CN110662029A

CN110662029A - Data compression method and device for dynamically and adaptively limiting value range

Info

Publication number: CN110662029A
Application number: CN201910582454.7A
Authority: CN
Inventors: 林涛
Original assignee: Shanghai Tianhe Electronic Information Co Ltd
Current assignee: Shanghai Tianhe Electronic Information Co Ltd
Priority date: 2018-06-30
Filing date: 2019-07-01
Publication date: 2020-01-07

Abstract

The invention provides a data compression method and a data compression device for dynamically and adaptively limiting the value range of a data set, which dynamically and adaptively limit the value range of elements according to a predetermined rule and an interdependency system among the elements of the data set, namely limit the actually available value range of the elements to a subset of the original value range, thereby reducing the actual value range and improving the coding efficiency.

Description

Data compression method and device for dynamically and adaptively limiting value range

Technical Field

The present invention relates to an encoding and decoding (also called decoding) system for compressing data, and more particularly, to a method and apparatus for encoding and decoding a data set with a limited range of values.

Background

With the introduction of human society into the era of big data, cloud computing, mobile computing, cloud-mobile computing, ultra high definition (4K) and ultra high definition (8K) video image resolution, 4G/5G communication, and virtual reality, it becomes an indispensable technology to compress various raw data, including big data, image data, video data, audio data, voice data, and neural network data, with ultra high compression ratio and extremely high quality.

The problem of compression coding a set of data, referred to as a data set for short, D, is essentially to represent all the elements of D e D with as few bits as possible, where e represents "belonging". The value of the element (i.e., data) D of the data set D is referred to as the value taken by the element D, and is referred to as the value or value of D. The range of values of the total composition D in which D is allowed to take different values is denoted as R = { R [ k ]: k = 0-K-1 }, wherein r [ i ] ≠ r [ j ] when i ≠ j. Obviously, K is the size of the range R, i.e. the number of elements of R.

The coding efficiency for D is directly related to the size K of the range R of values for D. The larger K, the larger the set of codewords for encoding and decoding the elements of R, i.e., the larger the number of codewords, resulting in the longer the length of the codeword (simply referred to as codeword length), and the more bits consumed.

In the prior art, when data compression is performed on a data set D, the value range R is fixed, so the size K of the value range is also fixed. When K is large, a large number of bits is used and consumed for each data d to be encoded or decoded, so that improvement of encoding efficiency is greatly limited.

In the first example of a data set and its value range, in the compression of image and video data, a frame of image is usually divided into coding blocks, each of which is represented by a block number, and the coding blocks are encoded (and correspondingly decoded) one by one. If an encoded block is an intra-coded block, the intra-prediction mode is used to specify which intra-prediction is to be performed for the intra-coded block. Each intra prediction mode of intra coding is represented by an intra prediction mode number. In this case, the range R of values of the data set D is the set of all intra prediction modes or their numbers. Obviously, in this example, the size K of the range R represents a total of K predetermined intra prediction modes. In the HEVC international video coding standard, K = 35. In the next generation international video coding standard currently in production, K may increase to 65.

A second example of a data set and its value range is that in the compression of image and video data, a color transform (also called color transform or color transform) is usually also required for an encoded block to eliminate the correlation between the components of a pixel. To adapt to the situation that different areas in images and videos often have unique color characteristics and unique color correlations, there are also many color transformations, which are represented by color transformation patterns, respectively. When encoding and decoding a coding block, a color transform mode is used to specify which color transform the coding block is to be subjected to. Each color conversion pattern is represented by a color conversion pattern number. In this case, the range R of values of the data set D is a set of all color transformation patterns or numbers thereof. Obviously, in this example, the size K of the range R indicates a total of K predetermined color conversion patterns. In the prior art, the types of color conversion used in practice are often more than 100, i.e. K > 100.

In a third example of a data set and its value range, in compression of image and video data, a technique called string matching or string prediction is often used to divide a coding block into a plurality of pixel strings, and matching or prediction is performed in units of strings. Each string is characterized by string encoding parameters including string length, offset (i.e., the position offset between the current string and the reference string). Within a coding block, each string has a string number. In this case, the range of values R of the data set is the set of possible values that the string coding parameter allows to take. Example 1: assuming that the number of pixels of a coding block is N (usually N is 64, 256, 1024, 4096, 16384), the string length, which is a possible value allowed by the string coding parameter, is 1, 2, … …, N-1, N, for a total of N values, so that the size K of the range R of values of D is equal to N. Example 2: assuming that the reference string is confined to a predetermined area of the picture, the possible values allowed to be taken by this string coding parameter are the values of the offsets that make the reference string fall within said predetermined area.

A fourth example of a data set and its value range is that in the compression of image and video data, a block matching or block prediction technique is often used, and for a current coding block, a reference block between frames or within frames is used to match (also called predict) the current coding block. Each current coding block represents the position of a reference block by the frame number (reference frame number for short) and the motion vector of the frame in which the reference block is located (reference frame for short). In this case, the range of values R of the data set is the set of possible values that the reference frame or its number and the motion vector allow to take. Such as: given that the reference block is confined to a number of predetermined areas of several frames, the possible values allowed to be taken by the reference frame or its number and the motion vector are the values of the reference frame or its number and the motion vector that make the reference block fall within said predetermined area.

Disclosure of Invention

In order to solve the problem that when the value range of a data set is large, the number of bits used for data encoding and decoding is large, and encoding efficiency is low, the invention provides a data compression method and a data compression device for dynamically and adaptively limiting the value range of the data set.

The invention is characterized in that the inherent natural correlation between the current coding and decoding element D and other elements of the data set D and/or the correlation formed in the coding and decoding process are fully utilized, and a dynamically and self-adaptively changed sub-range Rs is specified in the value range R according to a preset ruleAnd R is taken as the value range of d. Since Rs is smaller than R, the number of bits required to be consumed for representing d is also reduced when data compression is performed, thereby improving the encoding efficiency. Of course, for each d or group of d, it may need to consume some number of bits to represent what the Rs is. However, despite this negative effect, the present invention, as a whole, undoubtedly has the effect of improving coding efficiency, as long as Rs is properly selected and specified.

In the present invention, preferably, the interrelationship is a grouping relationship, that is, the data set D is divided into a plurality of groups, and the elements have a same-group relationship or a non-same-group relationship.

In the present invention, preferably, the correlation relationship is a multi-layer grouping relationship, that is, the data set D is divided into a plurality of 1 st layer groups, each 1 st layer group can be divided into a plurality of 2 nd layer groups, and so on, each L th layer group can be divided into a plurality of L +1 th layer groups, and the elements have the same or different L th layer group relationship.

In the present invention, preferably, when the elements of the data set D have an inherent natural spatial (e.g., 1-dimensional space or 2-dimensional space or N-dimensional space) positional relationship, the mutual relationship is a spatial positional relationship, such as a neighborhood relationship, a neighborhood relationship with a predetermined condition, a certain distance relationship, or a certain distance relationship with a predetermined condition.

In the present invention, preferably, the inter-correlation system is a sequential relationship in the encoding and decoding process, such as a relationship of several nearest elements, a relationship of several nearest elements satisfying some predetermined conditions.

In the present invention, it is preferable that the correlation system is a combination, composite, mixture, or fusion of the above-described various grouping relationships, multilayer grouping relationships, spatial position relationships, order relationships, and the like.

The most basic unique technical characteristic of the encoding method or device is that the inherent natural mutual correlation among the elements of the data set and/or the mutual correlation (such as grouping relation, spatial position relation and sequence relation) formed in the encoding and decoding process are fully mined and utilized, the current actual value range of the elements is dynamically and adaptively determined according to the preset rules, the current actual value range is smaller than the original value range, the reduced value range is used for carrying out related encoding function and operation, and a compressed data code stream at least containing information which can be used for directly or indirectly or directly and indirectly mixedly representing or deriving the current actual value range is generated. Fig. 1 is a schematic diagram of an encoding method or apparatus of the present invention.

The most basic special technical characteristic of the decoding method or device is that a compressed data code stream at least containing information which can be used for directly or indirectly or directly and indirectly mixedly representing or deriving what the current actual value range is analyzed, the current actual value range of the element which is smaller than the original value range is dynamically and adaptively determined according to a preset rule by utilizing the internal natural mutual correlation between the elements of the data set and/or the mutual correlation (such as grouping relationship, spatial position relationship and sequential relationship) formed in the encoding and decoding process, and the reduced value range is used for carrying out related decoding function and operation. Fig. 2 is a schematic diagram of a decoding method or apparatus of the present invention.

