CN112532990B

CN112532990B - String length parameter coding and decoding method and device

Info

Publication number: CN112532990B
Application number: CN202011324094.XA
Authority: CN
Inventors: 赵利平; 林涛; 周青阳; 周开伦; 焦孟草; 胡珂立; 沈士根
Original assignee: University of Shaoxing
Current assignee: University of Shaoxing
Priority date: 2020-11-23
Filing date: 2020-11-23
Publication date: 2024-03-08
Anticipated expiration: 2040-11-23
Also published as: CN112532990A

Abstract

The invention discloses a string length parameter coding and decoding method and device. The invention adopts the next residual number as the string length parameter to write the compressed data code stream, thereby further improving the compression rate. According to the current common string-primitive coding mode, the value of the next remaining number is usually smaller than the string length value, so that the compressed data code stream consumption is reduced. The preferred scheme has a relatively remarkable effect of improving the compression rate when the number of strings is less than 1/4 of the current block size.

Description

String length parameter coding and decoding method and device

Technical Field

The invention belongs to the field of data compression, and in particular relates to a string length parameter coding and decoding method and device.

Background

As human society has entered the era of artificial intelligence, big data, virtual reality, augmented reality, mixed reality, cloud computing, mobile computing, cloud-mobile computing, ultra-high definition 4K and ultra-high definition 8K video image resolution, 4G/5G communication, various data sets including big data, image data, video data, and various new forms of data have become major consumers of internet bandwidth, and ultra-high compression ratio and ultra-high quality data compression thereof have become indispensable technologies.

A dataset is a collection of finite data (e.g., a one-dimensional data queue, a two-dimensional data file, a frame of image, a video sequence, a transform domain, a transform block, multiple transform blocks, a three-dimensional scene, a sequence of continuously changing three-dimensional scenes) that is organized into a spatial (one-dimensional, two-dimensional, or multi-dimensional) shape, consisting of samples of data, also called elements (e.g., bytes, bits, pixels, pixel components, spatial sampling points, transform domain coefficients). When encoding (and correspondingly decoding) a data set, in particular a two-dimensional or more data set, the data set is typically divided into a number of subsets having a predetermined shape, called encoded blocks (from the decoding point of view, i.e. decoded blocks, collectively encoded blocks or decoded blocks), and a block is encoded or decoded in a predetermined temporal order in units of encoded blocks or decoded blocks. At any one time, the coding block being coded is referred to as the current coding block. At any one time, the decoding block being decoded is referred to as the current decoding block. The current encoded block or the current decoded block is collectively referred to as a current encoded block or a decoded block or simply a current block. The sample being encoded or decoded is referred to as the current encoded sample or current decoded sample, simply referred to as the current sample, also referred to as the current element.

For an encoded block or a decoded block having a certain shape, it is necessary in many cases to divide it into finer cells (base units), and to encode or decode one cell by one cell in a predetermined time sequence. The same type of encoding or decoding operation is typically performed on all samples within a primitive. At any one time, the primitive being encoded or decoded is referred to as the current primitive. The result of encoding a primitive is one or more encoding parameters that ultimately produce a compressed data stream containing the encoding parameters. Decoding a primitive is to parse the compressed data stream to obtain one or more coding parameters, and recover samples of reconstructed data from the one or more coding parameters. Examples of primitives include coded or decoded blocks (whole block as a primitive), sub-blocks, micro-blocks, strings, byte strings, alpha (alpha) strings, pixel unit strings, pixel strings, sample strings, index strings, lines.

The size of an encoded block or a decoded block is the number of elements or the equivalent number of pixels contained in the encoded block or the decoded block. For example, in the case where one element contains 4×1 pixels, the size of one encoded block or decoded block containing 32×32 pixels, i.e., 8×32 elements, is 1024 pixels, i.e., 256 elements.

Typical representations of string-primitive coding modes include the latest international video standard VVC (Versatile Video Coding) standard and the national AVS3 (the third generation of Audio Video Coding Standard) standard, and the Palette (PLT) mode and string matching (also known as string prediction or intra-frame string replication, ISC) mode in the previous version of the standard.

In PLT mode, a Palette is used to hold representative colors, also called primary colors, of the current CU or the nearest CU. For one input CU, first, pixels in the CU are subjected to a palette generation method to obtain palettes, then, main colors in each palette are represented by index (index), and finally, pixels in the CU are mapped to indexes corresponding to the main colors. The PLT mode represents pixels of the entire CU with a main color, a non-main color, and an index table in the palette, and thus the main color, the non-main color, and the index table need to be encoded. The coding of the index table adopts a certain scanning sequence, uses two index copying modes, namely a copying left mode and a copying upper mode, carries out entropy coding on the index copying mode and the length (length), and writes the code stream. The non-primary color directly transmits the pixel value itself.

The ISC mode is similar to the PLT mode, and is also a mode using strings as primitives, and uses a sample, also called an element (such as a pixel unit or a pixel), as a minimum copy unit, in a given reference pixel set, according to a predetermined scanning sequence, if an optimal reference sample (such as a pixel unit or a pixel) string can be found in the reference pixel set, a sample (such as a pixel unit or a pixel) string copy relationship parameter displacement vector and a length (length) are used to respectively represent the position relationship between the reference sample string (abbreviated as a reference string) and the current sample string to be coded (also called a current string), otherwise, the current incompletely matched sample (such as an incompletely matched pixel unit, which is a pixel unit whose mxn pixels or components or indexes have at least one unmatched pixel or component or index in which the reference sample cannot be found) is coded. Typically, the displacement vector is equal to the reference string starting position minus the current string starting position.

Taking an example of a coding block or decoding block of size 32x32 pixels (1024 pixels in size of the coding block or decoding block) divided into n pixel unit strings (such as those in the ISC mode, each pixel unit is composed of 4x1 pixels, each coding block or decoding block is composed of 256 pixel units, i.e., 256 elements in size of the coding block or decoding block) and/or m incompletely matched pixel units, the length values of the n strings are length in turn ₁ 、length ₂ 、......length _n (wherein,and->) A sequence of string lengths is formed. Since the sequence of string lengths has some inherent regularity and correlation: taking the example of a string of pixels, the probability of a string having a string length equal to the size of the entire encoded block is relatively high, and the sum of all string lengths and the number of unmatched pixels is equal to the size of the encoded block. Thus, redundancy exists in the sequence of string lengths, stringsThe length sequence direct coding or decoding scheme does not remove this redundancy, losing coding efficiency.

Disclosure of Invention

Aiming at the defects or improvement demands of the prior art, the invention provides a string length parameter coding and decoding method and device, which aim at adopting the next residual number as the string length parameter to write in a compressed data code stream, further improving the compression rate, thereby solving the technical problem that the existing string length is taken as the string length parameter to directly code the compression rate to be improved.

To achieve the above object, according to one aspect of the present invention, there is provided a string length parameter encoding method or apparatus, which includes at least steps or modules for performing the following functions and operations:

(1) Based at least on current block size, current string length _i And the sum of the lengths of the currently encoded elements curTlen, calculates the next remaining number nrp _i String length parameters as current string;

the next remaining number nrp _i For the number of elements remaining in the current block after encoding the current string;

(2) At least information representing the string length parameter values is written into the compressed data stream.

According to another aspect of the present invention, there is provided a string length parameter decoding method including at least the steps of:

(A) Analyzing the compressed data code stream in the current block to at least obtain the information containing the string length parameter value of the current string; the string length parameter is the next remaining number nrp _i ；

The next remaining number nrp _i For the number of elements remaining in the current block after decoding the current string;

(B) Obtaining a current string length according to a preset string length calculation mode at least according to the current block size curize, the sum curTlen of the number of elements which are decoded currently and the next residual number _i 。

Preferably, the string length parameter decoding method obtains a current string length according to a predetermined string length calculation mode _i The method specifically comprises the following steps:

Current string length _i The method is as follows:

length _i ＝cuSize-curTlen-nrp _i

or alternatively

length _i ＝R _i -nrp _i

Wherein R is _i The number of current remaining elements; preferably the number R of the current remaining elements _i For the difference between the current block size curize and the sum curTlen of the lengths of the currently decoded elements, calculating the number R of the current remaining elements in the current block _i The following are provided:

R _i ＝cuSize-curTlen

preferably, the value range of the next remaining number is [0, R _i -1]。

Preferably, the string length parameter decoding method, when R _i When=1, then nrp _i The value is set to 0, and the next remaining number nrp does not exist in the compressed data code stream _i Is a binary symbol string of (2); otherwise, i.e. R _i > 1, there is the next remaining number nrp in the compressed data stream _i Is inverse binarized the next remaining number nrp _i The next remaining number nrp obtained for the binary symbol string of (2) _i 。

Preferably, the string length parameter decoding method, the inverse binarizing the next remaining number nrp _i The next remaining number nrp obtained for the binary symbol string of (2) _i The method specifically comprises the following steps:

nrp the next remaining number in the compressed data stream of the current string in the current block _i Is decoded to include the next remaining number nrp according to a preset inverse binarization scheme _i For marking the next remaining number nrp, including the number of the segment in which the value is located _i A parameter vector of values; according to a preset next remaining number nrp _i The value range is divided into a plurality of linksSuccessive schemes of sequentially numbered segments will include the next remaining number nrp _i For marking the next remaining number nrp, including the number of the segment in which the value is located _i Conversion of the parameter vector of values to the next remaining number nrp _i Is a value of (2); preferably the preset next remaining number nrp _i The scheme of dividing the value range into a plurality of consecutive sequentially numbered segments is based on the next remaining number nrp _i Value range determination, preferably the one remaining number nrp _i The value range is based on the number R of the current residual elements _i And (5) determining.

