CN104811731A

CN104811731A - Multilayer sub-block matching image compression method

Info

Publication number: CN104811731A
Application number: CN201510000780.4A
Authority: CN
Inventors: 林涛
Original assignee: Shanghai Tianhe Electronic Information Co Ltd
Current assignee: Shanghai Tianhe Electronic Information Co Ltd
Priority date: 2014-01-03
Filing date: 2015-01-04
Publication date: 2015-07-29

Abstract

The invention provides an image provides an image compression method and an image compression device. When matching coding is carried out on a coding unit, the coding unit is segmented into a plurality of different levels of divisions, each level comprises one or more divisions, the sub-blocks are taken as units, a matching sub-block is searched in a reconstructed reference pixel sample value set according to predetermined evaluation criteria, and a corresponding displacement vector is acquired. Optimum coding of multilayer sub-block matching is carried out on the coding unit by selecting the level whose overall coding performance is the best, the divisions of the level and the corresponding group of displacement vectors after all levels of predetermined searching are completed.

Description

Multi-level sub-block matching image compression method

Technical field

The present invention relates to a kind of compression of digital video coding and the coding of decode system, particularly computer screen image and video and the method for decoding.

Background technology

The natural form of the digital video signal of image is the sequence of image.The rectangular area that one two field picture is normally made up of some pixels, and digital video signal is exactly the sequence of video images be made up of tens frames to thousands of two field picture, sometimes also referred to as video sequence or sequence.Digital encoding video signal is encoded to image one by one exactly.

Up-to-date international video compression standards HEVC(High Efficiency Video Coding) in, when one two field picture is encoded, one two field picture is divided into the subgraph of some pieces of MxM pixels, be called " coding unit (Coding Unit is called for short CU) ", take CU as basic coding unit, subgraph one piece a piece is encoded.The size of conventional M is 8,16,32,64.Therefore, carrying out encoding to a sequence of video images is exactly encode successively to each coding unit of each two field picture and CU.At any one time, the CU encoded is called present encoding CU.Equally, be also that each coding unit and CU are decoded successively during decoding, finally reconstruct whole sequence of video images.At any one time, the CU decoded is called current decoding CU.Present encoding CU or current decoding CU is referred to as current C U.

For adapting to the different of each several part picture material and character in a two field picture, carry out most efficient coding targetedly, in a two field picture, the size of each CU can be different, has plenty of 8x8, has plenty of 64x64, etc.In order to enable the CU of different size seamless spliced, one two field picture is always first divided into size identical " maximum coding unit (Largest Coding Unit is called for short LCU) " with NxN pixel, then each LCU again Further Division become the phasing multiple not of uniform size of tree with CU.Therefore, LCU is also referred to as " code tree unit (Coding Tree Unit is called for short CTU) ".Such as, a two field picture is first divided into the LCU(N=64 of the identical 64x64 pixel of size).Wherein certain LCU is made up of the CU of CU and 4 16x16 pixel of 3 32x32 pixels, and such 7 become the CU of tree to form a CTU.And another LCU is made up of the CU of CU and 20 8x8 pixel of the CU of 2 32x32 pixels, 3 16x16 pixels.Such 25 become the CU of tree to form another CTU.Encoding to a two field picture, is exactly encode to the CU one by one in CTU one by one successively.

A colour element has 3 component (component) compositions usually.The GBR color format that the most frequently used two kinds of pixel color forms (pixel color format) are made up of green component, blue component, red component and the YUV color format be made up of brightness (luma) component and two colourity (chroma) components, be commonly referred to as that the color format of YUV is actual comprises multicolour form, as YCbCr color format.Therefore, when a CU is encoded, a CU can be divided into 3 component planes (G plane, B plane, R plane or Y plane, U plane, V plane), 3 component planes are encoded respectively; Also 3 of a pixel component binding combination can be become 3 tuples, the CU entirety be made up of these 3 tuples is encoded.The arrangement mode of front a kind of pixel and component thereof is called the planar format (planar format) of image (and CU), and then the arrangement mode of a kind of pixel and component thereof is called the stacked package form (packed format) of image (and CU).The GBR color format of pixel and YUV color format are all 3 component representation formats of pixel.

Except 3 component representation formats of pixel, the representation formats of the prior art that the another kind of pixel is commonly used is palette index representation formats.In palette index representation formats, the numerical value of a pixel also can show with the index of palette.Store in palette space and need by the numerical value of 3 components of the pixel showed or numerical approximation, the address of palette is called as the index of the pixel stored in this address.An index can show the one-component of pixel, and an index also can show 3 components of pixel.Palette can be one, also can be multiple.In the situation of multiple palette, in fact complete index is numbered by palette and index two parts of palette of this numbering form.The index representation formats of pixel is exactly that index of reference is to show this pixel.The index representation formats of pixel is also referred to as indexed color (indexed color) or imitative color (pseudo color) representation formats of pixel in the prior art, or is usually directly called index pixel (indexed pixel) or imitative pixel (pseudo pixel) or pixel index or index.Index is also sometimes referred to as index.Its index representation formats of pixel is showed also referred to as indexation or indexation.

The pixel performance form of other conventional prior art comprises CMYK representation formats and expressing gradation form.