According to an aspect of the present invention, there is provided an encoding method or apparatus or a decoding method or apparatus for compressing a data set, comprising at least the steps or modules of:

establishing an interrelationship between elements of the data set, including an interrelationship based on inherent natural relationships of various data or parameters and/or an interrelationship formed during a coding/decoding process;

dynamically (i.e., continuously) determining or selecting or forming a current actual value range of the element using the correlation;

using the current actual value range to perform related coding or decoding functions and operations;

the compressed data code stream at least contains information which can be used for directly or indirectly or directly and indirectly representing or deriving what the current actual value range is.

From a first aspect, the invention provides a coding method for compressing a data set and its data, characterized in that it comprises at least the following steps:

1) establishing an interrelationship between elements of the data set, including an interrelationship based on inherent natural associations of various data or parameters and/or an interrelationship formed during an encoding process;

2) dynamically (i.e. continuously) determining or selecting or forming the current actual value range of the data in the current code by utilizing the correlation system;

3) encoding data in the current encoding by using the current actual value range;

4) and generating a compressed data code stream at least containing information which can be used for directly or indirectly or directly and indirectly mixedly representing or deriving what the current actual value range is.

From a second aspect, the present invention provides a decoding method for compressing a data set and its data, characterized by comprising at least the following steps:

1) analyzing the compressed data code stream to obtain information which can be used for directly or indirectly or directly and indirectly mixedly representing or deriving what the current actual value range is;

2) establishing an interrelationship between elements of the data set, including an interrelationship based on inherent natural relationships of various data or parameters and/or an interrelationship formed during a decoding process;

3) dynamically (i.e. continuously) determining or selecting or forming a current actual value range of the data in the current decoding according to the information and the correlation relationship;

4) and decoding the data in the current decoding by using the current actual value range.

From a third perspective, the present invention provides an encoding apparatus for compressing a data set and data thereof, characterized by comprising at least the following modules:

1) a data set element correlation system construction module for establishing the correlation system among the elements of the data set, which comprises the correlation system based on the inherent natural connection of various data or parameters and/or the correlation system formed in the encoding process;

2) a current actual value range forming module, which dynamically (i.e. continuously changes) determines or selects or forms the current actual value range of the data in the current code by using the correlation system;

3) the core coding module is used for coding the data in the current coding by using the current actual value range;

4) and the compressed data code stream generating module is used for generating a compressed data code stream at least containing information which can be used for directly or indirectly or directly and indirectly mixedly representing or deriving what the current actual value range is.

From a fourth perspective, the present invention provides a decoding apparatus for compressing a data set and data thereof, characterized by comprising at least the following modules:

1) the compressed data code stream analyzing module is used for analyzing the compressed data code stream to obtain information which can be used for directly or indirectly or directly and indirectly mixedly representing or deriving what the current actual value range is;

2) a data set element correlation system construction module for establishing correlation systems among elements of the data set, including the correlation systems based on the inherent natural relations of various data or parameters and/or the correlation systems formed in the decoding process;

3) a current actual value range forming module, which dynamically (i.e. continuously changes) determines or selects or forms the current actual value range of the data in the current decoding according to the information and the correlation system;

4) and the core decoding module is used for decoding the data in the current decoding by using the current actual value range.

In the case where the original data is image or video data, when compressing the image and video data, a frame of image is usually divided into coding _ regions or compression _ regions having a predetermined constant size (e.g., 64 × 64 pixels or 128 × 128 pixels) or having one or several rectangular region sizes (e.g., K × L + M × N pixels, where K, L, M, N are positive integers), each coding block usually has a unique coding block number, and each coding block is further divided into coding blocks (also called decoding blocks from the decoding point of view, often collectively called coding blocks or collectively called whole compression units) according to a quartering or bifurcating structure of one or more recursions, each coding block usually has a unique coding block number, the coding and decoding block number is allowed to be composed of two parts of a coding block number and a coding and decoding block relative number.

In the present invention, where the original data is image or video data, the elements of the data set include one of the following codec parameters appearing in the image or video codec or a combination, composite or fusion of various forms or relationships thereof, including parallel, cross, include, dependent, merge, partial merge, cross merge, union, intersection, difference, complement:

1) a code tree branching pattern whose value range includes part or all of the following set 1 and/or set 2 as represented by enumeration:

the set 1 ═ log of the logarithm to the base 2 of the partSize, since the size of the partition is always an index of 2, of the partSize, of the partstructure, { various branching structures which first or several bifurcations, then of the bifurcations or trifurcations }, any of which can obviously be expressed in terms of the number of times the coding tree unit starts a recursive bifurcations and the number of times a recursive bifurcations or trifurcations are carried out subsequently, or equivalently in terms of the size of the split after a recursive bifurcations, and the size of the split after a recursive bifurcations or trifurcations subsequently, of the partSize₂(partSize) to indicate that,

the set 2 is { no bifurcation, four bifurcations, horizontal trifurcations, vertical trifurcations, horizontal bifurcations, vertical bifurcations };

2) a coding tree luminance component bifurcation pattern whose value range includes a part or all of the following set 1 and/or set 2, which are expressed by enumeration:

the set 1 is { various bifurcated structures that bifurcate one or more times first and bifurcate one or more times second or third },

3) a coding tree chroma component bifurcation pattern whose range of values includes a portion or all of set 1 and/or set 2 as enumerated:

set 2 ═ nondividing, quartering, horizontal trifurcating, vertical trifurcating, horizontal bifurcating, vertical bifurcating }

4) And encoding a tree luminance and chrominance fused bifurcation mode, wherein the value range of the tree luminance and chrominance fused bifurcation mode comprises part or all of the set 1 and/or the set 2 which are expressed by enumeration as follows:

set 1 ═ luminance and chrominance have various branching structures in common, i.e., a single code tree, luminance and chrominance have various branching structures in respective code trees, i.e., the luminance code tree and the chrominance code tree are at least partially different, luminance and chrominance have various branching structures in a code tree having a common, i.e., a single trunk portion, but having respective branch portions, and luminance and chrominance have various branching structures in a code tree having a common, i.e., a single one-time or multiple-time four-branch trunk portion, but having respective one-time or multiple-time two-branches or three-branch portions, respectively },

set 2 ═ luminance and chrominance are neither split, luminance is four-split but chrominance is not split, luminance is horizontal is three-split but chrominance is not split, luminance is vertical is three-split but chrominance is not split, luminance is horizontal is two-split but chrominance is not split, luminance is not split but chrominance is four-split, luminance is four-split but chrominance is four-split, luminance is horizontal is three-split but chrominance is four-split, luminance is vertical is three-split but chrominance is four-split, luminance is horizontal is two-split but chrominance is four-split, luminance is vertical is three-split but chrominance is four-split, luminance is four-split but chrominance is vertical is two-split, and so on };

5) coding modes of the coding and decoding blocks, wherein the value range of the coding modes comprises a part or all of the following sets represented by enumeration: { IPCM, intra prediction, inter prediction, string prediction, palette, block prediction, micro block prediction, strip prediction };

4) and 5), namely a luminance and chrominance bifurcation mode of the coding tree and a coding mode of the coding and decoding block, and the value range of the luminance and chrominance bifurcation mode of the coding tree and the coding mode of the coding and decoding block comprises a part or the whole of the following set expressed by enumeration: { intraframe coding mode with luminance and chrominance having various branching structures of a coding tree common to a single coding tree, intraframe coding mode with luminance and chrominance having various branching structures of a coding tree respective to a coding tree, intraframe coding mode with luminance and chrominance having various branching structures of a coding tree common to a single trunk portion but respective to branch portions, intraframe coding mode with luminance and chrominance having various branching structures of a coding tree common to a single one-time or multiple-time four-branch trunk portion but respective to one-time or multiple-time two-branch or three-branch portions, interframe coding mode with luminance and chrominance having various branching structures of a coding tree common to a single coding tree, interframe coding mode with luminance and chrominance having various branching structures of a coding tree respective to a single trunk portion but respective branch portions, inter-frame coding mode and luma and chroma have various branching structures of a coding tree that are common, i.e., a single one or more-time four-branch trunk portion, but have respective one or more-time two-branch or three-branch portions };

6) coding mode associated coding parameters of the coding and decoding block;

7) the intra luminance prediction mode of the coding and decoding block comprises a part or a whole of one of the following 7 sets respectively expressed by enumeration:

(1) { DC, Planar, direction 1, direction 2, and so on, direction 65 };

(2) { DC, Planar, Biliner, Direction 1, Direction 2, and so on, Direction M }, where M is typically a constant between 30 and 130;

(3) { DC, Planar, direction 1, direction 2, and so on, direction M, extension 1, extension 2, and so on, extension N }, where M and N are each typically constants between 30 and 130;

(4) a plurality of elements dynamically selected from a mode set (such as { DC, Planar, direction 1, direction 2, and so on, direction 65 }) according to a predetermined rule (such as the value of the intra-frame brightness prediction mode of a plurality of adjacent coding and decoding blocks of the current coding and decoding block or the value of the intra-frame brightness prediction mode of a plurality of coding and decoding blocks closest to the coding and decoding sequence);