Preferably, the string length parameter decoding method, the inverse binarizing the next remaining number nrp _i Is composed of at least two parts:

the first part adopts an inverse binarization mode of truncated unitary code or truncated binary code or truncated k-order exponential golomb code or fixed-length code derived from a maximum value;

the second part adopts the inverse binarization mode of the segmented truncated binary code, the segment number and/or the dividing point of the segment are derived from at least the current residual number, the segment number is allowed to be 0, and the second part is null or not.

Preferably, the string length parameter decoding method, the inverse binarizing the next remaining number nrp _i Is composed of at least three parts:

the first part adopts an inverse binarization mode with a maximum value, wherein the inverse binarization mode comprises a truncated unitary code or a truncated binary code or a truncated k-order exponential golomb code or a fixed-length code derived from a maximum value;

the second part adopts a segmented inverse binarization mode of a fixed-length code derived from a maximum value, the segment number and/or the demarcation point of the segment are derived from at least the current residual number, the segment number is allowed to be 0, and the second part is null or does not exist;

the third part adopts a reverse binarization mode of segmented fixed-length codes, the number of segments and/or the dividing points of the segments are derived from at least the first part and/or the second part, the number of segments is allowed to be 0, and the third part is empty, namely, is not present.

the next remaining number nrp is counted using a truncated code having a truncated value, i.e. a maximum maxVal _i Is inverse binarized, the truncated code comprising one or a combination of the following codes:

Truncating unary code, or

Truncated binary codes, i.e. truncated fixed-length codes, having at most two code lengths, or

The fixed-length code derived from said maximum value is preferably composed of the shorter of the truncated binary codes with said maximum value, or

Cutting off an exponential golomb code of a preset order, namely cutting off an exponential golomb code of k orders;

preferably, at least according to the number R of the current remaining elements _i Determining the maximum value maxVal; preferably, the maximum maxVal is allowed to be 0, which indicates that the corresponding code is a null code, i.e., a binary symbol string does not actually exist.

Preferably, the string length parameter decoding method, the preferred element is a pixel unit consisting of 4x1 pixels or components or indexes;

its said inverse binarizing the next remaining number nrp _i The next remaining number nrp obtained for the binary symbol string of (2) _i One of the following two schemes is applied;

scheme one:

the binary symbol string of the next remaining number comprises at least three parts, the inverse binarization of which comprises at least the steps of:

performing inverse binarization on a first part of the binary symbol string by adopting a truncated code comprising a truncated unary code or a truncated binary code or a truncated k-order exponential golomb code;

the second part of the binary symbol string is inverse binarized by selecting one of predetermined I+1 cut-off codes, each of which is a fixed-length code derived from a maximum value Or truncated binary codes with a maximum value; the maximum value is at least based on the number R of the current residual elements _i And said first portion is derived, the maximum allowed to be 0, indicating that said second portion is empty, i.e., not present;

performing inverse binarization on a third part of the binary symbol string by adopting a fixed-length code with a code length k, wherein the code length k is derived from at least the first part and/or the second part, the code length k is allowed to be 0, and the third part is null or is not present;

scheme II:

the binary symbol string of the next remaining number includes a flag bit, and at least three parts; the flag bit is a one-bit binary symbol, namely a fixed-length code with the code length of 1 indicates whether the next residual number is zero or not;

when the result of the inverse binarization of the binary symbol of the flag bit indicates that the next remaining number is not zero, that is, equal to or greater than 1, the inverse binarization of the portion other than the binary symbol string flag bit of the next remaining number is the next remaining number minus 1, and the inverse binarization is performed on the portion other than the binary symbol string flag bit of the next remaining number, including at least the following steps:

Selecting one of predetermined i+1 truncated codes for the second portion of the binary symbol string, each of the i+1 truncated codes being a fixed-length code derived from a maximum value or a truncated binary code having a maximum value, to inverse binarize the second portion of the binary symbol string; the maximum value is at least based on the number R of the current residual elements _i And said first portion is derived, the maximum allowed to be 0, indicating that said second portion is empty, i.e., not present;

and performing inverse binarization on a third part of the binary symbol string by adopting a fixed-length code with a code length of k, wherein the code length k is derived from at least the first part and/or the second part, and the code length k is allowed to be 0, so that the third part is null, namely, is not present.

Preferably, the next remaining number nrp _i Or the total value range of the next remaining number minus 1 is divided into a predetermined Imax+1 integer interval [ N0 ] _i ,N0 _i+1 )，i＝0～Imax；

The I in the second part is that the condition N0 is satisfied _I ≤R-1<N0 _I+1 I.e. N0 _I <R≤N0 _I+1 I of (2); in scheme one, R is equal to R _i The method comprises the steps of carrying out a first treatment on the surface of the Alternatively, in scheme two, R is equal to R _i -1；

The maximum value maxValPrefix of the truncated code adopted by the first part is equal to I, and the truncated code with the maximum value maxValPrefix is adopted to carry out inverse binarization on the first part of the binary symbol string to obtain a value a of the first part;

The second part selects an ith truncated code in the preset I+1 truncated codes to inverse binarize the second part of the binary symbol string; said i being equal to the value a of said first portion; the maximum value maxvallnix of the truncated code used by the second part is equal to =n _i+1 -N _i -1, wherein N _i ＝N0 _i ，N _i+1 ＝min(N0 _i+1 R), i.e. maxvallnfix=min (N0) _i+1 -N0 _i -1,R-N0 _i -1); the next remaining number nrp _i Or the next remaining number minus 1 is [ N ] _i ,N _i+1 )。

Preferably, in scheme one, imax=4, i is 0, 1, 2, 3, or 4, N0 ₀ 、N0 ₁ 、N0 ₂ 、N0 ₃ 、N0 ₄ 、N0 ₅ R is the maximum value of 0, 1, 5, 21, 277, R _i Is the maximum value of (2);

or,

in scheme two, imax= 3,i is 0, 1, 2, or 3, N0 ₀ 、N0 ₁ 、N0 ₂ 、N0 ₃ 、N0 ₄ R is the maximum value of 0, 4, 20, 276 and R respectively _i -a maximum value of 1.

Preferably, in scheme one, imax= 3,i is 0, 1, 2, or 3, N0 ₀ 、N0 ₁ 、N0 ₂ 、N0 ₃ 、N0 ₄ R is the maximum value of 0, 1, 5, 21, R _i Is the maximum value of (2);

or,

in scheme two, imax=2, i is 0, 1, or 2, N0 ₀ 、N0 ₁ 、N0 ₂ 、N0 ₃ R is the maximum value of 0, 4, 20 and R _i -a maximum value of 1.