YUV color format can be subdivided into several subformat according to whether carrying out down-sampling to chromatic component again: the YUV4:4:4 pixel color form that 1 pixel is made up of 1 Y-component, 1 U component, 1 V component; The YUV4:2:2 pixel color form that adjacent 2 pixels in left and right are made up of 2 Y-component, 1 U component, 1 V component; The YUV4:2:0 pixel color form that 4 pixels of the neighbouring 2x2 of the pressing spatial arrangements in left and right are made up of 4 Y-component, 1 U component, 1 V component.One-component generally represents by the numeral of 18 ~ 16 bit.YUV4:2:2 pixel color form and YUV4:2:0 pixel color form are all obtain the down-sampling of YUV4:4:4 pixel color form execution chromatic component.Pixel component is also referred to as a pixel sample value (pixel sample) or be called a sample value (sample) simply.

Most basic element when coding or decoding can be a pixel, and also can be a pixel component, also can be a pixel index (i.e. index pixel).Be referred to as a pixel sample value (sample) as a pixel of the most basic element of coding or decoding or a pixel component or an index pixel, be sometimes also commonly referred to as a pixel value, or be called a sample value simply.

In the present invention and patent application of the present invention, " pixel sample value ", " pixel value ", " sample value ", " index pixel ", " pixel index " are synonyms, based on context, clearly can represent " pixel " or represent " pixel component " or represent " index pixel " or represent that three's is arbitrary simultaneously.If can not be clear and definite from context, be exactly so represent that three's is arbitrary simultaneously.

At existing prediction code encoding/decoding mode, Block-matching code encoding/decoding mode, in transform coding and decoding mode, in order to improve code efficiency, a CU can be divided into upper and lower 2 or 2, left and right or 4 equal predicting unit (Prediction Unit is called for short PU) up and down, also a CU can be divided into the foursquare converter unit (Transform Unit is called for short TU) of 1/4th or 1/16th size.

Encoding block or decoding block refer to a region of in a two field picture, it being implemented to coding or decoding, comprise following at least one: maximum coding unit LCU, code tree unit CTU, the subregion of coding unit CU, CU, predicting unit PU, converter unit TU.

Along with taking remote desktop as the cloud computing of new generation of classic manifestations and the development of messaging model and platform and universal, interconnected between multiple stage computer, between other digital devices such as main frame and intelligent television, smart mobile phone, panel computer and between various all kinds of digital device has become a reality and has increasingly become a kind of main trend.This makes server end (cloud) become current active demand to the real time screen transmission of user side.Owing to needing the screen video data volume of transmission huge, the data compression of high-effect high-quality must be carried out to computer screen image.

Making full use of the feature of computer screen image, computer screen image is carried out to the compression of Ultra-High Efficiency, is also a main target of up-to-date international video compression standards HEVC and some other international standards, domestic standard, the industry standard formulated.

A distinguishing feature of computer screen image usually has much similar even identical pixel patterns (pixel pattern) in same two field picture.Such as, the Chinese of normal appearance in computer screen image or foreign language word, be all be made up of a few basic strokes, can find much similar or identical stroke in same two field picture.Menu common in computer screen image, icon etc., also have much similar or identical pattern.Infra-frame prediction (intra prediction) mode adopted in existing image and video compression technology, only with reference to adjacent pixel sample value, cannot utilize the similitude in a two field picture or homogeny to improve compression efficiency.Intra frame motion compensation of the prior art (intra motion compensation) mode copies (intra block copy) mode also referred to as intra block, with several fixed size (8x8,16x16,32x32,64x64 pixel) block carry out intra block coupling (intra block matching) coding, the meticulousr coupling with various different size and shape can not be reached.And another kind of String matching of the prior art (string matching) mode, although the meticulous coupling of various different size and shape effectively can be found, there is the problems such as complexity, amount of calculation, memory read/write bandwidth be all larger.

Therefore, new coding tools must be sought, fully can excavate and utilize the similar or identical pattern existed in computer screen image, to increase substantially compression effectiveness, again complexity, amount of calculation, memory read/write bandwidth etc. all being controlled in lower scope.

Summary of the invention

Technical characteristics of the present invention is when carrying out coupling coding to a present encoding block, described code block segmentation is become the division of some different levels, each level has one or more division (usually to have 1 to 8 division, usually 16 divisions are no more than at most), each division is made up of some sub-blocks, and the division in same level has same sub-block number usually.Particularly, the 0th level only has a division usually, and this division only has this 1 sub-block of present encoding block itself; The division of the 1st level has at least 2 sub-blocks; The sub-block number of the division of J level is greater than the sub-block number of the division of J-1 level usually.To each division of each level, with each sub-block of described division, being called that by matched sub-block or current sub-block be unit, in the history pixel sample value region (also claiming reconstructed reference pixel sample value district) completing coding and reconstruct (comprising Perfect Reconstruction and partial reconfiguration in various degree), searching for corresponding matched sub-block and reference sub-block respectively also referred to as predicting sub-block.The described pixel sample value of reconstructed reference district is by being positioned at same two field picture (namely have same picture order count picture order count and be called for short poc) with described present encoding block (from the angle of decoder also referred to as decoding block) or being positioned at reconstruct (comprising Perfect Reconstruction and/or partial reconfiguration in various degree) the pixel sample value of other encoding blocks of same band (slice) and the sample value of reconstructed pixel reconstructing sub-block of described present encoding block is formed.The described pixel sample value of reconstructed reference district comprises the sample value of reconstructed pixel of other encoding blocks described and has describedly reconstructed the part or all of of the sample value of reconstructed pixel of sub-block.

Fig. 1 (a) is an example of the division of 4 levels: the 0th level has 1 partition and corresponding 1 displacement vector (Displacement Vector is called for short DV), 1st level has two 2 partitions, each division has 2 corresponding displacement vectors, 2nd level has 4 partitions and 4 corresponding displacement vectors, and the 3rd level has 16 partitions and 16 corresponding displacement vectors..