(5) a mode set (such as { DC, Planar, direction 1, direction 2, and so on, direction 65 }) minus a difference set obtained by dynamically selecting a plurality of elements according to a predetermined rule (such as the value of the intra-frame brightness prediction modes of a plurality of adjacent coding and decoding blocks of the current coding and decoding block or the value of the intra-frame brightness prediction modes of a plurality of coding and decoding blocks closest to the coding and decoding sequence);

(6) taking the value ranges (such as { element 0, element 4, element 8, and the like } of all elements with the sequence numbers being multiples of a predetermined constant (such as: 4) from the difference set (5) as an ordered set { element 0, element 1, element 2, element 3, element 4, element 5, element 6, element 7, element 8, and the like });

(7) taking all value ranges (e.g., { element 1, element 2, element 3, element 5, element 6, element 7, element 8, and so on } with serial numbers not being multiples of a predetermined constant (e.g.,: 4) composed of elements from the difference set of (5) as an ordered set { element 0, element 1, element 2, element 3, element 4, element 5, element 6, element 7, and so on });

8) the prediction mode associated coding parameters of the coding and decoding block comprise motion vectors of inter prediction and/or intra block prediction, motion vector differences, offset of the motion vector differences, motion vector candidate set indexes, prediction values, interpolation of the prediction values and interpolation weights of the prediction values:

9) the prediction mode of the coding and decoding block between frames and/or the prediction mode containing the intra-frame block, the value range of which comprises a part or the whole of one of the following 3 sets which are respectively expressed by enumeration;

(1) { Skip, Merge candidate 1, Merge candidate 2, and so on, Merge candidate I, the conventional mode directly represented by motion vectors and/or motion vector differences and/or offsets of motion vector differences }, where I is typically a constant between 2 and 10;

(2) { Skip, Merge candidate 1, Merge candidate 2, and so on, Merge candidate I, Affine1, Affine2, and so on, AffineJ, conventional mode directly represented by motion vectors and/or motion vector differences and/or offsets of motion vector differences }, where I is typically a constant between 2 and 10, and J is typically a constant between 1 and 5;

(3) { Skip, Merge candidate 1, Merge candidate 2, and so on, Merge candidate I, escape 1, escape 2, and so on, AffineJ, extension 1, extension 2, and so on, extension K, conventional mode that directly represents the amount of motion or amount of motion with motion vectors and/or motion vector differences and/or offsets of motion vector differences }, where I is typically a constant between 0 and 10 (I of 0 means no Merge candidate element), J is typically a constant between 0 and 5 (J of 0 means no Merge element), K is typically a constant between 0 and 8 (K of 0 means no extension element), extension includes one or a combination of: the recent history sub-mode is that the related coding parameters of the recent one or a plurality of coding and decoding blocks appearing in the coding and decoding process are used as the related coding parameters of the current coding and decoding block, a line sub-mode, a half-line sub-mode and a micro-block sub-mode;

8) the inter prediction mode in the middle motion vector and 9), i.e., the multi-source motion vector predictor (including the case where the prediction difference is zero, i.e., the motion vector predictor is equal to the motion vector itself), whose value range includes a part of or the union of all the following sets or subsets thereof:

set 1-the motion vector and/or motion vector difference and/or offset candidate set of motion vector differences for the Skip sub-mode,

set 2 being the motion vector and/or motion vector difference and/or offset candidate set of motion vector differences for the Merge sub-mode,

set 3-the motion vector and/or motion vector difference and/or offset candidate set of motion vector differences for the most recent history sub-mode,

set 4-a predetermined default motion vector and/or motion vector difference and/or offset candidate set of motion vector differences,

set 5 is a set of conventional motion vectors directly denoted by (horizontal component, vertical component);

10) inter-component prediction mode of the coded block;

11) inter-component prediction mode-dependent coding parameters of the coded block;

12) inter-component prediction coefficients of the coded block;

13) the inter-component prediction coefficient of the coding and decoding block is associated with a coding parameter;

14) residual component inter-prediction mode of the coded block;

15) residual component inter-prediction coefficients of the coded and decoded blocks;

16) an inter-component color transform pattern of a coded block whose range of values includes a portion or all of the following set, as enumerated: { RGB-YCoCg transform, GBR-YCoCg transform, BRG-YCoCg transform, RGB-YCoCg transform variant 1, GBR-YCoCg transform variant 1, BRG-YCoCg transform variant 1, RGB-YCoCg transform variant 2, GBR-YCoCg transform variant 2, BRG-YCoCg transform variant 2, RGB-YCoCg transform variant 3, GBR-YCoCg transform variant 3, BRG-YCoCg transform variant 3, and so on, examples of variants including fine tuning of transform coefficients;

17) inter-component color transform mode dependent coding parameters for a coding/decoding block

18) Encoding and decoding inter-component color transform coefficients of the block;

19) the inter-component color transform coefficient of the coding and decoding block is associated with a coding parameter;

20) coding and decoding the residual component color transformation mode of the block, wherein the value range of the residual component color transformation mode comprises a part or the whole of the following set represented by enumeration: { RGB-YCoCg transform, GBR-YCoCg transform, BRG-YCoCg transform, RGB-YCoCg transform variant 1, GBR-YCoCg transform variant 1, BRG-YCoCg transform variant 1, RGB-YCoCg transform variant 2, GBR-YCoCg transform variant 2, BRG-YCoCg transform variant 2, RGB-YCoCg transform variant 3, GBR-YCoCg transform variant 3, BRG-YCoCg transform variant 3, and so on, Cb-Cr transposition after chroma residual inverse transformation, examples of variants including fine tuning of the transform coefficients };

21) the inter-residual component color transform coefficients of the coded block or their associated coding parameters, such as: chroma residual component transform coefficients;

22) the residual of block prediction (including intra prediction and inter prediction) of a coded block;

23) coding and decoding transform coefficients of a residual of a block prediction of a block;

24) encoding parameters of string prediction of the encoded and decoded block;

25) string length of string prediction (also called string matching or string replication) of a codec block;

26) string displacement of string prediction of the coded and decoded block;

27) a string displacement horizontal component of a string prediction of a coded block;

28) string displacement vertical component of string prediction of the coding and decoding block;

29) a string type flag of a string prediction of a codec block;

30) a sub-Reference Buffer (SRB) address (referred to as a string SRB address) of a string of string prediction of the codec block;

31) string SRB color repetition number of string prediction of the coded block;

32) encoding parameters of the image coordinate string;

33) the number of consecutive image coordinate strings;

34) length of the image coordinate string;

35) the number of repetitions of a pixel at an image coordinate in the image coordinate string;

36) the address of the image coordinate string in an image coordinate array (PCA);

37) a coding parameter of the offset string;

38) the number of consecutive offset strings;

39) the length of the offset string.

In the present invention, in the case where original data is image or video data and a frame of image is divided into coding blocks, which are an area of the image, the following cases are included: the coding method comprises the following steps of sub-images of an image, a stripe slice, a stripe group consisting of a plurality of stripes, a tile, a tile group consisting of a plurality of tiles, a brick crack, a brick group consisting of a plurality of bricks, a maximum coding unit LCU, a maximum coding unit group consisting of a plurality of maximum coding units, a coding tree unit CTU and a coding tree unit group consisting of a plurality of coding tree units.

In the present invention, in the case where original data is image or video data and a frame of image is further divided into coding and decoding blocks which are one coding area or one decoding area of the image, the following cases are included: a sub-picture of a picture, a macroblock, a largest coding unit LCU, a coding tree unit CTU, a coding unit CU, a sub-region of a CU, a sub-coding unit SubCU, a prediction unit PU, a sub-region of a PU, a sub-prediction unit SubPU, a transform unit TU, a sub-region of a TU, a sub-transform unit SubTU.

The technical features of the present invention are explained above by specific embodiments. Other advantages and effects of the present invention will be readily apparent to those skilled in the art from the disclosure herein. The invention is capable of other and different embodiments and of being practiced or of being carried out in various ways, and its several details are capable of modification in various respects, all without departing from the spirit and scope of the present invention.

Detailed Description

The following are further implementation details or variations of the present invention.