According to another aspect of the present invention, there is provided a string length parameter decoding apparatus including: the device comprises a compressed data code stream analysis module and a string length calculation module;

The compressed data code stream analysis module is used for storing at least the current block size cuSize and the sum curTlen of the current decoded element length, obtaining at least the information of the string length parameter value representing the current string, and submitting the information to the string length calculation module; the string length parameter is the next remaining number nrp _i The method comprises the steps of carrying out a first treatment on the surface of the The next remaining number nrp _i Is the number of elements remaining in the current block after decoding the current string;

the string length calculation module is used for obtaining a string length according to a preset string length calculation mode at least according to the current block size cuSize, the sum curTlen of the number of elements which are decoded currently and the next residual number _i 。

Preferably, the string length calculation module is configured to calculate a length according to the current string length _i And updating the sum curTlen of the lengths of the elements currently decoded; the string length information is the next remaining number nrp _i ；

Preferably, the string length is obtained according to a predetermined string length calculation mode _i The sum curTlen of the currently decoded element length subtracted from the current block size cuSize and the next remaining number nrp _i The method is characterized by comprising the following steps:

length _i ＝cuSize-curTlen-nrp _i

or alternatively

length _i ＝R _i -nrp _i

Wherein R is _i As the wayThe number of the front remaining elements;

Preferably, the obtaining the information of the string length parameter value representing the current string is specifically that the obtaining the binarized compressed data code stream of the current string in the current block, and the performing inverse binarization to obtain the code containing the string length parameter information of the current string is specifically that:

calculating the number R of the current residual elements according to the current block size cuSize and the sum curTlen of the lengths of the elements which are decoded currently _i ，

The number R of the current residual elements _i For the difference between the current block size curize and the sum curTlen of the lengths of the currently decoded elements, calculating the number R of the current remaining elements in the current block _i The following are provided:

R _i ＝cuSize-curTlen

preferably, the value range of the next remaining number is [0, R _i -1]；

Preferably, the element includes a pixel or a pixel component, a pixel index, an index, or a pixel unit; the pixel unit is an array of mxn pixels or components or indexes, m and n being predetermined positive integers including 1, 2, 4, 8, 16, respectively; when m and n are both 1, the pixel unit is a pixel or a pixel component or a pixel index or an index; the preferred element is a pixel cell consisting of 4x1 pixels or components or indices.

Preferably, when R _i When=1, then nrp _i The value is set to 0, and the next remaining number nrp does not exist in the compressed data code stream _i Is a binary symbol string of (2); otherwise, i.e. R _i > 1, then there is the next remaining number nrp in the compressed data stream _i Is inverse binarized to obtain the next remaining number nrp _i 。

Preferably the next remaining number nrp obtained by said inverse binarization _i The method specifically comprises the following steps:

according to the number R of the current residual elements _i Determining the next remaining number nrp _i Value range and obtain its next remaining number nrp _i The value range being divided into a plurality of successive valuesA scheme of sequentially numbered segments; according to a preset next remaining number nrp _i Scheme for dividing value range into a plurality of consecutive sequentially numbered segments, said next remaining number nrp being included _i For marking the next remaining number nrp, including the number of the segment in which the value is located _i Conversion of the parameter vector of values to the next remaining number nrp _i To decode a binarized compressed data stream of a current string in a current block to include the next remaining number nrp according to a preset inverse binarization scheme _i For marking the next remaining number nrp, including the number of the segment in which the value is located _i A parameter vector of values.

Preferably, the string length calculation module updates the sum curTlen of the lengths of the currently decoded elements to be the current string length _i And the sum of the currently decoded element lengths curTlen updates the sum of the currently decoded element lengths curTlen; the method comprises the following steps:

curTlen＝curTlen+length _i 。

the compressed data stream parsing module is preferably configured to use a truncated code with a truncated value, i.e. a maximum value maxVal, to inverse binarize the binary symbol string of the next remaining number, where the truncated code includes one or a combination of the following codes:

truncating unary code, or

A fixed-length code derived from the maximum value, consisting of shorter codes of truncated binary codes having the maximum value, or

preferably, at least according to the number R of the current remaining elements _i Determining the maximum value maxVal; the maximum maxVal is allowed to be 0, indicating that the corresponding code is a null code, i.e., the binary symbol string is not actually present

Preferably, the next remaining number of binary symbol strings and the inverse binarization operation thereof are composed of at least three parts;

when the number R of the current residual elements _i When the binary symbol string is equal to 1, all parts of the binary symbol string are empty, namely the binary symbol string does not exist in the compressed data code stream, and the value of the next remaining number is 0;

when the number R of the current residual elements _i Above 1, inverse binarization is performed using one of two schemes:

scheme one:

Scheme II:

or,

In general, the above technical solutions conceived by the present invention, compared with the prior art, enable the following beneficial effects to be obtained:

according to the current common string-primitive coding mode, the value of the next remaining number is usually smaller than the string length value, so that the compressed data code stream consumption is reduced. The preferred scheme has a more remarkable effect of improving the compression rate when the data of the string is smaller than 5/16 of the current block size, for example, smaller than 1/4 and 3/16.

In the preferred scheme, the number relation among the next residual numbers is considered, and the corresponding binarization scheme is adopted, so that the consumption of compressed data code streams can be further reduced, and the compression rate is improved.

Detailed Description

The present invention will be described in further detail with reference to the following examples in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention. In addition, the technical features of the embodiments of the present invention described below may be combined with each other as long as they do not collide with each other.

The invention provides a string length parameter coding method, which is suitable for a data compression process taking strings as primitives, and comprises the following steps:

specific examples are:

acquiring current block size cuSize and current string length _i And the sum of the lengths of the currently encoded elements curTlen, calculates the next remaining number nrp _i String length parameters as current string;

the next remaining number nrp _i For the number of remaining elements in the current block after encoding the current string, i.e. for the number of all elements in the current block after the current string, i.e. the current block size, cuSize minus the sum of the lengths of the currently encoded elements curTlen and the current string length _i The method is characterized by comprising the following steps:

nrp _i ＝cuSize-curTlen-length _i

where cuSize is the current block size, curTlen is the sum of the lengths of the currently encoded elements, length _i For the current string length, i=1, 2 … …;

the total number of strings in the current block does not exceed 3/16, 1/4, or 5/16 of the current block size.

(2) Writing at least information representing the string length parameter value into a compressed data stream; the method comprises the following steps:

writing at least information representing the string length parameter values obtained in said step (1) into a compressed data stream; one of the specific methods is to binarize the current string length parameter obtained in the step (1), and write the binarized data into the compressed data code stream;

the binarization scheme of the string length parameter, namely the next remaining number nrp _i According to the binarization scheme, binarization is preferably carried out according to the distribution probability of the next remaining number and according to the principle that the code after binarization is shorter when the distribution probability is higher, so that the string length parameter value with higher occurrence frequency consumes fewer bits, and the compression rate is further improved;

according to current coding methods, especially for image or video data, the elements (including pixels or pixel units) in one coding block are likely to be divided into the same string, so that the frequency of the next remaining number 0 is quite high, coding with as few bits as possible, specifically as follows:

the number R of the current residual elements _i For the difference between the current block size curize and the sum curTlen of the lengths of the currently encoded elements, calculating the number R of the current remaining elements in the current block _i The following are provided:

R _i ＝cuSize-curTlen

if R is _i =1, then nrp _i Only can take the value of 0, and the next residual number nrp does not exist in the compressed data code stream _i Is a binary symbol string of (2); otherwise, i.e. R _i > 1, then nrp _i The next remaining number nrp to be _i Binarization.

The next remaining number nrp _i Binarization, specifically includes:

according to the number R of the current residual elements _i Determining the next remaining number nrp _i Value field and the next remaining number nrp _i The value range is divided into a plurality of consecutive sequentially numbered segments to include the next remaining number nrp _i Parameter directions including the number of the segment in which the value is locatedQuantity marks the next remaining number nrp _i And for said value for marking the next remaining number nrp _i Including the next remaining number nrp _i Binarizing the parameter vector including the number of the segment in which the value is located as the next remaining number nrp _i Is encoded by the binarization of (a).

Preferably, the first segment of the sequentially numbered segments contains only one value of 0.

Due to nrp _i The value range can be defined by the number R of the current residual elements _i Determine, therefore, by reasonably dividing the next remaining number nrp _i The value range can effectively reduce the value to be encoded, thereby further reducing the number of bits consumed.

(3) Length in current string length _i And the sum of the currently encoded element lengths curTlen updates the sum of the currently encoded element lengths curTlen; the method specifically comprises the following steps:

curTlen＝curTlen+length _i

the invention adopts the length parameter of the next remaining number to carry out the coding scheme on at least one string length, so that the length parameter of the next remaining number is smaller than the value of the direct coding string length under partial conditions, thereby reducing the bit number consumed by entropy coding of a part of length values in the pixel unit string or the sequence of the pixel string length and further improving the coding efficiency of the string length sequence. The scheme has the advantages that the effect of improving the coding efficiency is more obvious under the condition that the number of strings is smaller, and the coding efficiency is obviously improved compared with the existing coding mode which directly takes the string length as the string length parameter under the condition that the total number of strings in the current block is not more than 5/16 of the current block size.