Fig. 1 (b) is an example of the division of 4 levels: the 0th level has 1 partition and 1 corresponding displacement vector, 1st level has two 2 partitions, each division has 2 corresponding displacement vectors, 2nd level has 4 partitions and 4 corresponding displacement vectors, 3rd level has two 8 partitions, and each division has 8 corresponding displacement vectors.

Fig. 1 (c) is an example of the division of 4 levels: the 0th level has 1 partition and 1 corresponding displacement vector, 1st level has six 2 partitions, each division has 2 corresponding displacement vectors, 2nd level has 4 partitions and 4 corresponding displacement vectors, 3rd level has six 8 or 9 or 16 partitions, and each division has 8 or 9 or 16 corresponding displacement vectors.

Fig. 1 (d) is an example of the division of 2 levels: the 0th level has 1 partition and 1 corresponding displacement vector, and the 1st level has two 8 partitions, and each division has 8 corresponding displacement vectors.

In the present invention and patent application of the present invention, " matched sub-block ", " reference sub-block ", " prediction sub-block " are synonyms.

In the present invention and patent application of the present invention, " by matched sub-block " and " current sub-block " is synonym.

Multi-level sub-block coupling coding has the few thus complexity of displacement vector number and the lower characteristic sum matched shape of memory read/write bandwidth requirement concurrently and varies thus the high feature of code efficiency.

The multi-level division represented in Fig. 1 can be the division of stacked package form, also can be the division of the one-component (sample value) of planar format.Thus method of the present invention, had both gone for the coding to the encoding block of stacked package form or the pixel of decoding block or decoding, also went for the coding to the pixel sample value of the encoding block of planar format or a plane of decoding block or decoding.

In coding method of the present invention, the most basic peculiar technical characteristic is exactly, pre-determine multiple division, comprise the division of some levels, each level level number represents, if having two or more division in a level, then each division also needs to represent by a mode number, level number and mode number form partition mode, and the division of same level has the division of same sub-block number and different levels usually to have different sub-block numbers to be beneficial to carry out sub-block coupling by different level to encode usually; When encoding to present encoding block, respectively divide at each level and carry out sub-block coupling coding, namely each sub-block in each level of described present encoding block respectively being divided, be called by matched sub-block and current sub-block, search in the history pixel sample value region (also claiming reconstructed reference pixel sample value district) completing coding, find and described each matched sub-block of respectively being matched by matched sub-block and reference sub-block; Each matched sub-block parameter is called that displacement vector represents that namely matching distance mates relative position; The original position of matched sub-block can be the position of pixel sample value arbitrarily among reconstructed reference pixel sample value district, and have nothing to do with the size of sub-block, namely displacement vector is in units of pixel sample value, and its scope is the part or all of of view picture present image; After the search completing predetermined level and division, that level of binary encoding best performance and that group displacement vector of division and correspondence, exactly described present encoding block is carried out to the optimum code selection of multi-level sub-block coupling coding, characterize with the partition mode be made up of level number, mode number (if the level selected has more than one division) and these two parameters of displacement vector group, and write compressed video stream.The described pixel sample value of reconstructed reference district is formed by with the sample value of reconstructed pixel reconstructing sub-block that described present encoding block has in the sample value of reconstructed pixel of other encoding blocks in the image of same poc or band and described present encoding block.The described pixel sample value of reconstructed reference district comprises the sample value of reconstructed pixel of other encoding blocks described and has describedly reconstructed the part or all of of the sample value of reconstructed pixel of sub-block.

In coding/decoding method of the present invention, the most basic peculiar technical characteristic is exactly when the compressed bit stream decoding data to current decoding block, resolves the partition mode and displacement vector group that obtain comprising level number and mode number (if the level selected has more than one division) from bit stream data.Then, the position of each reference sub-block among reconstructed reference pixel sample value district is gone out from the position of current decoding block and each displacement Vector operation.Then, each reference sub-block is copied among reconstructed reference pixel sample value district, each reference sub-block is moved and pastes the position of each current sub-block in current decoding, namely the direct or indirect assignment of numerical value of the pixel sample value of described each reference sub-block is given described each current sub-block, restore the pixel sample value of whole current decoding block.The same with situation during coding, the original position of reference sub-block can be the position of pixel sample value arbitrarily among reconstructed reference pixel sample value district, have nothing to do with the size of sub-block, namely displacement vector is in units of pixel sample value, and its scope is the part or all of of view picture present image.The described pixel sample value of reconstructed reference district is formed by with the sample value of reconstructed pixel reconstructing sub-block that described current decoding block has in the sample value of reconstructed pixel of other decoding blocks in the image of same poc or band and described current decoding block.The described pixel sample value of reconstructed reference district comprises the sample value of reconstructed pixel of other decoding blocks described and has describedly reconstructed the part or all of of the sample value of reconstructed pixel of sub-block.

By some specific instantiations, technical characteristic of the present invention is described above.Those skilled in the art can understand other advantages of the present invention and effect easily by content disclosed in the present specification.The present invention can also be implemented or be applied by embodiments different in addition, and the every details in this specification also can based on different viewpoints and application, carries out various modification or change not deviating under spirit of the present invention.Vocabulary used in the present invention, also can replace with having other vocabulary of same meaning, as " level " also can replace by " type " or " kind ".

