piece of current image or the reference data of this piece itself to predict in the image; The prediction piece is to utilize the reference data of other reference pictures outside the current image to predict between image.The predictive mode of prediction piece is further divided into 5 kinds of patterns again in the image:

Pattern 0: vertical prediction, promptly utilize the reference data of the top of current block to predict from top to bottom;

Pattern 1: horizontal forecast, promptly utilize the reference data on the left side of current block to predict from left to right;

Pattern 2: mean prediction, promptly utilize the average of the reference data on the top of current block and the left side to come current block is predicted;

Mode 3: upper right-lower-left prediction, promptly utilize the data of top, top right-hand side and the left side, the left lower side of current block carry out along upper right-lower left to prediction;

Pattern 4: upper left-bottom right prediction, promptly utilize the data of the left side of current block and top carry out along upper left-lower right to prediction.

The concrete steps of adaptive scanning method embodiment of the present invention are as follows in this video encoding and decoding standard:

1) obtain and set up 6 scan tables according to various visual Forecasting Methodologies statistics in video encoding and decoding standard, as Fig. 3～shown in Figure 8, Arabic numerals are represented the scanning sequency of this pixel among the figure respectively.Wherein, under the coding situation on the scene, use scan table 1 during pattern 0, as shown in Figure 3, use scan table 2 during pattern 1, as shown in Figure 4, all use scan table 3 when pattern 2, mode 3, pattern 4, as shown in Figure 5; Under frame coding situation, use scan table 4 during pattern 0, as shown in Figure 6, use scan table 5 during pattern 1, as shown in Figure 7, all use scan table 6 when pattern 2, mode 3, pattern 4, as shown in Figure 8.

2) 1 bit is set as a token of in video code flow, this adaptive scanning mode is adopted in 1 expression in the present embodiment, the scan mode of acquiescence is adopted in 0 expression, if in video code flow, do not adopt adaptive scanning based on Forecasting Methodology, being the scan table 5 of Fig. 7 under Mo Ren the scan table coding situation on the scene then, is the scan table 6 of Fig. 8 under frame coding situation.

Other step in the present embodiment Code And Decode process all adopts conventional method.

Claims

1, the adaptive scan method in a kind of coding and decoding video is characterized in that, specifically comprises:

2, the adaptive scan method in the coding and decoding video according to claim 1 is characterized in that the coding method of described current block is the predictive mode in the cataloged procedure.

3, the adaptive scan method in the coding and decoding video according to claim 1, it is characterized in that, also be included in one or more bits are set in the code stream, be used to indicate the scan method of selecting to adopt acquiescence and still adopt adaptive scan method based on coding method.