Implementation or variant 1 (embodiment on the correlation system) said encoding method or device or said decoding method or device, said correlation system being one of the following relationships or a combination, composite or fusion of their various forms (for example: parallel, cross, include, dependent, merge, partial merge, cross-merge, union, intersection, difference, complement):

grouping relationships, i.e., the data set is divided into groups, elements have a same group relationship or a non-same group relationship,

or a multi-layer grouping relationship, i.e. the data set is divided into a plurality of 1 st layer groups, each 1 st layer group can be divided into a plurality of 2 nd layer groups, and so on, each L < th > layer group can be divided into a plurality of L < th > +1 < th > layer groups, and the elements have the same L < th > (L =1, 2,3, … …) layer group relationship or the same L < th > layer group relationship,

or spatial positional relationships, i.e., natural spatial (e.g., 1-dimensional or 2-dimensional or N-dimensional) positional relationships that elements of the data set may have, including

Adjacent relationships (e.g., an element in a lower left adjacent position, an element in a left adjacent position, an element in an upper right adjacent position of an element), elements in an adjacent relationship are referred to as adjacent elements, elements in a lower right adjacent position, elements in an upper left adjacent position, elements in an upper right adjacent position, elements in,

A neighborhood relationship having a predetermined condition (e.g., a condition related to a grouping relationship),

Proximity relations (e.g., adjacent neighbors), elements with proximity relations referred to as neighboring elements,

A proximity relation having a predetermined condition,

Distance relationships (e.g. distance being one of a number of predetermined values),

There is a distance relationship of a predetermined condition,

or coding and decoding order relation, namely the appearance sequence relation of each element in the coding and decoding process, comprising

A closest relation, i.e. the closest element or elements occurring before an element, the elements having the closest relation being called the closest elements,

A nearest N relation having a predetermined condition (e.g., a condition relating to an adjacent relation),

The most recent N relation, i.e. the most recent N (N being a predetermined number, e.g. 7) elements occurring before an element, the element having the most recent N relation being called the most recent N element,

The nearest N relationship having a predetermined condition (e.g., a condition related to a neighbor relationship).

EXAMPLES OR VARIATIONS 2 (examples relating to the interrelationship)

In the encoding method or apparatus or the decoding method or apparatus, the inter-correlation relationship is formed of a grouping relationship (referred to as an original data-based grouping relationship) formed at least in the following manner:

the original data is divided into original data blocks (e.g., coded blocks of image or video data), each of which typically has a unique original data block number (e.g., a coding block number),

each original data block is further divided into original data blocks (e.g., coded blocks of image or video data), each original data block typically having a unique original data block number (e.g., a coded block number),

the original data block number is typically comprised of two parts, an original data block number and an original data block relative number (e.g., a codec block relative number),

the number of the original data block, the number of the original data block and the relative number of the original data block are generally increased from small to large respectively according to the sequence of encoding and decoding,

obviously, two original data blocks having the same original data block number belong to the same original data block;

in the original data, every C (C is usually a predetermined constant between 1 and 50) original data blocks constitute one original data block group (for example, every sxt encoding block constitutes one encoding block group), and it is obvious that when C =1, each original data block group has only one original data block;

each original data block group typically has a unique original data block group number (e.g., coding block group number), which typically increases from small to large with the increasing relationship of the smallest original data block number within the original data block group;

each element of the data set corresponds to an original data block or a number thereof;

the elements of the data set form a group relationship and are endowed with group numbers, element numbers and element relative numbers according to the original data blocks corresponding to the elements of the data set or the original data block groups to which the numbers of the elements belong, namely the elements of the original data blocks corresponding to the elements of the data set or the numbers of the elements of the original data blocks in the same original data block group have the same group relationship, and the elements which are not in the same original data block group have non-same group relationship; the group number, the element number and the element relative number are respectively equal to the original data block group number, the original data block number and the original data block relative number.

EXAMPLES OR VARIATIONS 3 (examples relating to the interrelationship)

In the encoding method or apparatus or the decoding method or apparatus according to embodiment or variation 2, the original data is image or video data, the original data block is a coded block, and the original data block is a coded block;

c is sxt (S and T are typically predetermined constants between 1 and 10, respectively).

Examples of embodiment or variants 4 (correlation relationship)

In the encoding method or apparatus or the decoding method or apparatus, the inter-correlation relationship is formed of a multi-layer packet relationship (referred to as an original data-based multi-layer packet relationship) formed at least in the following manner:

in the original data, every C (C is usually a predetermined constant between 1 and 50) original data blocks constitute one layer 1 original data block group (e.g., every sxt encoding block constitutes one layer 1 encoding block group), and it is obvious that when C =1, each layer 1 original data block group has only one original data block;

decomposing said C into N₁×C₁(e.g., N)₁ = C，C₁= 1), further divide layer 1 original data block group into N₁A group of layer 2 original data blocks, obviously, each group of layer 2 original data blocks is formed by C₁The original data blocks are formed;

subjecting said C to₁Decomposition to N₂×C₂(e.g., N)₂ = C₁，C₂= 1), the layer 2 original data block group is further divided into N₂A group of layer 3 original data blocks, obviously, each group of layer 3 original data blocks is formed by C₂The original data blocks are formed;

and so on, the C_L(L ═ 1, 2,3, … …) decomposition to N_L+1×C_L+1(e.g., N)_L+1 = C_L，C_L+1= 1), further dividing L-th layer original data block group into N_L+1An L +1 layer original data block group, obviously, each of the L +1 layer original data block groups is composed of C_L+1The original data blocks are formed;

each L-th layer original data block group typically has a unique L-th layer original data block group number (e.g., L-th layer coding block group number), which typically increases from small to large with increasing relation to the smallest original data block number within the L-th layer original data block group;

the elements of the data set form an L-th layer grouping relationship according to the corresponding original data blocks or the L-th layer original data block group to which the numbers of the original data blocks belong to form an L-th layer group, and the L-th layer group is formed, and the number of the L-th layer group, the group number, the element number and the element relative number are given, that is, the corresponding original data blocks or the elements of which the numbers are in the same L-th layer original data block group have an L-th layer same group relationship, and the elements which are not in the same L-th layer original data block group have a different L-th layer group relationship, wherein L =1, 2,3, … …; the number of the L-th layer group, the group number, the element number and the element relative number are respectively equal to the number of the L-th original data block group, the number of the original data block and the element relative number.

EXAMPLES OR VARIATIONS 5 (examples relating to the interrelationship)

In the encoding method or apparatus or the decoding method or apparatus according to embodiment or variation 4, the original data is image or video data, the original data block is a coded block, and the original data block is a coded block;

c is S × T (S and T are typically predetermined constants between 1 and 10, respectively);

s is equal to U₁×S₁(e.g., U)₁ = S，S₁= 1), said T being equal to V₁×T₁(e.g., V)₁ = T，T₁= 1), said N₁Is equal to U₁×V₁Said C is₁Is equal to S₁×T₁；

Said S₁Is equal to U₂×S₂(e.g., U)₂ = S₁，S₂ = 1），The T is₁Is equal to V₂×T₂(e.g., V)₂ = T₁，T₂= 1), said N₂Is equal to U₂×V₂Said C is₂Is equal to S₂×T₂；

And so on, S_L(L ═ 1, 2,3, … …) equal to U_L+1×S_L+1(e.g., U)_L+1 = S_L，S_L+1= 1), said T_LIs equal to V_L+1×T_L+1(e.g., V)_L+1 = T_L，T_L= 1), said N_L+1Is equal to U_L+1×V_L+1Said C is_L+1Is equal to S_L+1×T_L+1。

EXAMPLES OR VARIATIONS 6 (examples relating to the interrelationship)

In the encoding method or apparatus or the decoding method or apparatus, the original data is image or video data, each frame of image of the image or video is divided into encoding blocks, and each encoding block is further divided into encoding and decoding blocks;

each element of the data set corresponds to a codec block or number thereof;

in a picture, every sxt (S and T are typically a predetermined constant between 1 and 10, respectively) coding blocks constitute one coding block group, obviously, when S = T =1, there is only one coding block per coding block group;

the elements of the data set form a grouping relation according to the coding block group to which the corresponding coding and decoding block or the number thereof belongs, namely the elements of the coding and decoding block or the number thereof corresponding to the elements of the data set in the same coding block group have the same group relation, and the elements not in the same coding block group have the non-same group relation;

the inter-correlation is made up of at least the grouping relationships.

EXAMPLES OR VARIATIONS 7 (examples relating to the interrelationship)

In the encoding method or apparatus or the decoding method or apparatus,

the original data is image or video data, each frame of image of the image or video is divided into coding blocks, each coding block is divided into coding and decoding blocks, and the coding and decoding blocks are divided into V primitives with relative sequence numbers V (V is more than or equal to 1 and less than or equal to V) (the types of the primitives comprise strings, lines, bars, micro blocks, sub-regions of CU, sub-coding units SubCU, sub-regions of PU, sub-prediction units SubPU, sub-regions of TU and sub-transformation units SubTU; V is a constant or a variable, and V =1 represents that only one primitive which is equal to the coding and decoding block exists;

each element of the dataset corresponds to a primitive or its sequence number;

the elements of the data set form a grouping relation according to the coding region block group to which the corresponding primitive or the serial number belongs, namely the corresponding primitives or the elements with the serial numbers in the same coding region block group have the same group relation, and the elements not in the same coding region block group have non-same group relation;

the inter-correlation is made up of at least the grouping relationships.