The schematic flow sheet of coding method of the present invention as shown in Figure 2.Coding method of the present invention includes but not limited to following steps:

1) carry out multi-level sub-block coupling coding to the pixel sample value of encoding block (including but not limited to CU) of an input, namely the partition mode be made up of level number and mode number (if the level of described level number has more than to divide) producing (1) optimum mates relative position with the displacement vector of (2) corresponding one group of optimum; Comprise: travel through predetermined each level and each division, in one among reconstructed reference pixel sample value district predetermined hunting zone, according to predetermined assessment level, search obtains each displacement vector of an optimum partition mode (comprising optimum level and its optimal dividing) and each current sub-block of correspondence and the optimum of each reference sub-block; Described each displacement vector is the difference of the position coordinates searching for each current sub-block described in the position coordinates of described each reference sub-block optimum among the pixel sample value of the reconstructed reference district that determines and described encoding block; The unit of institute's displacement vector is the min coordinates unit of pixel sample value (Integer Pel sample value or 1/2nd, 1/4th, 1/8th pixel sample values); The reference pixel that institute's displacement vector points to is positioned at same image or same band, and its position range is the part or all of of the reconstruction region of view picture present image; The output of this step is the partition mode of described optimum, and the level comprising optimum level and division is number with mode number (if described optimum level has more than to divide), described each displacement vector (also referred to as displacement vector group) and mate residual error; Described coupling residual error is the difference of the numerical value of the pixel sample value of each reference sub-block described in the numerical value of the pixel sample value of each current sub-block described in described encoding block and the described pixel sample value of reconstructed reference district;

2) remaining conventional coding and reconstruction step, as Point matching, palette coding, infra-frame prediction, inter prediction, conversion, quantification, inverse transformation, inverse quantization, entropy code, block-eliminating effect filtering, sample value adaptive equalization (Sample Adaptive Offset) in rectangle coupling, frame in String matching, frame in bar coupling, frame in microlith coupling, frame in intra block coupling, frame; The input of this step is above-mentioned steps 1) output and input pixel sample value; The output of this step is reconstructed pixel (comprising Perfect Reconstruction pixel and partial reconfiguration pixel in various degree) and the compressed bit stream containing partition mode (including but not limited to level information and optional mode number information) and displacement vector group and other coding results; Described reconstructed pixel puts into reconstructed reference pixel sample value working area (namely reconstructed reference pixel sample value district), as the reference pixel required for follow-up parallel multi-level sub-block coupling coding step, all the other conventional codings and reconstruction step; Described compressed bit stream is also the last output of this coding method.

Point matching in rectangle coupling, frame in String matching, frame in bar coupling, frame in microlith coupling, frame in multi-level sub-block coupling, intra block coupling, frame above, from the angle of decoding, be commonly referred to that multi-level sub-block copies, intra block copies, microlith copies in frame, bar copies in frame, string copies in frame, rectangle copies in frame, point copies in frame.

The feature of coding/decoding method of the present invention is, comprises

Resolve video code flow, obtain partition mode and the displacement vector group of decoding block, and other supplemental characteristics of described decoding block; Wherein, described partition mode includes but not limited to the level information of the division that multi-level sub-block copies and optional mode number information, represent the one in predetermined multiple division, be used to indicate and described decoding block is divided into one or more decoding sub-block, institute's displacement vector group is used to indicate the position of the reference sub-block of homographic solution numeral block;

According to described partition mode and displacement vector group, in the pixel sample value of the reconstructed reference district point to institute's displacement vector corresponding for described decoding sub-block, give the prediction sub-block of described decoding sub-block with the pixel of the reconstructed reference sample value assignment in described decoding sub-block same shape overlay area; Wherein, reference pixel sample value in the described pixel sample value of reconstructed reference district is set to be arranged in other decoding blocks of identical image or same strap or the decoding and reconstituting value of other sub-block pixels of decoding with described decoding block, or described decoding and reconstituting value process after filtering after value; Wherein, the reference pixel that institute's displacement vector points to is positioned at same image or same band, and its position range is the part or all of of the reconstruction region of view picture present image;

According to the prediction sub-block of described decoding sub-block in described decoding block and other supplemental characteristics of described decoding block, construct the reconstructed pixel sample value of described decoding block.

The schematic flow sheet of coding/decoding method of the present invention as shown in Figure 3.Coding/decoding method of the present invention includes but not limited to following steps:

1) compressed bit stream containing the partition mode and displacement vector group and other coding results that include level information and optional mode number information is resolved, export 1) partition mode of the level information that includes that obtains and optional mode number information and displacement vector group, 2) all the other decoding parametric obtained and data; The described partition mode including level information and optional mode number information represents the one in predetermined multiple division;

2) the multi-level sub-block using the partition mode including level information and optional mode number information and displacement vector group to carry out a decoding block copies decoding, comprise: respectively calculated by the position of matched sub-block (also known as current sub-block) in the described decoding block that partition mode is specified according to each displacement vector of described mobile vector group and determine the position of each reference sub-block, from the position of the described each reference sub-block reconstructed reference pixel sample value working area (namely reconstructed reference pixel sample value district), copy the sample value of described each reference sub-block, and described each reference sub-block is moved and is pasted the position of each current sub-block described in described decoding block, namely the direct or indirect assignment of numerical value of the pixel sample value of described each reference sub-block is given described each current sub-block, restore the pixel sample value of described decoding block, the unit of described each displacement vector is the min coordinates unit of pixel sample value (Integer Pel sample value or 1/2nd, 1/4th, 1/8th pixel sample values), the reference pixel of described each displacement vectors directed is positioned at same image or same band, and its position range is the part or all of of the reconstruction region of view picture present image,

3) remaining conventional decoding and reconstruction step, as intra block copy, in frame microlith copy, in frame bar copy, in frame string copy, in frame rectangle copy, in frame point copy, palette decoding, infra-frame prediction, inter prediction, inverse transformation, inverse quantization, entropy decoding, block-eliminating effect filtering, sample value adaptive equalization (Sample Adaptive Offset); The input of this step is above-mentioned steps 2) output and above-mentioned steps 1) output 2) namely all the other resolve the data that obtain; The output of this step is reconstructed pixel (comprising Perfect Reconstruction pixel and partial reconfiguration pixel in various degree); Described reconstructed pixel puts into reconstructed reference pixel sample value working area (namely reconstructed reference pixel sample value district), copies the reference pixel required for decoding step, all the other conventional decodings and reconstruction step as subsequent multi-layer time sub-block; Described Perfect Reconstruction pixel is also the last output of this coding/decoding method.