How to efficiently encode or decode length sequences in coding modes (such as PLT mode and ISC mode) that are primitive with pixel cell strings or pixel strings is a critical issue. Existing string-length sequence encoding or decoding schemes directly to length ₁ 、length ₂ 、......length _n Encoding or decoding is performed. One specific example is: let length ₁ 、length ₂ 、......length _n The sequence coding scheme of the existing string length is to directly code or decode four length values of 200, 8, 45 and 3. An important step in encoding or decoding is binarization or inverse binarization. The process of obtaining a string of binary symbols representing a value of a variable or parameter from this value is called binarization. The process of obtaining a value of a variable or parameter from a string of binary symbols representing the value is called inverse binarization. In general, the larger the value of a variable or parameter, the greater the number of bits required for binarization. Representing the value of the variable or parameter with a higher probability of occurrence by a smaller numerical value or reducing the value of the variable or parameter can further improve the encoding efficiency. The invention takes the residual string length of the string as the coding parameter, and the difference between the residual string length of the previous string and the residual string length of the current string is the current string length.

Coding modes with strings as primitives have greater prediction flexibility and higher prediction accuracy, but also bring about some drawbacks, such as high hardware implementation complexity. To solve the complexity problem of hardware implementation, it is often necessary to limit the total number of pixel unit strings into which the entire coding block is divided, for example, the total number of the entire pixel unit strings does not exceed one quarter of the entire coding block. Note that there is some inherent regularity and correlation of the sequence of string lengths: taking the example of pixel strings, the probability of strings having a string length equal to the size of the entire encoded block is relatively high. Therefore, there is redundancy in the string-length sequence, and the string-length sequence direct encoding or decoding scheme does not remove such redundancy, losing encoding efficiency.

The string is formed by arranging samples, also called elements, in a predetermined scanning manner. The samples include pixels or pixel components or pixel indices or pixel units. The pixel unit is an array of mxn pixels or components or indices, m and n being predetermined positive integers including 1, 2, 4, 8, 16, respectively. When m and n both take a value of 1, the pixel unit is a pixel or a pixel component or a pixel index or an index.

The invention provides a string length parameter decoding method, which comprises the following steps:

the method specifically comprises the following steps: sequentially obtaining a binarized compressed data code stream of a current string in a current block, and performing inverse binarization to obtain a code containing string length parameter information of the current string; the string length parameter is the next remaining number nrp _i The method comprises the steps of carrying out a first treatment on the surface of the More specifically:

R _i ＝cuSize-curTlen

if R is _i =1, then nrp _i The value is set to 0, since 0 is the only possible value, there is no next remaining number nrp in the compressed data stream _i Is a binary symbol string of (2); otherwise, i.e. R _i > 1, there is the next remaining number nrp in the compressed data stream _i Is inverse binarized to obtain the next remaining number nrp _i 。

The inverse binarization yields the next remaining number nrp _i The method specifically comprises the following steps:

decoding the binarized compressed data code stream of the current string in the current block into a code stream comprising the next remaining number nrp according to a preset inverse binarization scheme _i For marking the next remaining number nrp, including the number of the segment in which the value is located _i A parameter vector of values; according to a preset next remaining number nrp _i Scheme for dividing value range into a plurality of consecutive sequentially numbered segments, said next remaining number nrp being included _i For marking the next remaining number nrp, including the number of the segment in which the value is located _i Conversion of the parameter vector of values to the next remaining number nrp _i Values of (2)The method comprises the steps of carrying out a first treatment on the surface of the The preset next remaining number nrp _i The scheme of dividing the value range into a plurality of consecutive sequentially numbered segments is based on the next remaining number nrp _i Value range determination, the one remaining number nrp _i The value range is based on the number R of the current residual elements _i And (5) determining.

Preferably, the next remaining number of binary symbol strings are inverse binarized using a truncated code having a truncated value, i.e. a maximum value maxVal, the truncated code comprising one or a combination of the following codes:

truncating unary code, or

scheme one:

the inverse binarization of the string of binary symbols for the next remaining number consists of at least the following three part operations:

the first part adopts a truncated code comprising a truncated unary code or a truncated binary code or a truncated k-order exponential golomb code to carry out inverse binarization on the first part of the binary symbol string;

The second part selecting one of predetermined I+1 cut-off codes to inverse binarize the second part of the binary symbol string, each of the I+1 cut-off codes being a fixed-length code derived from a maximum value or a cut-off binary code having a maximum value; the maximum value is at least based on the number R of the current residual elements _i And said first portion is derived, the maximum allowed to be 0, indicating that said second portion is empty, i.e., not present;

the third part is subjected to inverse binarization by adopting a fixed-length code with a code length k, wherein the code length k is derived from at least the first part and/or the second part, the code length k is allowed to be 0, and the third part is null or does not exist;

scheme II:

firstly, a flag, namely a one-bit binary symbol, namely a fixed-length code with the code length of 1 is adopted to represent whether the next residual number is zero or not;

if the result of the inverse binarization of the binary symbols of the flag indicates that the next remaining number is equal to or greater than 1, subtracting one from the next remaining number, referred to as subtracting 1 from the next remaining number, and continuing to inverse binarize the string of binary symbols of the next remaining number minus 1, the inverse binarization of the next remaining number minus 1 consisting of at least three operations:

the third part is inverse binarized with a fixed-length code of code length k derived from at least the first and/or second part, the code length k being allowed to be 0, indicating that the third part is empty, i.e. not present.

or,

Preferably, in scheme one, imax= 3,i is 0, 1, 2,Or 3, N0 ₀ 、N0 ₁ 、N0 ₂ 、N0 ₃ 、N0 ₄ R is the maximum value of 0, 1, 5, 21, R _i Is the maximum value of (2);

or,

(B) Based at least on the current block size cuSize, the sum of the number of currently decoded elements curTlen, and the next remaining number nrp _i Obtaining the current string length according to a preset string length calculation mode _i . The method specifically comprises the following steps:

calculating the current string length according to the code containing string length parameter information obtained in the step (A) _i The method comprises the steps of carrying out a first treatment on the surface of the The string length information is the next remaining number nrp _i ；

The string length _i The sum curTlen of the currently decoded element length subtracted from the current block size cuSize and the next remaining number nrp _i The method is as follows:

length _i ＝cuSize-curTlen-nrp _i

or alternatively

length _i ＝R _i -nrp _i

R _i ＝cuSize-curTlen

preferably, the value range of the next remaining number is [0, R _i -1]；

(C) Length in current string length _i And the sum of the currently decoded element lengths curTlen updates the sum of the currently decoded element lengths curTlen; the method specifically comprises the following steps:

curTlen＝curTlen+length _i 。

the present invention provides a string length parameter encoding device, comprising: the device comprises a string length parameter acquisition module and a compressed code stream generation module;

the string length parameter obtaining module stores the current block size cuSize, obtains the current string and reads the current string length _i And the sum curTlen of the lengths of the currently encoded elements, and reading the current string length in parallel according to the current block size cuSize, the sum curTlen of the lengths of the previously encoded elements, and the current string _i Calculate the next remaining number nrp _i And the number of current remaining elements R _i The next remaining number nrp _i As string length parameter and number of current remaining elements R _i Submitting the compressed code stream to a compressed code stream generation module;

the next remaining number nrp _i For the number of remaining elements in the current block after encoding the current string, i.e. the sum of the lengths of all strings after the current string in the current block, i.e. the current block size, cuSize minus the sum of the lengths of the currently encoded elements curTlen and the current string length _i The method is characterized by comprising the following steps:

nrp _i ＝cuSize-curTlen-length _i

the compressed code stream generating module is at least used for writing information representing the string length parameter value into a compressed data code stream;

in particular for the number R of current remaining elements _i Binarizing the string length parameter, writing the binarized data into a compressed code stream, and updating the sum curTlen of the lengths of the elements which are coded currently;

R _i ＝cuSize-curTlen

if R is _i =1, then nrp _i The value is set to 0, since 0 is the only possible value, there is no next remaining number nrp in the compressed data stream _i Is a binary symbol string of (2); otherwise, i.e. R _i > 1, the next remaining number nrp _i Binarization.

The next remaining number nrp _i Binarization, specifically includes:

according to the number R of the current residual elements _i Determining the next remaining number nrp _i Value field and the next remaining number nrp _i The value range is divided into a plurality of consecutive sequentially numbered segments to include the next remaining number nrp _i The next remaining number nrp is marked by a parameter vector including the number of segments in which the value is located _i And for said value for marking the next remaining number nrp _i Including the next remaining number nrp _i Binarizing the parameter vector including the number of the segment in which the value is located as the next remaining number nrp _i Is encoded by the binarization of (a).

The sum curTlen of the length of the element which is coded currently is updated to be the length of the current string _i And the sum of the currently encoded element lengths curTlen updates the sum of the currently encoded element lengths curTlen.