EXAMPLES OR VARIATIONS 8 (examples relating to the interrelationship)

In the encoding method or apparatus or the decoding method or apparatus,

the original data is image or video data, each frame of image of the image or video is divided into coding blocks, and each coding block is further divided into coding and decoding blocks;

each element of the data set corresponds to a codec block or number thereof;

the elements of the data set form an adjacent relationship according to the relative positions of the corresponding coding and decoding blocks or the numbers thereof in the image (such as an element positioned at the lower left adjacent position, an element positioned at the upper right adjacent position and an element positioned at the upper right adjacent position of one element);

the elements of the data set form a closest relationship according to the sequence of the corresponding coding and decoding blocks or the numbers thereof in the coding and decoding process (such as the latest N elements appearing before one element, wherein N is a preset number, such as 7);

the inter-relation is made up of at least one of the grouping relationship, the adjacent relationship, the closest relationship, or a combination, composite, or fusion thereof.

Examples of embodiment or variants 9 (examples of actual value ranges at present)

In the encoding method or apparatus or the decoding method or apparatus,

the current actual value range is a subset of the total value range of the data set, and is composed of or generated by one of a plurality of types of sub-ranges or combination or composite or fusion (for example, parallel, cross, including, dependent, combined, partially combined, cross combined, intersected, difference set and complementary set) thereof; the several types include at least the following four types:

sub-ranges of fixed type: a predetermined number of fixed subsets of the total range of values (e.g. a predetermined I number of fixed subsets R of the total range of values R)₁

R，R₂

R，R₃

R，……，R_I

R), each fixed subset is typically referred to as a fixed sub-range,

or sub-ranges of adjacent types: a set of values taken by neighbouring elements of an element (e.g. a neighbouring sub-range R of values taken by a number of predetermined neighbouring elements of element d_neighbor(d) Which may itself obviously vary with d), is often referred to as a neighboring sub-range,

or a sub-range of the most recent N types: a set of values taken by the most recent N elements of an element (e.g., a most recent N subrange R of values taken by a number of predetermined most recent N elements of an element d_recentN(d) The nearest N subrange itself may obviously vary with d), often referred to as the nearest N subrange;

or subranges of close-valued types: a set of similar values (e.g., a set of values that differ by ± 1 and/or ± 2), often referred to as a similar value sub-range;

or subranges of like ordinal types: the total range is an ordered set, each element has a sequence number, the elements are arranged according to the sequence number, and a set (for example, a set consisting of values with sequence numbers differing by + -1 and/or + -2) which are close to each other (for example, the absolute value of the sequence number difference is smaller than a constant) is generally called a close sequence number sub-range;

and coding and decoding the elements by using the code word set adaptive to the size of the current actual value range.

Examples of embodiment or variants 10 (examples of actual value ranges at present)

In the encoding method or apparatus or the decoding method or apparatus according to embodiment or variation 9, the current actual value range is one of the following cases:

the current actual value range is a fixed sub-range;

the current actual value range is an adjacent sub-range;

the current actual value range is a nearest N sub-range;

the current actual value range consists of adjacent sub-ranges and one (or a plurality of) fixed sub-ranges;

the current actual value range is composed of adjacent sub-ranges and similar value sub-ranges (for example, is a union of the adjacent sub-ranges and the similar value sub-ranges of the elements of the adjacent sub-ranges);

the current actual value range consists of adjacent sub-ranges and close sequence number sub-ranges (for example, is a union of the adjacent sub-ranges and the close sequence number sub-ranges of the elements of the adjacent sub-ranges);

the current actual value range consists of adjacent sub-ranges and the nearest N sub-ranges;

the current actual value range consists of a nearest N sub-range and one (or a plurality of) fixed sub-ranges;

the current actual value range is composed of adjacent sub-ranges, the nearest N sub-ranges and one (or a plurality of) fixed sub-ranges.

Examples or variants 11 (examples of the actual value ranges at present)

In the encoding method or apparatus or the decoding method or apparatus, the current actual value range is a variation or an extension of the current actual value range in implementation or variation example 11, for example, a variation obtained by replacing a fixed sub-range and/or an adjacent sub-range and/or a nearest N sub-range and/or a similar value sub-range with another type of sub-range.

Examples of embodiment or variations 12 (examples in which the actual value range and the correlation relationship are relevant at present)

In the encoding method or device or the decoding method or device, the current actual value range of the element is at least related to the correlation between the elements.

Examples of implementation or modification 13 (examples in which the actual value range and the correlation are relevant at present)

In the encoding method or apparatus or the decoding method or apparatus, the current actual value range of the element is at least associated with an interrelation between the elements, and the association includes one or a combination of the following cases:

associated case 1: the elements at least have grouping relation, the elements have element numbers, the element numbers are increased from small to large according to the sequence of coding and decoding, in one group,

if the element number is equal to one or several predetermined values

{ its current actual value range R_aIs the total range of values of the data set or a subset of the total range of values,

using a compound of formula (I) and (II)_aThe adaptive size codeword set encodes and decodes the elements,

else

{ its current actual value range R_bOne or a combination of the following conditions is satisfied:

R_bis smaller than R_aThe size of (a) is (b),

R_a－R_b（R_sand R_bDifference set of) is R_aIs a proper subset of the total number of cells,

using a compound of formula (I) and (II)_bThe adaptive size codeword set encodes and decodes the elements,

}

examples of the one or several predetermined values include the following cases:

case 1 of the predetermined value: the minimum value in the group, that is, the element is the first element to be coded and decoded in the group to which the element belongs;

case 2 of the predetermined value: 4 times of;

case 3 of the predetermined value: a multiple of a predetermined integer;

the R is_aAnd R_bExamples of (c) include the following:

R_aand R_bCase 1: r_aIs the total value range, R, of the data set_bIs an adjacent subrange;

R_aand R_bCase 2: r_aIs the total value range, R, of the data set_bIs a contiguous sub-range ≦ a fixed sub-range;

R_aand R_bCase 3: r_aIs the total value range, R, of the data set_bIs a sub-range of adjacent sub-ranges, U.S. a fixed sub-range, U.S. a near value sub-range;

associated case 2: the elements of the data set D are divided into M groups: d_mWherein M is more than or equal to 1 and less than or equal to M;

defining I sub-ranges R in said value range R₁

R，R₂

R，R₃

R，……，R_I

R；

D_mThe current actual value ranges of all the elements d are all sub-ranges R with the same subscript (namely, sub-range index) being i (m)_i(m)I.e. by

d ∈D_mThe current actual value range of d is R_i(m)；

If D ∈ D_mThen, then

{ d is currently the actual value range R_i(m)，

Using a compound of formula (I) and (II)_i(m)D is coded and decoded by the code word set with the adaptive size; }

The compressed data code stream at least contains information directly or indirectly or directly and indirectly mixed to represent the value of the M sub-range indexes i (M), and the information directly representing the value of the sub-range indexes i (M) is formed by one or more than one compressed data code streamA bit string (bit string) consisting, for example, of an identification code or syntax element called ctu _ library _ index; said information indirectly representing the value of the sub-range index i (m) is information derived from other coding parameters and/or codec variables and/or other syntax elements of the compressed data stream, e.g. from said D_mInformation derived from values of sub-range indices of groups having a neighboring relationship or a nearest N relationship and/or other coding parameters and/or other codec variables; the information directly and indirectly mixedly representing the value of the sub-range index i (m) is information partially directly (i.e. consisting of one or more bit strings in the compressed data code stream) and partially indirectly (i.e. derived from other coding parameters and/or codec variables and/or other syntax elements of the compressed data code stream) mixedly representing the value of the sub-range index i (m);

and the like, performing operations of dividing layer groups 2,3, 4 and … …, specifying sub-ranges of the layers 3, 4 and … …, and limiting the current actual value range of the element to a sub-range for encoding and decoding the element.

Examples of embodiment or modification 14

In the encoding method or apparatus or the decoding method or apparatus according to embodiment or variation 13, the data set D is divided into small groups;

the group or first group of groups D_mConsists of C of said subgroups;

c is equal to N₁×C₁(e.g., N)₁ = C，C₁= 1), second layer group D_mnFrom C₁(ii) each of said subgroups;

and so on, the C_L＝N_L+1×C_L+1(L-1, 2,3, … …) group L +1 group is composed of C_LEach of said subgroups.