Above provided diagram only illustrates basic conception of the present invention in a schematic way, only the assembly directly relevant with the present invention is shown but not component count, shape and size when implementing according to reality is drawn in diagram, it is actual when implementing, and the kenel of each assembly, quantity and ratio can be a kind of change arbitrarily, and its assembly layout kenel also may be more complicated.

Below more implementation detail of the present invention and variant.

the embodiment 1 of predetermined level and division

Described predetermined multiple division includes but not limited to level and the division with following 3 features simultaneously: 1) have 2 or 3 levels, 2) level number be 0 level only have a size to be equal to the sub-block of described present encoding block or decoding block, 3), in all the other levels, a level is had to have 4 partitions of non-4 deciles up and down at least.

the embodiment 2 of predetermined level and division

Described predetermined multiple division includes but not limited to level and the division with following 2 features simultaneously: 1) level number be 0 level only have a size to be equal to the sub-block of described present encoding block or decoding block, 2), in all the other levels, a level is had to have division more than 4 sub-blocks.

the embodiment 3 of predetermined level and division

Described predetermined multiple division includes but not limited to level and the division with following 4 features simultaneously: 1) level number be 0 level only have a size to be equal to the sub-block of described present encoding block or decoding block, 2) level is had to have 2 partitions, 3) level is had to have 4 partitions, 4) there is a level to have division more than 4 sub-blocks.

the embodiment 4 of predetermined level and division

Described predetermined multiple division includes but not limited to level and the division with following 2 features simultaneously: 1) level number be 0 level only have a size to be equal to the sub-block of described present encoding block or decoding block, 2) have a level to have 8 partitions.

the embodiment 5 of predetermined level and division

Described predetermined multiple division includes but not limited to level and the division with following 2 features simultaneously: 1) level number be 0 level only have a size to be equal to the sub-block of described present encoding block or decoding block, 2) have a level to have 16 partitions.

the embodiment 6 of predetermined level and division

Described predetermined multiple division includes but not limited to the part or all of of following level and division:

Level number is level i.e. the 0th level of 0, has 1 unique partition, is equal to described present encoding block or decoding block, non-mode number;

Level number is level i.e. the 1st level of 1, there are two 2 partitions, the mode number of 2 partitions is 1, by described present encoding block or decoding block, i.e. unique division of described 0th level, left and right 2 decile obtains, the mode number of another 2 partition is 2, by described present encoding block or decoding block, i.e. unique division of described 0th level, upper and lower 2 deciles obtain;

Level number is level i.e. the 2nd level of 2, has 4 unique partitions, by described present encoding block or decoding block, i.e. and unique division of described 0th level, 4 deciles obtain up and down, non-mode number;

Level number is level i.e. the 3rd level of 3, has 16 unique partitions, and by described present encoding block or decoding block, i.e. unique division of described 0th level, decile upper and lower 8 deciles in left and right 2 obtain, non-mode number.

the embodiment 7A of predetermined level and division

As shown in Fig. 1 (a), described predetermined multiple division includes but not limited to the part or all of of following level and division:

Level number is level i.e. the 1st level of 1, there are two 2 partitions, the mode number of 2 partitions is 1, unique division left and right 2 deciles of described 0th level are obtained, its 2 sub-blocks are called the 1st sub-block and the 2nd sub-block of the division one of the 1st level, the mode number of another 2 partition is 2, is obtained by upper and lower 2 deciles of unique division of described 0th level, and its 2 sub-blocks are called the 1st sub-block and the 2nd sub-block of the division two of the 1st level;

Level number is level i.e. the 2nd level of 2, there are 4 unique partitions, by unique division of described 0th level up and down 4 deciles obtain, non-mode number, its 4 sub-blocks are called the 1st sub-block, the 2nd sub-block, the 3rd sub-block, the 4th sub-block of unique division of the 2nd level;

Level number is level i.e. the 3rd level of 3, has 16 unique partitions, is obtained by upper and lower 2 deciles of unique division left and right 8 deciles of described 0th level, non-mode number.