The present invention provides a string length parameter decoding device, comprising: the device comprises a compressed data code stream analysis module and a string length calculation module;

The compressed data code stream analysis module is used for storing at least the current block size cuSize and the sum curTlen of the current decoded element length, obtaining at least the information of the string length parameter value representing the current string, and submitting the information to the string length calculation module; the string length parameter is the next remaining number nrp _i The method comprises the steps of carrying out a first treatment on the surface of the The next remaining number nrp _i Is the number of elements remaining in the current block after decoding the current string; in particular, the method comprises the steps of,

storing at least the current block size cuSize, obtaining at least the current string and the current stringFront string length _i And the sum of the lengths of the currently encoded elements curTlen, at least in terms of the current block size cuSize, the sum of the lengths of the currently encoded elements curTlen, and the current string length _i Calculate the next remaining number nrp _i At least the next remaining number nrp _i Submitting the string length parameter to the compressed code stream generating module as a string length parameter; the next remaining number nrp _i For the number of elements remaining in the current block after encoding the current string; preferably comprising an inverse binarization sub-module:

the inverse binarization sub-module is used for reading the current block size cuSize and the sum of the lengths of the elements which are decoded currently, and is used for acquiring a binarized compressed data code stream of a current string in the current block, performing inverse binarization to acquire information of the current string containing string length parameters and submitting the information to the string length acquisition module; the string length parameter is the next remaining number nrp _i ；

The method for obtaining the binary compressed data code stream of the current string in the current block and performing inverse binarization to obtain the code containing string length parameter information of the current string specifically comprises the following steps:

R _i ＝cuSize-curTlen

if R is _i =1, then nrp _i The value is set to 0, since 0 is the only possible value, there is no next remaining number nrp in the compressed data stream _i Is a binary symbol string of (2); otherwise, i.e. R _i At > 1, there is a next remaining number nrp in the compressed data stream _i Is inverse binarized to obtain the next remaining number nrp _i 。

The inverse binary valueObtaining the next remaining number nrp _i The method specifically comprises the following steps:

according to the number R of the current residual elements _i Determining the next remaining number nrp _i Value range and obtain its next remaining number nrp _i A scheme in which the value range is divided into a plurality of consecutive sequentially numbered segments; according to a preset next remaining number nrp _i Scheme for dividing value range into a plurality of consecutive sequentially numbered segments, said next remaining number nrp being included _i For marking the next remaining number nrp, including the number of the segment in which the value is located _i Conversion of the parameter vector of values to the next remaining number nrp _i To decode a binarized compressed data stream of a current string in a current block to include the next remaining number nrp according to a preset inverse binarization scheme _i For marking the next remaining number nrp, including the number of the segment in which the value is located _i A parameter vector of values.

The string length obtaining module is at least used for obtaining a string length according to a preset string length calculation mode according to the current block size cuSize, the sum curTlen of the number of elements which are decoded currently and the next residual number _i . Specifically:

for calculating a current string length from the string length parameter _i Calculating the current string length _i And updating the sum curTlen of the lengths of the elements currently decoded; the string length information is the next remaining number nrp _i ；

length _i ＝cuSize-curTlen-nrp _i

updating the sum curTlen of the currently decoded element lengths to be at the current string length _i And the sum of the currently decoded element lengths curTlen updates the sum of the currently decoded element lengths curTlen; the method specifically comprises the following steps:

curTlen＝curTlen+length _i 。

the compressed data code stream parsing module is configured to use a truncated code with a truncated value, i.e. a maximum value maxVal, to perform inverse binarization on the binary symbol string of the next remaining number, where the truncated code includes one or a combination of the following codes:

truncating unary code, or

scheme one:

the second part inverse binarizes the second part of the binary symbol string by selecting one of predetermined I+1 cut-off codes, each of the I+1 cut-off codes being a fixed-length code derived from a maximum value or a cut-off two having a maximum valueA meta code; the maximum value is at least based on the number R of the current residual elements _i And said first portion is derived, the maximum allowed to be 0, indicating that said second portion is empty, i.e., not present;

Scheme II:

Preferably, the next remaining number nrp _i Or the total value range of the next remaining number minus 1 is divided into a predetermined Imax+1 segmentInteger interval [ N0 ] _i ,N0 _i+1 )，i＝0～Imax；

or,

The invention is suitable for encoding and decoding the data in a lossy compression way, is also suitable for encoding and decoding the data in a lossless compression way, and is used for lossless compression when the dividing rule of the strings divides adjacent elements with the same element value into the same string, and is used for lossy compression when the dividing rule divides adjacent elements with similar element value into the same string, the quality of the data in the lossy compression way is lower according to the definition stricter degree of the similar element value, the higher the data quality is, the higher the loose compression rate is, and the worse the data quality is.

The method is applicable to the encoding and decoding of one-dimensional data such as character string data or byte string data or one-dimensional graphics or fractal graphics, and is also applicable to the encoding and decoding of two-dimensional or more data such as images or video data.

In the present invention, the data to which the data compression is applied includes one or a combination of the following types of data: one-dimensional data; two-dimensional data; multidimensional data; a pattern; a split dimension graph; an image; a sequence of images; video; a three-dimensional scene; a sequence of continuously varying three-dimensional scenes; a virtual reality scene; images in the form of sequence pixels of a continuously changing virtual reality scene; transform domain data of the image; a set of two or more bytes; a set of two or more bits; a set of pixels; a set of three-component pixels (Y, U, V); a set of three-component pixels (Y, cb, cr); a set of three-component pixels (Y, cg, co); a set of three-component pixels (R, G, B); a set of four-component pixels (C, M, Y, K); a set of four-component pixels (R, G, B, a); a set of four-component pixels (Y, U, V, a); a set of four-component pixels (Y, cb, cr, a); a set of four-component pixels (Y, cg, co, a).

In the present invention, in the case where the data is an image, a sequence of images, a video, or the like, the encoded block or the decoded block is one encoded region or one decoded region of the image, including at least one of: the method comprises the steps of a whole image, a sub-image of the image, a slice, a tile, a macroblock, a maximum coding unit LCU, a coding tree unit CTU, a coding unit CU, a sub-region of the CU, a sub-coding unit sub-CU, a prediction unit PU, a sub-region of the PU, a sub-prediction unit sub-PU, a transformation unit TU, a sub-region of the TU and a sub-transformation unit sub-TU.

In the present invention, the primitives of the encoded block or the decoded block involved in data compression include one or a combination of the following cases: a coded or decoded block, sub-block, micro-block, string, byte string, alpha (alpha) string, pixel cell string, pixel string, sample string, index string, line, matching block, matching sub-block, matching micro-block, matching string, matching pixel cell string, matching sample string, matching index string, matching strip, matching line, offset string, coordinate string, unpredictable pixel string, incompletely matched string, incompletely predictable string, coordinate or unpredictable pixel string, equivalent string, string where one component of a base vector string is zero, base vector string where the absolute value of a component is 1.

The following are examples:

example 1

A string length parameter encoding method comprising the steps of:

the pixel unit is an array of mxn pixels or pixel components or pixel indices, m and n being predetermined positive integers comprising 1, 2, 4, 8, 16, respectively. When m and n take values of 1, the pixel unit is a pixel or a pixel component or a pixel index or an index.

Repeating the following steps until the pixel unit string or the sequence of the pixel string length (length ₁ 、length ₂ 、......length _n ) And (3) finishing encoding:

(1) Acquiring current block size cuSize and current string length _i And the sum of the lengths of the currently encoded elements curTlen, calculates the next remaining number nrp _i String length parameters as current string;

the next remaining number nrp _i A sum of lengths of all strings after the current string in the current block, that is, a current block size, cuSize minus a sum of lengths of currently encoded elements curTlen and a current string length _i The method is characterized by comprising the following steps:

nrp _i ＝cuSize-curTlen-length _i

the total number of strings in the current block does not exceed 1/4 of the current block size.

(2) Binarizing the current string length parameter obtained in the step (1), and writing the binarized data into a compressed data code stream;

According to the current coding method, elements (pixels or pixel arrays) in one coding block are likely to be divided into the same string, so that the frequency of the next remaining number of 0 is quite high, and the coding is performed with as few bits as possible, specifically as follows:

R _i ＝cuSize-curTlen

if R is _i =1, then nrp _i The value is set to 0, and the next remaining number nrp does not exist in the compressed data code stream _i Is a binary symbol string of (2); otherwise, i.e. R _i > 1, the next remaining number nrp _i Binarization.

The next remaining number nrp of the pairs _i Binarization, specifically includes:

according to the number R of the current residual elements _i Determining the next remaining number nrp _i Value field and the next remaining number nrp _i The value range is divided into a plurality of consecutive sequentially numbered segments as follows:

the next remaining number nrp _i Is divided into a predetermined integer interval of I+1 section [ N ] _i ,N _i+1 )，i＝0～I；

Determination nrp _i A segment number a of the segment preceding the segment where the value of (a) is located, and will nrp _i The expression is as follows: nrp _i ＝N _a +b

And (3) respectively binarizing the a and/or the b, and then writing the binary codes into a code stream, wherein a truncated unary code, a truncated binary code, a truncated k-order exponential Columbus code or a fixed-length code derived from a maximum value can be respectively adopted.

For example: one example of binarizing a and/or b, respectively, is:

and (3) adopting a binarization mode of truncated unary codes for the first part a, obtaining a first partial maximum value maxValPrefix from at least the current residual number according to at least a preset first partial maximum value calculation mode, carrying out binarization on truncated unary codes with the truncated value of synElVal being a, namely with the maximum value maxVal being maxValPrefix, and adopting a table look-up mode or a calculation mode to obtain binary symbol strings. An example of a look-up table is table 1:

TABLE 1 relation of synElVal and binary symbol string (truncated unary code)

If a is equal to 0, then there is no second portion and no remaining portion in the compressed data stream, and the next remaining number nrp is directly taken _i The value of (2) is set to 0; otherwise, namely a is larger than 0, continuing binarization of the segmented truncated binary code of the following second part;

binarization of the segmented truncated binary code is used for the second part b:

at least according to the currentPrevious remaining number, next remaining number nrp _i The value range of (1) is divided into a preset Imax+1 section of integer interval [ Ni, ni+1), i=0-Imax;

determining a value of a segment i=a based at least on the value a of the first portion;

based at least on the value i of the segment, the second partial maximum maxvallnix is obtained by maxvallnix=n _i+1 -N _i –1；

B is binarized using a truncated binary code whose maximum value maxVal is maxvallnix.

Where imax=4, i is 0, 1, 2, 3, or 4, N0 ₀ 、N0 ₁ 、N0 ₂ 、N0 ₃ 、N0 ₄ 、N0 ₅ Respectively 0, 1, 5, 21, 277, the maximum value of the current remaining number R.

curTlen＝curTlen+length _i

example 2

A string length parameter decoding method, corresponding to embodiment 1, comprising the steps of:

repeating the following steps until the pixel unit string or the sequence of the pixel string length (length) ₁ 、length ₂ 、......length _n ) And (3) finishing decoding:

(A) Sequentially obtaining a binarized compressed data code stream of a current string in a current block, and performing inverse binarization to obtain a code containing string length parameter information of the current string; the string length parameter is the next remaining number nrp _i The method comprises the steps of carrying out a first treatment on the surface of the The method specifically comprises the following steps:

R _i ＝cuSize-curTlen

if R is _i =1, then nrp _i The value is set to 0, and the next remaining number nrp does not exist in the compressed data code stream _i Is a binary symbol string of (2); otherwise, i.e. R _i At > 1, there is a next remaining number nrp in the compressed data stream _i Is inverse binarized to obtain the next remaining number nrp _i 。

decoding the binarized compressed data code stream of the current string in the current block into a code stream comprising the next remaining number nrp according to a preset inverse binarization scheme _i For marking the next remaining number nrp, including the number of the segment in which the value is located _i A parameter vector of values; according to a preset next remaining number nrp _i Scheme for dividing value range into a plurality of consecutive sequentially numbered segments, said next remaining number nrp being included _i For marking the next remaining number nrp, including the number of the segment in which the value is located _i Conversion of the parameter vector of values to the next remaining number nrp _i Is a value of (2); the preset next remaining number nrp _i The scheme of dividing the value range into a plurality of consecutive sequentially numbered segments is based on the next remaining number nrp _i Value range determination, the one remaining number nrp _i The value range is based on the number R of the current residual elements _i And (5) determining.

If the number of current remaining pixel units or pixels in the decoded block is referred to as the current remaining number R _i Equal to 1, the next remaining number nrp does not exist in the compressed data stream _i Directly to nrp _i The value of (2) is set to 0; otherwise, i.e. R _i >1, continue to follow the predetermined inverse binarization scheme pair nrp _i Inverse binarization was performed to obtain nrp _i The values of (2) are as follows:

nrp _i the binary symbol string is composed of at least two parts a and b, and is converted and decoded into values of a and b by the compressed code stream according to a preset inverse binarization scheme:

according to the number R of the current residual elements _i Determining the next residueNumber of remainders nrp _i Value field and the next remaining number nrp _i The value range is divided into a plurality of consecutive sequentially numbered segments as follows:

Calculation nrp _i The method comprises the following steps: nrp _i ＝N _a +b。

(B) Calculating the current string length according to the code containing string length parameter information obtained in the step (A) _i The method comprises the steps of carrying out a first treatment on the surface of the The string length information is the next remaining number nrp _i ；

length _i ＝cuSize-curTlen-nrp _i

for example:

nrp _i is composed of at least two parts:

the value a of the first part is obtained from at least the current residual number by adopting an inverse binarization mode of a truncated unary code and at least according to a preset first part maximum value calculation mode, and the value a of the first part is obtained by adopting the inverse binarization mode of the truncated unary code with the truncated value, namely, the maximum value maxVal being maxValPrefix; the inverse binarization method of truncated unary code using truncated value, i.e. maximum value maxVal is maxValPrefix, can adopt a table look-up method or a calculation method. One example of a look-up table approach: obtaining synElVal from table 1 using a truncated value, i.e., maximum value maxVal, as maxValPrefix and binary string symbols, a being equal to synElVal;

If a is equal to 0, then there is no second portion and no remaining portion in the compressed data stream, and the next remaining number nrp is directly taken _i The value of (2) is set to 0; otherwise, namely a is larger than 0, continuing the inverse binarization of the segmented truncated binary code of the following second part;

the value b of the second part is inverse binarized with a segmented truncated binary code:

based at least on the current remaining number, the next remaining number nrp _i The value range of (2) is divided into a preset Imax+1 section of integer interval [ N ] _i ,N _i+1 )，i＝0～Imax；

Determining the value i of the segment based at least on the value a of the first portion in such a way that i=a;

Obtaining a value b of the second part by using a reverse binarization mode of a truncated binary code with a truncated value, namely a maximum value maxVal, of maxValInfinix;

obtaining the next remaining number nrp based at least on the value i of the segment and/or the value b of the second portion _i The way of value of (a) is nrp _i ＝N _i +b。

curTlen＝curTlen+length _i 。

Example 3

A string length parameter encoding method comprising the steps of:

the pixel unit is an array of mxn pixels or pixel components, m and n being predetermined positive integers comprising 1,2, 4, 8, 16, respectively. When m and n take values of 1, the pixel unit is a pixel or a pixel component or a pixel index or an index.

(1) Obtaining the current block size cuSize and the currentString length _i And the sum of the lengths of the currently encoded elements curTlen, calculates the next remaining number nrp _i String length parameters as current string;

nrp _i ＝cuSize-curTlen-length _i

R _i ＝cuSize-curTlen

if R is _i =1, then nrp _i The value is set to 0, and the next remaining number nrp does not exist in the compressed data code stream _i Is a binary symbol string of (2); otherwise, i.e. R _i > 1, there is the next remaining number nrp in the compressed data stream _i Is inverse binarized to obtain the next remaining number nrp _i 。

s1, the next remaining number nrp of the pairs _i Expressed in terms of first, second, and third parts:

nrp _i ＝N _a +(b<<k)+c-(2 ⁿ -maxValInfix-1)×k

wherein a is the value of the first part, b is the value of the second part, maxValInfix is the maximum value of the second part, and n is the length of the fixed-length code of the second part; k is the length of the third part fixed-length code; the method comprises the following steps:

the next remaining number nrp _i The value range of (2) is divided into a preset Imax+1 section of integer interval [ Ni, ni+1 ], i=0-Imax; for example, imax=4, i is 0, 1, 2, 3, or 4, N0 ₀ 、N0 ₁ 、N0 ₂ 、N0 ₃ 、N0 ₄ 、N0 ₅ Respectively 0, 1, 5, 21, 277, the maximum value of the current remaining number R.

Determination nrp _i A segment number a of a segment preceding the segment where the value of (a) is located;

determining a second partial maximum as: maxvallnix=n _a+1 –N _a -1, the length n of the second partial fixed-length code is: n=ceil (Log (maxvallnix+1));

the second part b has the value: when k=0, b= nrp _i -N _a The method comprises the steps of carrying out a first treatment on the surface of the Otherwise, k=1, b= (nrp _i –N _a +1+(1<<n))>>1, wherein>>For the right-shift operation,<<is left shift operation; wherein k is the length of the third partial fixed-length code, and the calculation method is as follows: nrp _i Na is less than2 ⁿ -maxvallnix-1 or maxvallnix equals 0, then the value of k is set to 0; otherwise, the value of k is set to 1.

The third part c has the value: c= nrp _i -Na-(b<<1)+((1<<n)-(maxValInfix+1))＝nrp _i –Na+1-(b<<1)+(1<<n)；

S2, respectively binarizing a and/or b and/or c and then writing the binary codes into a code stream, wherein the binarization of a can adopt a truncated unary code, a truncated binary code, a truncated k-order exponential Columbus code or a fixed-length code derived from a maximum value; and b and c adopt fixed-length codes. The method comprises the following steps:

for the first part: the value a of the first part is binarized by adopting a truncated unary code and then written into a code stream;

for the second part: if the length n of the fixed-length code of the second part is calculated to be equal to or greater than 1, the b value of the second part is binarized by adopting the fixed-length code with the length n and then written into the code stream; otherwise, i.e. n is smaller than 1, the second part binary symbol string is empty, i.e. is not actually present in the compressed data code stream;

for the third part: if k is not equal to 0, the value c of the third part is binarized by adopting a fixed-length code with the length of 1 and then written into the code stream; otherwise, i.e. k is equal to 0, the third partial string of binary symbols is empty, i.e. virtually not present in the compressed data stream.

curTlen＝curTlen+length _i

Example 4

A string length parameter decoding method, corresponding to embodiment 3, comprising the steps of:

(A) Sequentially obtaining binarized compressed data code stream of current string in current block, performing inverse binarization to obtain current stringEncoding containing string length parameter information; the string length parameter is the next remaining number nrp _i The method comprises the steps of carrying out a first treatment on the surface of the The method comprises the following steps:

R _i ＝cuSize-curTlen

if R is _i =1, then nrp _i The value is set to 0, and the next remaining number nrp does not exist in the compressed data code stream _i Is a binary symbol string of (2); otherwise, i.e. R _i > 1, then there is the next remaining number nrp in the compressed data stream _i Is inverse binarized to obtain the next remaining number nrp _i 。

If the number of current remaining pixel units or pixels in the decoded block is referred to as the current remaining number R _i Equal to 1, the next remaining number nrp does not exist in the compressed data stream _i Directly to nrp _i The value of (2) is set to 0; otherwise, i.e. R _i >1, continue to follow the predetermined inverse binarization scheme pair nrp _i Performing inverse binarization:

nrp _i inverse binarization consists of at least three parts:

for the value of the first part a: from at least the current remaining number R, at least in a predetermined first partial maximum calculation, using inverse binarization of truncated unary codes _i Obtaining a first partial maximum value maxValPrefix, and obtaining a value a of the first partial by using a truncated unary code inverse binarization mode of which the maximum value maxVal is maxValPrefix; if a is equal to 0, the second portion, the third portion and the remaining portion are not present in the compressed data stream, and the next remaining number nrp is directly taken _i The value of (2) is set to 0; otherwise, namely a is larger than 0, continuing the inverse binarization of the following second part and third part;

for the value b of the second part: according to the current residual number R _i The next remaining number nrp _i The value range of (2) is divided into a preset Imax+1 section of integer interval [ N ] _i ,N _i+1 ) I=0 to Imax, and the second partial maximum maxValInfix, maxValInfix =n is determined _a+1 –N _a -1 and deriving the length n of the second partial fixed-length code, n=ceil (Log (maxvallnix+1)); if n is equal to or greater than 1, the value b of the second part is obtained by using the inverse binarization mode of the fixed-length code with the length of n, otherwise, namely, n is less than 1, the binary symbol string is empty and is not actually existed in the compressed data code stream, so that the value b of the second part is directly set to 0 without analyzing the compressed data code stream; for example, where imax=4, i is 0, 1, 2, 3, or 4, N0 ₀ 、N0 ₁ 、N0 ₂ 、N0 ₃ 、N0 ₄ 、N0 ₅ Respectively 0, 1, 5, 21, 277, the maximum value of the current remaining number R.

For the third part: obtaining the length k of the third part fixed-length code according to the value b of the second part, if b is smaller than 2 ⁿ -maxvallnix-1 or maxvallnix is equal to 0, the value of k is set to 0, otherwise the value of k is set to 1; if k is equal to or greater than 1, the length is usedThe value c of the third part is obtained in an inverse binarization mode of the fixed-length code of k, otherwise, namely k is smaller than 1, the binary symbol string is empty, namely the binary symbol string is not actually present in the compressed data code stream, so that the value c of the third part is directly set to 0 without analyzing the compressed data code stream;

obtaining the next remaining number nrp based on the value a of the segment and the value b of the second portion, the length n of the second portion fixed-length code, the value c of the third portion, and the length k of the third portion fixed-length code _i The values of (2) are as follows:

nrp _i ＝Na+(b<<k)+c-(2 ⁿ -maxValInfix-1)×k

length _i ＝cuSize-curTlen-nrp _i

(C) Length in current string length _i And the sum of the currently decoded element lengths curTlen updates the sum of the currently decoded element lengths curTlen; the method comprises the following steps:

curTlen＝curTlen+length _i 。

it will be readily appreciated by those skilled in the art that the foregoing description is merely a preferred embodiment of the invention and is not intended to limit the invention, but any modifications, equivalents, improvements or alternatives falling within the spirit and principles of the invention are intended to be included within the scope of the invention.

Claims

1. A method for decoding string length parameters, comprising at least the steps of:

(B) Obtaining a current string length according to a preset string length calculation mode at least according to the current block size curize, the sum curTlen of the number of elements which are decoded currently and the next residual number _i ；

The current string length is obtained according to a preset string length calculation mode _i The method specifically comprises the following steps:

current string length _i The method is as follows:

length _i ＝R _i -nrp _i

wherein R is _i The number of current remaining elements; the number R of the current residual elements _i For the difference between the current block size curize and the sum curTlen of the lengths of the currently decoded elements, calculating the number R of the current remaining elements in the current block _i The following are provided:

R _i ＝cuSize-curTlen；

when R is _i When=1, then nrp _i The value is set to 0, and the next remaining number nrp does not exist in the compressed data code stream _i Is a binary symbol string of (2); otherwise, i.e. R _i > 1, there is the next remaining number nrp in the compressed data stream _i Is inverse binarized the next remaining number nrp _i The next remaining number nrp obtained for the binary symbol string of (2) _i ；

The inverse binarizing the next remaining number nrp _i The next remaining number nrp obtained for the binary symbol string of (2) _i The method specifically comprises the following steps:

nrp the next remaining number in the compressed data stream of the current string in the current block _i Is decoded to include the next remaining number nrp according to a preset inverse binarization scheme _i For marking the next remaining number nrp, including the number of the segment in which the value is located _i A parameter vector of values; according to a preset next remaining number nrp _i The value range being divided into a plurality of successive values A scheme of sequentially numbered segments including the next remaining number nrp _i For marking the next remaining number nrp, including the number of the segment in which the value is located _i Conversion of the parameter vector of values to the next remaining number nrp _i Is a value of (2).

2. The string length parameter decoding method according to claim 1, wherein the value range of the next remaining number is [0, r _i -1]。

3. The string length parameter decoding method of claim 1, wherein the element comprises a pixel or a pixel component simply component or a pixel index simply index or a pixel unit; the pixel unit is an array of mxn pixels or components or indexes, m and n being predetermined positive integers including 1, 2, 4, 8, 16, respectively; when m and n both take a value of 1, the pixel unit is a pixel or a pixel component or a pixel index or an index.

4. The string length parameter decoding method as claimed in claim 1, wherein the preset next remaining number nrp _i The scheme of dividing the value range into a plurality of consecutive sequentially numbered segments is based on the next remaining number nrp _i And (5) determining a value range.

5. The string length parameter decoding method according to claim 4, wherein the next remaining number nrp _i The value range is based on the number R of the current residual elements _i And (5) determining.