Implementation or variation 15 implementation or variation 14 the encoding method or apparatus or the decoding method or apparatus wherein the original data is image or video data, and C is sxt (S and T are typically predetermined constants between 1 and 10, respectively);

Said S₁Is equal to U₂×S₂(e.g., U)₂ = S₁，S₂= 1), said T₁Is equal to V₂×T₂(e.g., V)₂ = T₁，T₂= 1), said N₂Is equal to U₂×V₂Said C is₂Is equal to S₂×T₂；

Examples of embodiment or modification 16

In the encoding method or apparatus or the decoding method or apparatus, the original data is image or video data, each frame of the image or video is divided into encoding tree units CTUs, each encoding tree unit CTU is further divided into encoding and decoding blocks according to a quartering or trisections or dichotomy or non-bifurcation structure of one or more recursions, each element of the data set corresponds to one encoding and decoding block or a number thereof, obviously, each encoding and decoding block belongs to one encoding tree unit;

the elements of the data set form a grouping relationship according to the coding tree unit to which the corresponding coding and decoding blocks belong, namely the elements of the corresponding coding and decoding blocks in the same coding tree unit have the same group relationship, the elements not in the same coding tree unit have the non-same group relationship, and the inter-group relationship is formed by at least the grouping relationship;

the data to be coded and decoded is inter-frameIn coded frames (also called inter-coded pictures, i.e. frames or pictures allowing to encode/decode a coded block using both inter-coding and intra-coding modes)Coding module of coding and decoding block Formula (II)And the value range includes a part or all of the following set represented by enumeration:

{ Intra-frame coding mode, inter-frame coding mode, string prediction coding mode, palette coding mode, Intra-frame Block matching coding mode }

A particular practical range a is one of the following:

the particular actual value range A is { intra-coding mode }

Or the specific actual value range A is { intra-frame coding mode, string prediction coding mode }

Or the specific actual value range A is { intra-frame coding mode, palette coding mode }

Or the specific actual value range a is { intra coding mode, intra block matching coding mode }

Or the specific actual value range A is { intra-frame coding mode, string prediction coding mode, palette coding mode }

Or the specific actual value range A is { intra-frame coding mode, string prediction coding mode, intra-frame block matching coding mode }

Or the specific actual value range A is { intra coding mode, palette coding mode, intra block matching coding mode }

Or the specific actual value range a is { intra-frame coding mode, string prediction coding mode, palette coding mode, intra-frame block matching coding mode };

the compressed data code stream at least contains information which can be used for directly or indirectly or directly and indirectly representing or deriving whether the current actual value range is the specific actual value range A or not in a mixed manner.

Examples or modifications 17

In the encoding method or apparatus or the decoding method or apparatus according to implementation or modification example 16, the compressed data code stream at least includes an encoding mode flag bit of an encoding tree unit, where the flag bit indicates or derives whether a current actual value range of an encoding mode of an encoding block or a decoding block of the encoding tree unit is the specific actual value range a in the following manner:

if the flag bit is equal to a predetermined value (e.g., a value of 1), the current actual value range of the coding mode of the coding/decoding block of the coding tree unit is the specific actual value range a.

Implementation or modification 18 implementation or modification 16 said encoding method or apparatus or said decoding method or apparatus, wherein the specific actual value range a is an { intra-frame coding mode }, the compressed data code stream at least includes a coding mode flag bit of a coding tree unit, and the flag bit indicates or derives whether a current actual value range of a coding mode of a coding block of the coding tree unit is the specific actual value range a in the following manner:

if the flag bit is equal to a predetermined value (e.g., 1), all the coding blocks in the coding tree unit adopt the intra-coding mode.

Examples or modifications 19

the data to be coded and decoded is an inter-coded frame (also called an inter-coded picture, i.e., a frame or picture allowing coding and decoding of a coded block in both an inter-coding mode and an intra-coding mode) InCoding tree luminance and chrominance bifurcation Brightness chroma coding tree structure with mode fused with coding mode of coding and decoding block and coding modeAnd the value range includes a part or all of the following set represented by enumeration:

{ intra-coding mode and luma and chroma have various bifurcated structures that are common or a single coding tree,

intra coding mode and luma and chroma have respective coding tree branching structures,

intra coding mode and luminance and chrominance have various branching structures of a coding tree that are common i.e. a single trunk portion but have respective branch portions,

intra-coding modes and luminance and chrominance have various branching structures of a coding tree which are common, i.e. a single one or more quarter-branch trunk portions, but have respective one or more bifurcations or three-branch portions,

inter coding modes and luma and chroma have various branching structures that are common or a single coding tree,

inter coding mode and luma and chroma have respective branched structures of the coding tree,

inter-coding mode and luma and chroma have various branching structures of a coding tree that have a common i.e. single trunk portion but separate branch portions,

inter-coding mode and various branching structures of coding trees having common, i.e. single, one or more, four-branched trunk portions for luminance and chrominance, but respective, one or more, two-branched or three-branched branch portions

}

A particular practical range a is one of the following:

the specific actual value range A is { the intra-frame coding mode and the various branch structures of the luma and chroma having a common, i.e., single, coding tree }

Or the specific actual value range A is { the intra-frame coding mode and the respective branch structures of the luminance and chrominance having respective coding trees }

Or the specific actual value range A is { various branch structures of a coding tree in intra-frame coding mode, where luminance and chrominance have a common, i.e., single, trunk portion, but have respective branch portions }

Or the specific actual value range A is { the intra-frame coding mode and the brightness and the chroma have various bifurcation structures of a coding tree which is a common one-time or multi-time four-bifurcation trunk part but respectively has one-time or multi-time two-bifurcation or three-bifurcation branch parts };

Implementation or modification example 20 implementation or modification example 19 the encoding method or apparatus or the decoding method or apparatus, where the compressed data code stream at least includes an encoding mode flag of an encoding tree unit, where the flag indicates or derives whether a current actual value range of an encoding mode of an encoding/decoding block of the encoding tree unit is the specific actual value range a in the following manner:

if the flag is equal to a predetermined value (e.g., a value of 1), the current actual value range of the coding mode and the luma coding tree structure of the coding tree unit is the specific actual value range a.

Examples of embodiment or modification 21

In the encoding method or apparatus or the decoding method or apparatus according to embodiment or variation 19, the specific actual value range a is { an intra-frame encoding mode, and the luminance and the chrominance have respective branch structures of the encoding trees, that is, at least a part of the luminance encoding tree and at least a part of the chrominance encoding tree are different }, the compressed data code stream at least includes a luminance-chrominance encoding tree structure and an encoding mode flag bit of an encoding tree unit, and the flag bit represents or derives whether the current actual value range of the encoding mode of the encoding and decoding block of the encoding tree unit is the specific actual value range a in the following manner:

if the flag is equal to a predetermined value (e.g., 1), all of the coding blocks in the coding tree unit adopt the intra-coding mode and at least a portion of the luma coding tree and the chroma coding tree of the coding tree unit are different.

Examples of embodiment or modification 22

the data to be coded and decoded is a multi-source motion vector predicted value of a coding and decoding block adopting an inter-frame prediction mode in an inter-frame coding image, and the value range of the multi-source motion vector predicted value comprises the following sets or a part of or the union of all the sets:

set 1 is the motion vector candidate set for the Skip sub-mode,

set 2 is the candidate set of motion vectors for the Merge sub-mode,

set 3 is the candidate set of motion vectors for the recent history sub-pattern,

set 4-a predetermined default motion vector candidate set,

the compressed data code stream at least contains information which can be used for directly or indirectly or directly and indirectly representing or deriving whether the current actual value range is a specific sub-range.

Examples of embodiment or variations 23

the elements of the data set form a grouping relationship according to the coding tree units to which the corresponding coding and decoding blocks belong, namely the elements of the corresponding coding and decoding blocks in the same group of C coding tree units, namely a coding tree unit group consisting of C coding tree units have the same group relationship, the elements not in the same coding tree unit group have the non-group relationship, the inter-correlation system is formed by at least the grouping relationship, wherein C is a predetermined constant between 1 and 50;

the data to be coded and decoded is an index (also commonly referred to as an address, an index, or the like) MergeIndex, which is an integer ranging from 0 to MaxMergeIndex-1, that is, MergeIndex, that is, 0. ltoreq. MergeIndex < MaxMergeIndex, where MaxMergeIndex is a predetermined constant, of a merge candidate list MergeCandidateList of motion vector predictors of coded and decoded blocks in an inter-coded picture using an inter prediction mode;

the current actual range of the MergeIndex of the codec blocks in a current coding tree unit group is an integer from 0 to ctuMaxMergeIndex-1, namely MergeIndex satisfies 0 ≦ MergeIndex < ctuMaxMexMergeIndex, wherein ctuMaxMergeIndex is an integer less than or equal to MaxMergeIndex, obviously, different coding tree unit groups are allowed to have different ctuMaxMergeIndex;

the compressed data code stream contains at least one or a combination of the following information:

1) information usable to directly or indirectly or directly and indirectly mixedly represent or derive a value of C,

2) information that can be used to directly or indirectly or directly indirectly mixedly represent or derive a value for maxmergesendex,

3) information that can be used to directly or indirectly or directly indirectly mixedly represent or derive the value of the corresponding ctuMaxMergeIndex for each group of coding tree elements.

Examples of implementation or modification 24

In the encoding method or apparatus or the decoding method or apparatus, the original data is image or video data, each frame of the image or video is divided into strips or strip groups or slice blocks or slice groups or brick block groups or coding tree units or coding tree unit groups or maximum coding units or maximum coding unit groups which are collectively called coding blocks, each coding block is composed of one or a plurality of coding tree units, each coding tree unit is further divided into coding and decoding blocks according to a quartering or non-branching structure of one or a plurality of recursions firstly and a trisecting or bifurcating or non-branching structure of one or a plurality of recursions secondly, each element of the data set corresponds to one coding and decoding block or a number thereof, obviously, each coding and decoding block belongs to one coding block;

the elements of the data set form a grouping relation according to the coding region blocks to which the corresponding coding and decoding blocks belong, namely the elements of the corresponding coding and decoding blocks in the same coding region block have the same group relation, the elements which are not in the same coding region block have the non-same group relation, and the mutual relation is formed by at least the grouping relation;

the data to be coded and decoded is the division size generated by dividing the coding tree unit according to the four-fork structure of one or more recursions or the equivalent logarithm base 2 of the coding tree unit, and the range of the value of the PartSizeInBit is from

Base 2 logarithm of minimum size for codec block MinCuSizeInBit

To

Base 2 logarithm of the size of a coding tree unit CtuSizeInBit

Or a subset or fraction thereof,

that is, part SizeInBit is satisfied

MinCuSizeInBit ≤ partSizeInBit ≤ CtuSizeInBit

Or a subset or portion thereof, where MinCuSizeInBit and ctussizeinbit are two predetermined constants in one image or video being encoded or decoded;

the current actual value range of the partSizeInBit of the coding/decoding block in a current coding block is from

The coding tree unit is divided into the minimum size of the generated coding and decoding block or the equivalent base 2 logarithm of the coding and decoding block according to a four-fork structure of one or more recursions

To

Base 2 logarithm of the size of a coding tree unit CtuSizeInBit

The integer between, i.e. partSizeInBit satisfies

crMinQtSizeInBit ≤ partSizeInBit ≤ CtuSizeInBit，

Where crMinQtSizeInBit is an integer equal to or greater than MinCuSizeInBit, it is clear that different coding blocks allow for different crminqtsizinbits.

Embodiment or variant example 25

In the encoding method or apparatus or the decoding method or apparatus according to embodiment or variation 24, the compressed data stream at least contains information that can be used to directly or indirectly or directly and indirectly mixedly represent or derive a value of the crMinQtSizeInBit;

the directly represented or derived information consists of one or more bit strings (bit strings) in a compressed data code stream; said indirectly represented or derived information is information derived from other coding parameters and/or codec variables and/or other syntax elements of the compressed data stream; the directly indirectly mixedly represented or derived information is information that is partially directly (i.e. composed of one or more bit strings in the compressed data stream) and partially indirectly (i.e. derived from other encoding parameters and/or codec variables and/or other syntax elements of the compressed data stream) mixedly represented or derived.

Example of embodiment or variation 26

In the encoding method or apparatus or the decoding method or apparatus according to embodiment or modification 25, the information directly indicating or deriving the value of crMinQtSizeInBit is the following information present in the coding block header informationCoarse italicSyntax elements of the representation:

minimum four-branch partition size of coding interval minus minimum size cr _ MinQtSizeInBit _ minus _ \ of coding and decoding block MinCuSizeInBit

Fromcr_MinQtSizeInBit_minus_MinCuSizeInBitThe value cr _ MinQtSizeInBit _ minus _ MinCuSizeInBit calculates the value crMinQtSizeInBit and the minimum size crMinQtSize of the codec block generated by the coding tree unit being divided in a quarterworking structure of one or more recursions as follows:

crMinQtSizeInBit ＝ MinCuSizeInBit ＋ cr_MinQtSizeInBit_minus_MinCuSizeInBit

crMinQtSize ═ (1 < < crMinQtSizeInBit), where < < is a 2-bit left shift operation.

Embodiment or modification 27 in the encoding method or apparatus or the decoding method or apparatus of embodiment or modification 26, at least under the condition that the part sizeinbit is greater than the MinCuSizeInBit, the compressed data code stream at least contains information that can be used for directly or indirectly or directly and indirectly mixedly representing or deriving whether to perform the next quartering structure division;

Examples of embodiment or variations 28

In the encoding method or apparatus or the decoding method or apparatus according to embodiment or modification 27, the information directly indicating or deriving whether to perform the next quartering structure division is as followsCoarse italicSyntax elements of the representation:

quad split flag qt _ split _ flag

qt_split_flagThe value qt _ split _ flag of (1) represents that the next quartering structure division is performed;qt_split_ flaga value qt _ split _ flag of '0' indicates that the quad structure division is terminated.

EXAMPLES OR VARIATIONS 29 (combinations of EXAMPLES or VARIATIONS 24 to 28)

Base 2 logarithm of minimum size for codec block MinCuSizeInBit

To

Base 2 logarithm of the size of a coding tree unit CtuSizeInBit

Or a subset or fraction thereof,

that is, part SizeInBit is satisfied

MinCuSizeInBit ≤ partSizeInBit ≤ CtuSizeInBit

To

Base 2 logarithm of the size of a coding tree unit CtuSizeInBit

The integer between, i.e. partSizeInBit satisfies

crMinQtSizeInBit ≤ partSizeInBit ≤ CtuSizeInBit，

Where crMinQtSizeInBit is an integer equal to or greater than MinCuSizeInBit, it is clear that different coding blocks allow for different crminqtsizeinbits;

the compressed data code stream at least contains information which can be used for directly or indirectly or directly and indirectly mixedly representing or deriving the value of the crMinQtSizeInBit;

the directly represented or derived information consists of one or more bit strings (bit strings) in a compressed data code stream; said indirectly represented or derived information is information derived from other coding parameters and/or codec variables and/or other syntax elements of the compressed data stream; the information directly and indirectly mixedly represented or derived is information that is partially directly (i.e., composed of one or more bit strings in the compressed data code stream) and partially indirectly (i.e., derived from other encoding parameters and/or codec variables and/or other syntax elements of the compressed data code stream) mixedly represented or derived;

the information directly representing or deriving the value of crMinQtSizeInBit is the following information present in the coding block header informationCoarse italicSyntax elements of the representation:

minimum four-way split partitioning between code binsSize minus minimum size cr _ MinQtSizeInBit _ minus _ @, of coding and decoding block MinCuSizeInBit

crMinQtSizeInBit ＝ MinCuSizeInBit ＋ cr_MinQtSizeInBit_minus_MinCuSizeInBit

crMinQtSize ═ (1 < < crMinQtSizeInBit), where < < is a 2-bit left shift operation;

at least under the condition that the part SizeInBit is larger than the MinCuSizeInBit, the compressed data code stream at least contains information which can be used for directly or indirectly or directly and indirectly mixed representing or deriving whether to carry out next four-fork structure division;

said information directly indicating or deriving whether to proceed with the next quartering structure division is as followsCoarse italicSyntax elements of the representation:

quad split flag qt _ split _ flag

Examples of embodiment or variations 30

In the encoding method or apparatus or the decoding method or apparatus, the original data is image or video data, each frame of the image or video is divided into strips or strip groups or tiles or tile groups or bricks or tile groups or coding tree units or coding tree unit groups or maximum coding units or maximum coding unit groups which are collectively called coding blocks, each coding block is composed of one or several coding tree units, each coding tree unit is further divided into coding and decoding blocks, each element of the data set corresponds to one coding and decoding block or the number thereof, obviously, each coding and decoding block belongs to one coding block;

data to be coded/decoded is an offset set _ mvd _ offset in units of a quarter pixel, a half pixel, or a whole pixel of a motion vector difference mvd in the set mode of a coded/decoded block having a motion vector mv, and the range of the offset set _ mvd _ offset is R = {2 =ⁱ: i=0, 1, 2, 3, 4, 5, 6, 7, 8, 9}；

The current actual value range of merge _ mvd _ offset of the coding/decoding block in a current coding block is R₀={2ⁱI =0, 1, 2,3, 4, 5, 6, 7} or R₁={2ⁱI =2, 3, 4, 5, 6, 7, 8, 9}, it is obvious that different coding blocks are allowed to have different current actual value ranges;

the following uses exist in the coding block header information of the compressed data code streamCoarse italicThe syntax elements of the representation represent or within which sub-range the values of the data corresponding to the elements of the coding block are restricted:

coded block merge motion vector difference offset flag cr _ merge _ mvd _ offset _ flag

cr_merge_mvd_offset_flagThe value cr _ merge _ mvd _ offset _ flag of "1" indicates that the current actual value range of the data corresponding to the elements of the coding block is R₁(ii) a A cr _ merge _ mvd _ offset _ flag value of '0' indicates that the current actual value range of the data corresponding to the elements of the coding block is R₀。

Examples or modifications 31

the data to be coded and decoded is interpolation weight when a predicted value of intra prediction is calculated, and the value range is R = { k is 0 ≦ k ≦ 31 };

the current actual value range of the interpolation weight of the coding and decoding block in a current coding block is R₀K =0 or R₁= R, obviously different coding blocks are allowed to have different current actual value ranges;

the following uses exist in the coding block header information of the compressed data code streamCoarse italicThe syntax elements represented represent or derive the values of the data corresponding to the elements of the coding block within which the current actual value range is limited:

intra prediction interpolation flag cr _ intra _ pred _ interpolation _ flag for coded block

cr_intra_pred_interpolation_flagA value of cr _ intra _ pred _ interpolation _ flag of '1' indicatesThe current actual value range of the data corresponding to the elements of the coding block is R₁Namely actually carrying out interpolation operation; a cr _ intra _ pred _ interpolation _ flag value of '0' indicates that the current actual value range of the data corresponding to the elements of the coding block is R₀I.e. no interpolation is actually performed.

Examples of implementation or modification 32

data to be coded and decoded is a chroma residual component transform coefficient, and the value range of the chroma residual component transform coefficient is R = {1, -1 };

the current actual value range of the chroma residual component transform coefficients of a coded block in a current coded block is R₀=1 or R₁=1, obviously, different coding blocks allow different current actual value ranges;

coded block chroma residual component transform coefficient flag cr _ chroma _ resi _ comp_trans_coeff_flag

cr_chroma_resi_comp_trans_coeff_flagThe value cr _ chroma _ resi _ comp _ trans _ coeff _ flag of '0' indicates that the current actual value range of the data corresponding to the elements of the coding block is R₀Namely the chroma residual component transformation coefficient is 1; a cr _ intra _ pred _ interpolation _ flag value of '1' indicates that the current actual value range of the data corresponding to the elements of the coding block is R₁I.e., the chrominance residual component transform coefficient is-1.

Examples or modifications 33

the data to be coded and decoded is a component conversion mode after a predetermined chroma residual inverse component conversion is carried out, and the value range of the component conversion mode is R = { the Cb-Cr is not converted after the chroma residual inverse component conversion, and the Cb-Cr is converted after the chroma residual inverse component conversion };

the current actual value range of the component swapping pattern of a codec block in a current coding block is R₀= { Cb-Cr is not converted after chroma residual inverse component conversion } or R₁= Cb-Cr transform after inverse component transform of the chroma residual, it is clear that different coding blocks allow different current actual fetchesA range of values;

flag cr _ chroma _ resi _ post-inv-comp-tran \/after inverse transform of chroma residual of coded block exchange_flag

cr_chroma_resi_post-inv-comp-tran_exchange_flagThe value cr _ chroma _ resi _ post-inv-comp-tran _ exchange _ flag of '0' indicates that the current actual value range of the data corresponding to the elements of the coding block is R₀Namely, Cb-Cr is not exchanged after the chroma residual inverse component is transformed; a cr _ chroma _ resi _ post-inv-comp-tran _ exchange _ flag value of '1' indicates that the current actual value range of the data corresponding to the elements of the coding block is R₁Namely Cb-Cr conversion after the chroma residual reverse component conversion.

Description of the drawings fig. 1 is a schematic diagram of an encoding method or apparatus of the present invention.

Fig. 2 is a schematic diagram of a decoding method or apparatus of the present invention.

Claims

1. An encoding method for compressing a data set and its data, characterized by comprising at least the following steps:

2. An encoding device for compressing a data set and its data, characterized by comprising at least the following modules:

3. A decoding method for compressing a data set and its data, comprising at least the steps of:

4. A decoding device for compressing data sets and data thereof, characterized by comprising at least the following modules:

5. The encoding method according to claim 1 or the encoding apparatus according to claim 2 or the decoding method according to claim 3 or the decoding apparatus according to claim 4, characterized in that:

the raw data is image or video data;

each frame of the image or video is divided into a coding block, which is a region of the image, including the following cases: the method comprises the following steps that (1) a sub-image of an image, a stripe slice, a stripe group consisting of a plurality of stripes, a tile, a tile group consisting of a plurality of tiles, a brick crack, a brick group consisting of a plurality of bricks, a maximum coding unit (LCU), a maximum coding unit group consisting of a plurality of maximum coding units, a Coding Tree Unit (CTU) and a coding tree unit group consisting of a plurality of coding tree units are formed;

the coding block is further divided into coding and decoding blocks, the coding and decoding blocks are a coding area or a decoding area of the image, and the following situations are included: a sub-picture of a picture, a macroblock, a largest coding unit LCU, a coding tree unit CTU, a coding unit CU, a sub-region of a CU, a sub-coding unit SubCU, a prediction unit PU, a sub-region of a PU, a sub-prediction unit SubPU, a transform unit TU, a sub-region of a TU, a sub-transform unit SubTU.

6. The encoding method according to claim 1 or the encoding apparatus according to claim 2 or the decoding method according to claim 3 or the decoding apparatus according to claim 4, characterized in that:

the original data is image or video data, each frame of the image or video is divided into strips or strip groups or tiles or tile groups or bricks or tile groups or coding tree units or coding tree unit groups or maximum coding units or maximum coding unit groups which are collectively called coding blocks, each coding block is composed of one or a plurality of coding tree units, each coding tree unit is further divided into coding and decoding blocks according to a quartering or non-branching structure of one or a plurality of recursions firstly and then a trisectional or bifurcating or non-branching structure of one or a plurality of recursions firstly, each element of the data set corresponds to one coding and decoding block or a number thereof, obviously, each coding and decoding block belongs to one coding block;

Base 2 logarithm of minimum size for codec block MinCuSizeInBit

To

Base 2 logarithm of the size of a coding tree unit CtuSizeInBit

Or a subset or fraction thereof,

that is, part SizeInBit is satisfied

MinCuSizeInBit ≤ partSizeInBit ≤ CtuSizeInBit

To

Base 2 logarithm of the size of a coding tree unit CtuSizeInBit

The integer between, i.e. partSizeInBit satisfies

crMinQtSizeInBit ≤ partSizeInBit ≤ CtuSizeInBit，

crMinQtSizeInBit ＝ MinCuSizeInBit ＋ cr_MinQtSizeInBit_minus_MinCuSizeInBit

quad split flag qt _ split _ flag

7. The encoding method according to claim 1 or the encoding apparatus according to claim 2 or the decoding method according to claim 3 or the decoding apparatus according to claim 4, characterized in that:

the original data is image or video data, each frame of image of the image or video is divided into strips or strip groups or tiles or tile groups or bricks or tile groups or coding tree units or coding tree unit groups or maximum coding units or maximum coding unit groups which are collectively called coding blocks, each coding block is composed of one or a plurality of coding tree units, each coding tree unit is further divided into coding and decoding blocks, each element of the data set corresponds to one coding and decoding block or the number thereof, obviously, each coding and decoding block belongs to one coding block;

cr_merge_mvd_offset_flagThe value cr _ merge _ mvd _ offset _ flag of "1" indicates the current actual value of the data corresponding to the element of the coding blockRange is R₁(ii) a A cr _ merge _ mvd _ offset _ flag value of '0' indicates that the current actual value range of the data corresponding to the elements of the coding block is R₀。

8. The encoding method according to claim 1 or the encoding apparatus according to claim 2 or the decoding method according to claim 3 or the decoding apparatus according to claim 4, characterized in that:

coded block intra prediction interpolation flag cr _ intra _ pred_interpolation_flag

cr_intra_pred_interpolation_flagThe value cr _ intra _ pred _ interpolation _ flag of '1' indicates that the current actual value range of the data corresponding to the elements of the coding block is R₁Namely actually carrying out interpolation operation; a cr _ intra _ pred _ interpolation _ flag value of '0' indicates that the current actual value range of the data corresponding to the elements of the coding block is R₀I.e. no interpolation is actually performed.

9. The encoding method according to claim 1 or the encoding apparatus according to claim 2 or the decoding method according to claim 3 or the decoding apparatus according to claim 4, characterized in that:

coding block of compressed data code streamThe following uses exist in the header informationCoarse italicThe syntax elements represented represent or derive the values of the data corresponding to the elements of the coding block within which the current actual value range is limited:

coded block chroma residual component transform coefficient flag cr _ chroma _ resi _ comp _ trans _ coeff _ flag

10. The encoding method according to claim 1 or the encoding apparatus according to claim 2 or the decoding method according to claim 3 or the decoding apparatus according to claim 4, characterized in that:

the current actual value range of the component swapping pattern of a codec block in a current coding block is R₀= { Cb-Cr is not converted after chroma residual inverse component conversion } or R₁= { Cb-Cr transposition after inverse component transformation of chroma residual }, it is obvious that different coding blocks are allowed to have different current actual value ranges;