Above-described embodiment 6 the 0th, 1, the division of 2 levels and embodiment 7A the 0th, 1, the division of 2 levels is identical, and the 3rd level of embodiment 6 is identical from the sub-block number of unique division of the 3rd level of embodiment 7A but sub-block shape is different.

the embodiment 7B of predetermined level and division: with an embodiment of embodiment 7A equivalent

Level number is level i.e. the 2nd level of 2, there are 4 unique partitions, by the 1st sub-block of the division one of described 1st level and the 2nd sub-block respectively upper and lower 2 deciles obtain, non-mode number, its 4 sub-blocks are called the 1st sub-block, the 2nd sub-block, the 3rd sub-block, the 4th sub-block of unique division of the 2nd level;

Level number is level i.e. the 3rd level of 3, has 16 unique partitions, by the 1st sub-block of unique division of described 2nd level, the 2nd sub-block, the 3rd sub-block, the 4th sub-block respectively left and right 4 deciles obtain, non-mode number.

the embodiment 7C of predetermined level and division: with an embodiment of embodiment 7A equivalent

Level number is level i.e. the 2nd level of 2, there are 4 unique partitions, by the 1st sub-block of the division two of described 1st level and the 2nd sub-block respectively left and right 2 deciles obtain, non-mode number, its 4 sub-blocks are called the 1st sub-block, the 2nd sub-block, the 3rd sub-block, the 4th sub-block of unique division of the 2nd level;

the embodiment 7D of predetermined level and division: with an embodiment of embodiment 7A equivalent

Level number is level i.e. the 3rd level of 3, has 16 unique partitions, by the 1st sub-block of the division two of described 1st level and the 2nd sub-block respectively left and right 8 deciles obtain, non-mode number.

the embodiment 7E of predetermined level and division: with an embodiment of embodiment 7A equivalent

the embodiment 7F of predetermined level and division: with an embodiment of embodiment 7A equivalent

Level number is level i.e. the 1st level of 1, there are two 2 partitions, the mode number of 2 partitions is 1, unique division left and right 2 deciles of described 0th level are obtained, its 2 sub-blocks are called the 1st sub-block (left sub-block) and the 2nd sub-block (right sub-block) of the division one of the 1st level, the mode number of another 2 partition is 2, obtained by upper and lower 2 deciles of unique division of described 0th level, its 2 sub-blocks are called the 1st sub-block (upper sub-block) and the 2nd sub-block (lower sub-block) of the division two of the 1st level;

Level number is level i.e. the 2nd level of 2, there are 4 unique partitions, by unique division of described 0th level up and down 4 deciles obtain, non-mode number, its 4 sub-blocks are called the 1st sub-block (top left sub-block), the 2nd sub-block (upper right sub-block), the 3rd sub-block (lower-left sub-block), the 4th sub-block (bottom right sub-block) of unique division of the 2nd level;

Level number is level i.e. the 3rd level of 3, has 16 unique partitions, by 8 deciles about the 1st sub-block of the division two of described 1st level, and by the 3rd sub-block of unique division of described 2nd level, the 4th sub-block respectively left and right 4 deciles obtain, non-mode number.

an embodiment of a kind of variant of the embodiment 7G of predetermined level and division: embodiment 7A

Level number is level i.e. the 3rd level of 3, there are 29 partitions and 1 16 partition, any about 18 deciles (producing 29 partitions) in 2 sub-blocks of the division two of described 1st level or 2 all distinguished left and right 8 deciles (producing 1 16 partition) and obtain.

The difference of embodiment 7F and embodiment 7G is in the 3rd level, and unique division is changed into multiple division.

an embodiment of a kind of variant of the embodiment 7H of predetermined level and division: embodiment 7A

Level number is level i.e. the 3rd level of 3, there are 25 partitions, 18 partition, 29 partitions, 2 12 partitions and 1 16 partition, by any 1 (the only 1) left and right 8 deciles (producing 29 partitions) or upper and lower 4 deciles (producing 25 partitions) in 2 sub-blocks of the division two of described 1st level, or 2 all distinguish about 8 deciles or upper and lower 4 deciles (producing 18 partition, 2 12 partitions, 1 16 partition) obtain.

The difference of embodiment 7G and embodiment 7H is in the 3rd level, changes into " left and right 8 decile or upper and lower 4 deciles " from " left and right 8 decile " the bisecting method of the sub-block of the 1st level.

an embodiment of a kind of variant of the embodiment 7I of predetermined level and division: embodiment 7A

Level number is level i.e. the 3rd level of 3, there are 45 partitions, 28 partitions, 49 partitions, 4 12 partitions and 2 16 partitions, by any 1 (only 1) upper and lower 8 deciles (producing 29 partitions) in 2 sub-blocks of the division one of described 1st level or left and right 4 decile (producing 25 partitions), or all distinguish upper and lower 8 deciles or left and right 4 decile (produces 18 partition for 2, 2 12 partitions, 1 16 partition), by any 1 (the only 1) left and right 8 deciles (producing 29 partitions) in 2 sub-blocks of the division two of described 1st level or upper and lower 4 deciles (producing 25 partitions), or 2 all distinguish left and right 8 deciles or upper and lower 4 deciles (produce 18 partition, 2 12 partitions, 1 16 partition) obtain.

The difference of embodiment 7I and embodiment 7H is in the 3rd level, changes two divisions (division one and division two) to the decile of the sub-block of the 1st level into from a division (dividing two).

an embodiment of a kind of variant of the embodiment 7J of predetermined level and division: embodiment 7A

Level number is level i.e. the 3rd level of 3, has multiple division, is obtained by following manner:

By farthest decile or the left and right farthest decile up and down of at least 1 sub-block in 2 sub-blocks of the division one of described 1st level;

By farthest decile or up and down farthest decile about at least 1 sub-block in 2 sub-blocks of the division two of described 1st level;

By farthest decile or up and down farthest decile about at least 1 sub-block in 4 sub-blocks of unique division of described 2nd level.

Embodiment 7J is the summary of embodiment 7A ~ 7I.