6. The string length parameter decoding method of claim 1, wherein the inverse binarizing next remaining number nrp _i Is composed of at least two parts:

7. The string length parameter decoding method of claim 1, wherein the inverse binarizing next remaining number nrp _i Is composed of at least three parts:

8. The string length parameter decoding method of claim 7, wherein for the value a of the first portion: from at least the current remaining number R, at least in a predetermined first partial maximum calculation, using inverse binarization of truncated unary codes _i Obtaining a first partial maximum value maxValPrefix, and obtaining a value a of the first partial by using a truncated unary code inverse binarization mode of which the maximum value maxVal is maxValPrefix; if a is equal to 0, the second portion, the third portion and the remaining portion are not present in the compressed data stream, and the next remaining number nrp is directly taken _i The value of (2) is set to 0; otherwise, namely a is larger than 0, continuing the inverse binarization of the following second part and third part;

for the value b of the second part: according to the current residual number R _i Will leave nextNumber of remainders nrp _i The value range of (2) is divided into a preset Imax+1 section integer interval [ N0 ] _i ,N0 _i+1 ) I=0 to Imax, and the second partial maximum value maxValInfix, maxValInfix =n0 is determined _a+1 –N0 _a -1 and deriving the length n of the second partial fixed-length code, n=ceil (Log (maxvallnix+1)); if n is equal to or greater than 1, the value b of the second part is obtained by using the inverse binarization mode of the fixed-length code with the length of n, otherwise, namely, n is less than 1, the binary symbol string is empty and is not actually existed in the compressed data code stream, so that the value b of the second part is directly set to 0 without analyzing the compressed data code stream;

for the value c of the third part: obtaining the length k of the third part fixed-length code according to the value b of the second part, if b is smaller than 2 ⁿ -maxvallnix-1 or maxvallnix is equal to 0, the value of k is set to 0, otherwise the value of k is set to 1; if k is equal to or greater than 1, the value c of the third part is obtained by using the inverse binarization mode of the fixed-length code with the length of k, otherwise, if k is less than 1, the binary symbol string is empty and is not actually present in the compressed data code stream, so that the value c of the third part is directly set to 0 without analyzing the compressed data code stream;

nrp _i ＝N0 _a +(b<<k)+c-(2 ⁿ -maxValInfix-1)×k。

9. the string length parameter decoding method of claim 1, wherein the inverse binarizing next remaining number nrp _i The next remaining number nrp obtained for the binary symbol string of (2) _i The method specifically comprises the following steps:

truncating unary code, or

The exponential golomb code of the predetermined order is truncated, i.e., the k-order exponential golomb code is truncated.

10. The string length parameter decoding method of claim 9, wherein the number R of current remaining elements is at least based on _i The maximum maxVal is determined.

11. The string length parameter decoding method of claim 10, wherein the maximum value maxVal is allowed to be 0, indicating that the corresponding code is a null code, i.e., a binary symbol string does not actually exist.

12. The string length parameter decoding method of claim 10, wherein the element is a pixel unit consisting of 4x1 pixels or components or indices.

13. The string length parameter decoding method of claim 12, wherein the inverse binarizing the next remaining number nrp _i The next remaining number nrp obtained for the binary symbol string of (2) _i One of the following two schemes is applied;

scheme one:

selecting one of predetermined i+1 truncated codes for the second portion of the binary symbol string, each of the i+1 truncated codes being a fixed-length code derived from a maximum value or a truncated binary code having a maximum value, to inverse binarize the second portion of the binary symbol string; the maximum value is at least based on the currentNumber of remaining elements R _i And said first portion is derived, the maximum allowed to be 0, indicating that said second portion is empty, i.e., not present;

scheme II:

14The string length parameter decoding method of claim 13, wherein the next remaining number nrp _i Or the total value range of the next remaining number minus 1 is divided into a predetermined Imax+1 integer interval [ N0 ] _i ,N0 _i+1 )，i＝0～Imax；

15. The string length parameter decoding method according to claim 14, wherein in scheme one, imax=4, i is 0, 1, 2, 3, or 4, N0 ₀ 、N0 ₁ 、N0 ₂ 、N0 ₃ 、N0 ₄ 、N0 ₅ R is the maximum value of 0, 1, 5, 21, 277, R _i Is the maximum value of (2);

or,

16. The string length parameter decoding method according to claim 14, wherein in scheme one, imax= 3,i is 0, 1, 2, or 3, N0 ₀ 、N0 ₁ 、N0 ₂ 、N0 ₃ 、N0 ₄ R is the maximum value of 0, 1, 5, 21, R _i Is the maximum value of (2);

or,

17. A string length parameter decoding apparatus, comprising: the device comprises a compressed data code stream analysis module and a string length calculation module;

The string length calculation module is used for obtaining a string length according to a preset string length calculation mode at least according to the current block size cuSize, the sum curTlen of the number of elements which are decoded currently and the next residual number _i The method comprises the steps of carrying out a first treatment on the surface of the And is used for length according to the current string length _i And updating the sum curTlen of the lengths of the elements currently decoded; the string length information is the next remaining number nrp _i ；

The string length is obtained according to a preset string length calculation mode _i The sum curTlen of the currently decoded element length subtracted from the current block size cuSize and the next remaining number nrp _i The method is characterized by comprising the following steps:

length _i ＝R _i -nrp _i

wherein R is _i The number of current remaining elements;

the information for obtaining the string length parameter value representing the current string is specifically:

the number Ri of current remaining elements is calculated from the current block size cuSize and the sum curtlin of the lengths of the elements currently decoded,

R _i ＝cuSize-curTlen；

when R is _i When=1, then nrp _i The value is set to 0, and the next remaining number nrp does not exist in the compressed data code stream _i Is a binary symbol string of (2); otherwise, i.e. R _i > 1, then there is the next remaining number nrp in the compressed data stream _i Is inverse binarized to obtain the next remaining number nrp _i ；

According to the number R of the current residual elements _i Determining the next remaining number nrp _i Value range and obtain its next remaining number nrp _i A scheme in which the value range is divided into a plurality of consecutive sequentially numbered segments; according to a preset next remaining number nrp _i Scheme of value range division into a plurality of consecutive sequentially numbered segments will include the next remaining number nrp _i For marking the next remaining number nrp, including the number of the segment in which the value is located _i Conversion of the parameter vector of values to the next remaining number nrp _i To decode a binarized compressed data stream of a current string in a current block to include the next remaining number nrp according to a preset inverse binarization scheme _i For marking the next remaining number nrp, including the number of the segment in which the value is located _i A parameter vector of values.

18. The string length parameter decoding device according to claim 17, wherein the value range of the next remaining number is [0, r _i -1]。

19. The string length parameter decoding device of claim 17, wherein the element comprises a pixel or a pixel component simply component or a pixel index simply index or a pixel unit; the pixel unit is an array of mxn pixels or components or indexes, m and n being predetermined positive integers including 1, 2, 4, 8, 16, respectively; when m and n are both 1, the pixel unit is a pixel or a pixel component or a pixel index or an index; an element is a unit of pixel consisting of 4x1 pixels or components or indices.

20. The string length parameter decoding apparatus according to claim 17, wherein the string length calculation module updates a sum curtlien of lengths of elements currently decoded to be a current string length _i And the sum of the lengths of the elements currently decoded curTlen, in particular:

curTlen＝curTlen+length _i 。

21. the string length parameter decoding device according to claim 17, wherein the compressed data stream parsing module is configured to use a truncated code having a truncated value, i.e., a maximum maxVal, to inverse binarize the next remaining number of binary symbol strings, the truncated code including one or a combination of the following codes:

truncating unary code, or

22. The string length parameter decoding apparatus as claimed in claim 21, wherein the number R of at least current remaining elements is based on _i Determining the maximum value maxVal; the maximum value maxVal is allowed to be 0, indicating that the corresponding code is a null code, i.e., a binary symbol string is not actually present.

23. The string length parameter decoding device according to claim 22, wherein the next remaining number of binary symbol strings and their inverse binarization operations are composed of at least three parts;

scheme one:

scheme II:

24. The string length parameter decoding apparatus according to claim 23, wherein the next remaining number nrp _i Or the total value range of the next remaining number minus 1 is divided into a predetermined Imax+1 integer interval [ N0 ] _i ,N0 _i+1 )，i＝0～Imax；

25. The string length parameter decoding apparatus according to claim 24, wherein in scheme one, imax=4, i is 0, 1, 2, 3, or 4, N0 ₀ 、N0 ₁ 、N0 ₂ 、N0 ₃ 、N0 ₄ 、N0 ₅ R is the maximum value of 0, 1, 5, 21, 277, R _i Is the maximum value of (2);

or,

26. The string length parameter decoding apparatus according to claim 24, wherein in scheme one, imax= 3,i is 0, 1, 2, or 3, N0 ₀ 、N0 ₁ 、N0 ₂ 、N0 ₃ 、N0 ₄ R is the maximum value of 0, 1, 5, 21, R _i Is the maximum value of (2);

or,