Above-described embodiment 7A, the difference of 7B, 7C, 7D, 7E, 7F is only the difference of expression form, and the level finally obtained and division are all of equal value and identical mutually, as shown in Fig. 1 (a).Above-described embodiment 7G, 7H, 7I, 7J are the direct of embodiment 7A, 7B, 7C, 7D, 7E, 7F and simple variation.

the embodiment 8 of predetermined level and division

Level number is level i.e. the 1st level of 1, have two 2 partitions, the mode number of 2 partitions is 1, is obtained by unique division left and right 2 deciles of described 0th level, the mode number of another 2 partition is 2, is obtained by upper and lower 2 deciles of unique division of described 0th level;

Level number is level i.e. the 2nd level of 2, has 4 unique partitions, by unique division of described 0th level up and down 4 deciles obtain, non-mode number;

Level number is level i.e. the 3rd level of 3, has 16 unique partitions, is obtained by upper and lower 4 deciles of unique division left and right 4 deciles of described 0th level, non-mode number.

Above-described embodiment 6 the 0th, 1, the division of 2 levels and embodiment 8 the 0th, 1, the division of 2 levels is identical, and the 3rd level of embodiment 6 is identical from the sub-block number of unique division of the 3rd level of embodiment 8 but sub-block shape is different.

Similar with embodiment 7A ~ 7J, embodiment 8 also has execution mode and the variant of several equivalence.

the embodiment 9 of predetermined level and division

Level number is level i.e. the 3rd level of 3, has 8 unique partitions, is obtained by upper and lower 4 deciles of unique division left and right 2 deciles of described 0th level, non-mode number.

Above-described embodiment 6 the 0th, 1, the division of 2 levels and embodiment 9 the 0th, 1, the division of 2 levels is identical, and the 3rd level of embodiment 6 is all different with sub-block shape from the sub-block number of unique division of the 3rd level of embodiment 9.

Similar with embodiment 7A ~ 7J, embodiment 9 also has execution mode and the variant of several equivalence.

the embodiment 10 of predetermined level and division

Level number is level i.e. the 3rd level of 3, has 8 unique partitions, is obtained by upper and lower 2 deciles of unique division left and right 4 deciles of described 0th level, non-mode number.

Above-described embodiment 9 the 0th, 1, the division of 2 levels and embodiment 10 the 0th, 1, the division of 2 levels is identical, and the 3rd level of embodiment 9 is identical from the sub-block number of unique division of the 3rd level of embodiment 10 but sub-block shape is different.

Similar with embodiment 7A ~ 7J, embodiment 10 also has execution mode and the variant of several equivalence.

the embodiment 11 of predetermined level and division

Level number is level i.e. the 3rd level of 3, has 8 unique partitions, is obtained by upper and lower 8 deciles of unique division of described 0th level, non-mode number.

Above-described embodiment 9 the 0th, 1, the division of 2 levels and embodiment 11 the 0th, 1, the division of 2 levels is identical, and the 3rd level of embodiment 9 is identical from the sub-block number of unique division of the 3rd level of embodiment 11 but sub-block shape is different.

Similar with embodiment 7A ~ 7J, embodiment 11 also has execution mode and the variant of several equivalence.

the embodiment 12 of predetermined level and division

Level number is level i.e. the 3rd level of 3, has 8 unique partitions, is obtained by unique division left and right 8 deciles of described 0th level, non-mode number.

Above-described embodiment 9 the 0th, 1, the division of 2 levels and embodiment 12 the 0th, 1, the division of 2 levels is identical, and the 3rd level of embodiment 9 is identical from the sub-block number of unique division of the 3rd level of embodiment 10 but sub-block shape is different.

Similar with embodiment 7A ~ 7J, embodiment 12 also has execution mode and the variant of several equivalence.

containing the embodiment of the compressed bit stream of level number, mode number and displacement vector group and other coding results

Described encoding block containing in the compressed bit stream of partition mode (comprise level number, mode number) and displacement vector group and other coding results (from the angle of decoder also referred to as decoding block) part is made up of the syntactic element being loaded with following message:

Coding build, partition mode (comprising level number, mode number or sky), displacement vector 1 horizontal component, displacement vector 1 vertical component, displacement vector 2 horizontal component, displacement vector 2 vertical component ..., displacement vector K horizontal component, displacement vector K vertical component, other coding results;

Wherein, except described coding build syntactic element, it not is unique that the placement of every other institute's syntax elements in code stream puts in order, and can adopt any one predetermined rational order; Described partition mode syntactic element preferably can be splitted into a few part, is placed on places different in code stream respectively; Described level syntactic element preferably can be splitted into a few part, is placed on places different in code stream respectively; Described mode number syntactic element preferably can be splitted into a few part, is placed on places different in code stream respectively; Described level syntactic element and described mode number syntactic element can preferably be split and be reassembled into a few part, are placed on places different in code stream respectively; Partly or entirely can not preferably directly being placed in code stream of part or all of, the displacement vector syntactic element of part or all of, the mode number syntactic element of part or all of, the level syntactic element of described partition mode syntactic element, but derive from other codings or decoding parametric; Part or all of, the displacement vector syntactic element of part or all of, the mode number syntactic element of part or all of, the level syntactic element of described partition mode syntactic element partly or entirely can preferable through be placed in code stream after prediction computing or other computings; Other coding result syntactic elements described preferably can be splitted into a few part, are placed on places different in code stream respectively.

Accompanying drawing explanation

Fig. 1 is the example of predetermined level and division

Fig. 2 is coding method schematic flow sheet of the present invention

Fig. 3 is coding/decoding method schematic flow sheet of the present invention.

Claims

1. a method for encoding images, is characterized in that comprising the following steps:

The pixel sample value of step 1) to the block of an input carries out multi-level sub-block coupling coding, and namely what produce (1) optimum mates relative position by the displacement vector of the level number partition mode formed with optional mode number and (2) corresponding one group of optimum; Comprise: travel through predetermined each level and each division, in one among reconstructed reference pixel sample value district predetermined hunting zone, according to predetermined assessment level, search obtains each displacement vector of an optimum partition mode (comprising optimum level and its optimal dividing) and each current sub-block of correspondence and the optimum of each reference sub-block; Described each displacement vector is exactly the difference of the position coordinates searching for each current sub-block described in the position coordinates of described each reference sub-block optimum among the pixel sample value of the reconstructed reference district that determines and described encoding block; The unit of institute's displacement vector is the min coordinates unit of pixel sample value (Integer Pel sample value or 1/2nd, 1/4th, 1/8th pixel sample values); The reference pixel that institute's displacement vector points to is positioned at same image or same band, and its position range is the part or all of of the reconstruction region of view picture present image; The output of this step is the partition mode of described optimum, and the level comprising optimum level and division is number with optional mode number, described each displacement vector (also referred to as displacement vector group) and mate residual error; Described coupling residual error is exactly the difference of the numerical value of the pixel sample value of each reference sub-block described in the numerical value of the pixel sample value of each current sub-block described in described encoding block and the described pixel sample value of reconstructed reference district;

Step 2) remaining conventional coding and reconstruction step; The input of this step is above-mentioned steps 1) output and input pixel sample value; The output of this step is reconstructed pixel (comprising Perfect Reconstruction pixel and partial reconfiguration pixel in various degree) and the compressed bit stream containing partition mode (including but not limited to level information and optional mode number information) and displacement vector group and other coding results; Described reconstructed pixel puts into reconstructed reference pixel sample value working area, as the reference pixel required for subsequent multi-layer second son Block-matching coding step, all the other conventional codings and reconstruction step; Described compressed bit stream is also the last output of this coding method.

2. a picture decoding method, is characterized in that, comprises

3. a picture decoding method, is characterized in that comprising the following steps:

Step 1) is resolved compressed bit stream, obtains the partition mode and the displacement vector group that include level information and optional mode number information; Described partition mode represents the one in predetermined multiple division;

Step 2) the multi-level sub-block that uses partition mode and displacement vector group to carry out a decoding block copies decoding; Comprise: in the described decoding block that partition mode is specified according to each displacement vector of described mobile vector group, the position of each reference sub-block is determined in the position of each current sub-block, the direct or indirect assignment of numerical value of the pixel sample value of described each reference sub-block is given described each current sub-block; The reference pixel of described each displacement vectors directed is positioned at same image or same band, and its position range is the part or all of of the reconstruction region of view picture present image;

Step 3) remaining conventional decoding and reconstruction step.

4. coding method according to claim 1 or the coding/decoding method described in Claims 2 or 3, it is characterized in that: described encoding block or decoding block are a coding region or a decoding region of image, comprise following at least one: maximum coding unit LCU, code tree unit CTU, the subregion of coding unit CU, CU, predicting unit PU, converter unit TU.

5. coding method according to claim 1 or the coding/decoding method described in Claims 2 or 3, it is characterized in that: described predetermined level and division have 3 features simultaneously: 1) have 2 or 3 levels, 2) level number be 0 level only have a size to be equal to the sub-block of described encoding block or decoding block, 3), in all the other levels, a level is had to have 4 partitions of non-4 deciles up and down at least.

6. coding method according to claim 1 or the coding/decoding method described in Claims 2 or 3, it is characterized in that: described predetermined level and divide there are 2 features simultaneously: 1) level number be 0 level only have a size to be equal to the sub-block of described encoding block or decoding block, 2), in all the other levels, a level is had to have division more than 4 sub-blocks.

7. coding method according to claims 1 to 3 or coding/decoding method, is characterized in that: described predetermined multiple division includes but not limited to the part or all of of following level and division:

0th level, has 1 unique partition, is equal to described present encoding block or decoding block, non-mode number;

1st level, there are two 2 partitions, the mode number of 2 partitions is 1, unique division left and right 2 deciles of described 0th level are obtained, its 2 sub-blocks are called the 1st sub-block (left sub-block) and the 2nd sub-block (right sub-block) of the division one of the 1st level, the mode number of another 2 partition is 2, is obtained by upper and lower 2 deciles of unique division of described 0th level, and its 2 sub-blocks are called the 1st sub-block (upper sub-block) and the 2nd sub-block (lower sub-block) of the division two of the 1st level;

2nd level, there are 4 unique partitions, by unique division of described 0th level up and down 4 deciles obtain, non-mode number, its 4 sub-blocks are called the 1st sub-block (top left sub-block), the 2nd sub-block (upper right sub-block), the 3rd sub-block (lower-left sub-block), the 4th sub-block (bottom right sub-block) of unique division of the 2nd level;

3rd level, has 16 unique partitions, by 8 deciles about the 1st sub-block of the division two of described 1st level, and by the 3rd sub-block of unique division of described 2nd level, the 4th sub-block respectively left and right 4 deciles obtain, non-mode number.

8. coding method according to claims 1 to 3 or coding/decoding method, is characterized in that: described predetermined multiple division includes but not limited to the part or all of of following level and division:

9. coding method according to claims 1 to 3 or coding/decoding method, is characterized in that: described predetermined multiple division includes but not limited to the part or all of of following level and division:

10. coding method according to claims 1 to 3 or coding/decoding method, is characterized in that:

Encoding block part in described compressed bit stream is made up of the syntactic element being loaded with following message: