CN107483931A - A kind of HEVC intraframe codings depth division high-speed decision method based on significance - Google Patents
A kind of HEVC intraframe codings depth division high-speed decision method based on significance Download PDFInfo
- Publication number
- CN107483931A CN107483931A CN201710646984.4A CN201710646984A CN107483931A CN 107483931 A CN107483931 A CN 107483931A CN 201710646984 A CN201710646984 A CN 201710646984A CN 107483931 A CN107483931 A CN 107483931A
- Authority
- CN
- China
- Prior art keywords
- mrow
- lcu
- msub
- frame
- significance
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
- H04N19/146—Data rate or code amount at the encoder output
- H04N19/147—Data rate or code amount at the encoder output according to rate distortion criteria
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/103—Selection of coding mode or of prediction mode
- H04N19/11—Selection of coding mode or of prediction mode among a plurality of spatial predictive coding modes
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
- H04N19/136—Incoming video signal characteristics or properties
- H04N19/14—Coding unit complexity, e.g. amount of activity or edge presence estimation
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/182—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being a pixel
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
- H04N19/593—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving spatial prediction techniques
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The invention discloses a kind of HEVC intraframe codings depth based on significance to divide high-speed decision method, the notable angle value of current encoded frame and present frame in units of pixel is read in first, calculate the nonoverlapping notable angle value using 4 × 4 block of pixels as unit of present frame, the notable angle value of all maximum coding units of present frame (LCU) is calculated again and is normalized, then the notable angle value after present frame is normalized by LCU layers according to ratio is divided into low significance, three regions of middle significance and high significance, and the significance region according to belonging to LCU sets depth different during HEVC intraframe codings to divide scope for it, HEVC intraframe codings finally are carried out according to the depth bounds of setting to each LCU in present frame, it is optimal intraframe coding depth to select the minimum depth of rate distortion costs.The present invention by significance information to reduce HEVC intraframe codings when LCU optimal depth seeking scope, computation complexity and the scramble time of HEVC intraframe codings, and the loss very little of code efficiency can be effectively reduced.
Description
Technical field
The invention belongs to the fast coding technical field in HEVC Video codings, is related to a kind of HEVC frames based on significance
Interior coding depth divides high-speed decision method.
Background technology
Efficient video coding (High Efficiency Video Coding, HEVC) is that International video newest at present is compiled
Code standard, by using in the frame of flexible quad-tree partition structure and various modes with the technology such as inter prediction, greatly
Code efficiency is improved, but the computation complexity encoded also sharply increases, and greatly limit HEVC encoders in reality
Application in border.
Whole two field picture is divided into nonoverlapping maximum coding unit (Largest by especially HEVC first in intraframe coding
Coding Unit, LCU), LCU is the block of pixels of 64 × 64 sizes, travels through the four of different depth level in units of LCU afterwards
Fork tree partition structure, the interior division depth bounds of LCU frames is 0 to 3, and the block corresponded to respectively from 64 × 64 to 8 × 8 divides size, leads to
Cross calculate different depth partition structure rate distortion costs come determine optimal intra-frame encoding mode and division depth, cause in frame
Coding has high computation complexity.It is therefore desirable to study the high-speed decision for HEVC intraframe coding optimal dividing depth
Method, reduce the computation complexity of intraframe coding on the premise of code efficiency and subjective and objective video quality is not reduced as far as possible.
Existing HEVC intraframe codings fast algorithm is mostly based on the spatial coherence in the temporal correlation or frame of interframe,
Using the adjacent encoded LCU of the current LCU of previous frame same position LCU or same frames depth division and the system such as rate distortion costs
To predict, division depth, this kind of method reduce encoder complexity, but algorithm to meter information to a certain extent in current LCU frame
Itself need to store more statistical information, and hold due to not accounting for influence of the human visual system (HVS) to coding quality
Easily cause the decline of video quality.This is due to that HVS perceives to the quality of encoded video and degree of concern is not uniformly to be distributed
On whole two field picture, and often concentrate the high significance region for causing human eye to be paid close attention in the picture.
The content of the invention
It is an object of the present invention to provide a kind of HEVC intraframe codings depth based on significance to divide high-speed decision method, the party
Method can effectively reduce the optimal depth seeking scope in HEVC intraframe codings, and then reduce the computation complexity of intraframe coding,
Higher code efficiency and preferable video quality are kept simultaneously, and moving towards practical for HEVC encoders provides technical support.
In order to achieve the above object, the technical solution adopted by the present invention is deep for a kind of HEVC intraframe codings based on significance
Degree division high-speed decision method, specifically comprises the following steps:
Step 1:Current video frame images I (i, j) to be encoded is read in, (i, j) is pixel ranks coordinate position, and value is
I=[1,2 ..., N] and j=[1,2 ..., M], M and N are width and height of the image in units of pixel respectively;
Step 2:Read in current notable angle value S (i, j) of the frame to be encoded in units of pixel and obtain accurate pixel layer
Notable angle value;
Step 3:It is nonoverlapping with 4 × 4 pixels that the notable angle value S (i, j) of pixel layer obtained according to step 2 calculates present frame
Block is the notable angle value of unit, and makees normalized, first, is calculated according to formula (1) nonoverlapping with 4 × 4 pixels in frame
Block is the notable angle value S ' of unit4×4(k, l), (k, l) are 4 × 4 pieces of ranks coordinate, and value is k=[1,2 ..., K], l=
[1,2 ..., L], wherein K=N/4, L=M/4, i.e., whole frame have K × L=(M × N)/16 using 4 × 4 block of pixels as the aobvious of unit
Write angle value,
Afterwards, by present frame with 4 × 4 pieces of notable angle value S ' for unit4×4(k, l) is normalized between [0,1], is obtained
S4×4(k,l);
Step 4:Notable angle value S after each 4 × 4 pieces of normalization obtained according to step 34×4(k, l) calculates present encoding
All maximum coding unit LCU notable angle value in frame, and make normalized, first, institute in frame is calculated according to formula (2)
There is LCU notable angle value S 'LCU(p, q), (p, q) are LCU ranks coordinates, and LCU is that size is 64 × 64 block of pixels, each LCU
Contain the block of pixels of 16 × 16 4 × 4 sizes, therefore p=[1,2 ..., P], q=[1,2 ..., Q], wherein P=K/16, Q
=L/16, i.e., whole frame have P × Q=(K × L)/(16 × 16)=(M × N)/(64 × 64) individual using 64 × 64 block of pixels as unit
LCU notable angle value,
Afterwards, by all LCU of present frame notable angle value S 'LCU(p, q) is normalized between [0,1], obtains SLCU(p,
q);
Step 5:Notable angle value after current frame to be encoded is normalized by LCU layers according to ratio is divided into three regions,
I.e. low significance region SL, middle significance region SMWith high significance region SH;
Step 6:The significance region belonging to present frame each LCU obtained according to step 5, is set for each LCU
Different depth division scope during HEVC intraframe predictive codings;
Step 7:Depth bounds is divided in current encoded frame each the LCU frame set according to step 6, using HEVC standard
Method for choosing frame inner forecast mode all LCU in frame are encoded by order from left to right, from top to bottom, first, lead to
Crossing rate distortion costs selects LCU to divide the optimal intra prediction mode under depth, and then percent of pass distortion in each frame
Cost divides in the frame of setting to be selected to divide depth in optimal frame in depth bounds, thus obtains the optimal of each LCU
Partition mode in frame;
Step 8:In current encoded frame after the completion of all LCU codings, then present frame coding is completed, and reads in next frame afterwards
Image, each LCU of 1 to 7 pair of repeat step divided in frame setting and the coding of depth bounds.
Further, current frame to be encoded is divided by the notable angle value after the normalization of LCU layers according to ratio in above-mentioned steps 5
Specifically included for trizonal method:
A) sum for remembering current all LCU of frame to be encoded is NLCU=P × Q, (p, q) individual LCU are LCU (p, q);
B) by the notable angle value S after current all LCU normalization of frame to be encodedLCU(p, q) according to it is incremental i.e. from small to large
Order sorts;
C) N is definedLCUTwo-dimensional array S [the N that row 2 arrangesLCU] [2], first row storage label [1,2 ..., NLCU], secondary series is deposited
Store up the notable angle value after LCU layers sort from small to large;
D) the low significance threshold value of LCU layers is takenHigh significance threshold value WhereinExpression rounds downwards, i.e. threshold value SaWith SbIt is to be determined according to the quantitative proportion of whole frame LCU significances, rather than
Determined according to absolute significantly angle value;
E) magnitude relationship of notable angle value and two threshold values after being normalized according to each LCU, will have using formula (3)
The LCU of different significantly angle value is divided into low significance region SL, middle significance region SMWith high significance region SH:
Wherein ALCU(p, q) represents the significance region that LCU (p, q) is divided into.
Further, each LCU that is divided into above-mentioned steps 6 according to present frame significance region is set in HEVC frames in advance
The method for surveying depth division scope different during coding is as follows:
A) when HEVC carries out intra mode decision in units of LCU, depth { 0,1,2,3 } is divided totally in 4 frames, consideration
Depth is more, then what is divided in LCU frames is finer, and code efficiency is higher, but computation complexity is also higher, if can be in LCU frames
During division, its optimal division depth is predicted, then can reduce computation complexity while code efficiency is kept;First, it is right
In low significance region SLIn LCU, the region that its quantity accounts for the 10% of whole frame LCU and less paid close attention in human eye, set
In low significance region SLIn LCU frame in division depth bounds be { 0,1,2 };
B) similarly, in middle significance region SMLCU, its quantity accounts for the 70% of whole frame LCU and compares in human eye
The region of concern, setting are in middle significance region SMLCU frame in division depth bounds be { 1,2,3 };
C) in high significance region SHLCU, its quantity accounts for the 20% of whole frame LCU and paid special attention in human eye
Region, setting are in high significance region SHLCU frame in division depth bounds be { 2,3 };
Division depth bounds sets and can be expressed as follows with formula (4) in the frame of LCU in different significance regions:
Wherein DRLCU(p, q) represents that the depth of division in LCU (p, q) frame sets scope.
It should be noted that the specific method that the notable angle value of current frame pixel layer is calculated in step 2 is more accurate that can obtain
The notable angle value of pixel layer be condition.
Beneficial effect:
1st, HEVC intraframe coding depth division high-speed decision method of the present invention based on significance, passes through the notable of LCU layers
Degree expresses HVS to whole frame coding quality and the difference of degree of concern, and there is provided the division of different LCU intraframe codings depth accordingly
Scope, effectively reduce computation complexity and the scramble time of HEVC intraframe codings.
2nd, the present invention divides scope based on significance to reduce depth in frame, it is contemplated that perception of the HVS to video quality is special
Property, maintain higher code efficiency and preferable video quality.
3rd, the present invention sets depth in frame to divide scope based on significance, it is not necessary to encoded LCU system before storage
Information is counted, therefore there is relatively low computation complexity and storage demand.
Brief description of the drawings
Fig. 1 is the overall flow figure of the HEVC intraframe coding depth high-speed decision methods based on significance of the present invention.
Present frame is divided into trizonal detail flowchart by Fig. 2 for the present invention's by the notable angle value of LCU layers.
Embodiment
The invention is described in further detail with reference to Figure of description.
As depicted in figs. 1 and 2, the HEVC intraframe coding depth high-speed decision methods of the invention based on significance, have altogether
Body embodiment is as follows:
1) current video frame images I (i, j) to be encoded is read in, (i, j) is pixel ranks coordinate position, value i=
[1,2 ..., N] and j=[1,2 ..., M], M and N are width and height of the image in units of pixel.
2) current notable angle value S (i, j) of the frame to be encoded in units of pixel is read in.
3) it is nonoverlapping with 4 × 4 block of pixels that the notable angle value S (i, j) of pixel layer obtained according to step 2) calculates present frame
For the notable angle value of unit, and make normalized, first, calculated according to formula (1) nonoverlapping with 4 × 4 block of pixels in frame
For the notable angle value S ' of unit4×4(k, l), (k, l) are 4 × 4 pieces of ranks coordinate, and value is k=[1,2 ..., K], l=[1,
2 ..., L], wherein K=N/4, L=M/4, i.e., whole frame have K × L=(M × N)/16 significance using 4 × 4 block of pixels as unit
Value,
Afterwards, by present frame with 4 × 4 pieces of notable angle value S ' for unit4×4(k, l) is normalized between [0,1], is obtained
S4×4(k,l)。
4) the notable angle value S after each 4 × 4 pieces obtained according to step 3) normalize4×4(k, l) calculates current encoded frame
In all maximum coding unit (LCU) notable angle value, and make normalized, first, institute in frame calculated according to formula (2)
There is LCU notable angle value S 'LCU(p, q), (p, q) are LCU ranks coordinates, and LCU is the basic coding unit in HEVC, is big
Small is 64 × 64 block of pixels, and each LCU contains the block of pixels of 16 × 16 4 × 4 sizes, therefore p=[1,2 ..., P], q=
[1,2 ..., Q], wherein P=K/16, Q=L/16, i.e., whole frame have P × Q=(K × L)/(16 × 16)=(M × N)/(64 × 64)
The individual notable angle value using 64 × 64 block of pixels as the LCU of unit,
Afterwards, by all LCU of present frame notable angle value S 'LCU(p, q) is normalized between [0,1], obtains SLCU(p,
q);
5) the notable angle value after current frame to be encoded is normalized by LCU layers according to ratio is divided into three regions, i.e., low
Significance region SL, middle significance region SMWith high significance region SH.Including following sub-step:
A) sum for remembering current all LCU of frame to be encoded is NLCU=P × Q, (p, q) individual LCU are LCU (p, q);
B) by the notable angle value S after current all LCU normalization of frame to be encodedLCU(p, q) according to it is incremental i.e. from small to large
Order sorts;
C) N is definedLCUTwo-dimensional array S [the N that row 2 arrangesLCU] [2], first row storage label [1,2 ..., NLCU], secondary series is deposited
Store up the notable angle value after LCU layers sort from small to large;
D) the low significance threshold value of LCU layers is takenHigh significance threshold value WhereinExpression rounds downwards, i.e. threshold value SaWith SbIt is to be determined according to the quantitative proportion of whole frame LCU significances, without
It is according to definitely significantly angle value determines;
E) magnitude relationship of notable angle value and two threshold values after being normalized according to each LCU, will have using formula (3)
The LCU of different significantly angle value is divided into low significance region SL, middle significance region SMWith high significance region SH;
Wherein ALCU(p, q) represents the significance region that LCU (p, q) is divided into.
6) the significance region belonging to present frame each LCU obtained according to step 5), HEVC is set for each LCU
Different depth division scope during intraframe predictive coding.Including following sub-step:
A) first, in low significance region SLIn LCU, its quantity account for the 10% of whole frame LCU and in human eye not
The region paid close attention to very much, by being counted to multiple video sequence codings, it is found that division depth exists in this fraction LCU actual frame
Probability in the range of { 0,1,2 } is larger, therefore sets and be in low significance region SLIn LCU frame in division depth bounds be
{0,1,2};
B) similarly, in middle significance region SMLCU, its quantity accounts for the 70% of whole frame LCU and compares in human eye
The region of concern, find to divide probability of the depth in the range of { 1,2,3 } in this most of LCU actual frame by encoding statistics
Close to 1, and depth is almost 0 for the probability of { 0 }, therefore sets and be in middle significance region SMLCU frame in division it is deep
It is { 1,2,3 } to spend scope;
C) similarly, in high significance region SHLCU, its quantity account for the 20% of whole frame LCU and in human eye very
The region of concern, find that probability of the depth in the range of { 2,3 } is divided in this part LCU actual frame to be exceeded by encoding statistics
0.9, therefore set and be in high significance region SHLCU frame in division depth bounds be { 2,3 };
Division depth bounds sets and can be expressed as follows with formula (4) in the frame of LCU in different significance regions:
Wherein DRLCU(p, q) represents that the depth of division in LCU (p, q) frame sets scope.
7) division depth bounds in current encoded frame each the LCU frame set according to step 6), using HEVC standard
Method for choosing frame inner forecast mode is encoded to all LCU in frame by order from left to right, from top to bottom, first, is passed through
Rate distortion costs selection LCU divides the optimal intra prediction mode under depth, and then percent of pass distortion generation in each frame
Valency divides in the frame of setting to be selected to divide depth in optimal frame in depth bounds, thus obtains each LCU optimal frames
Interior partition mode.
8) in current encoded frame after the completion of all LCU codings, then present frame coding is completed, and reads in next frame figure afterwards
Picture, continue setting and the coding for beginning to use this method to carry out dividing depth bounds to each LCU in frame from step 1).
Claims (4)
- A kind of 1. HEVC intraframe codings depth division high-speed decision method based on significance, it is characterised in that methods described bag Include following steps:Step 1:Read in current video frame images I (i, j) to be encoded, (i, j) is pixel ranks coordinate position, value i= [1,2 ..., N] and j=[1,2 ..., M], M and N are width and height of the image in units of pixel respectively;Step 2:Read in current notable angle value S (i, j) of the frame to be encoded in units of pixel and obtain the aobvious of accurate pixel layer Write angle value;Step 3:The notable angle value S (i, j) of the pixel layer that is obtained according to step 2 calculate present frame it is nonoverlapping using 4 × 4 block of pixels as The notable angle value of unit, and make normalized, first, according to formula (1) calculate in frame it is nonoverlapping using 4 × 4 block of pixels as The notable angle value S ' of unit4×4(k, l), (k, l) are 4 × 4 pieces of ranks coordinate, and value is k=[1,2 ..., K], l=[1, 2 ..., L], wherein K=N/4, L=M/4, i.e., whole frame have K × L=(M × N)/16 significance using 4 × 4 block of pixels as unit Value,<mrow> <msub> <msup> <mi>S</mi> <mo>&prime;</mo> </msup> <mrow> <mn>4</mn> <mo>&times;</mo> <mn>4</mn> </mrow> </msub> <mrow> <mo>(</mo> <mi>k</mi> <mo>,</mo> <mi>l</mi> <mo>)</mo> </mrow> <mo>=</mo> <mfrac> <mn>1</mn> <mn>16</mn> </mfrac> <msubsup> <mi>&Sigma;</mi> <mrow> <mi>i</mi> <mo>=</mo> <mi>k</mi> <mo>&times;</mo> <mn>4</mn> <mo>+</mo> <mn>1</mn> </mrow> <mrow> <mi>k</mi> <mo>&times;</mo> <mn>4</mn> <mo>+</mo> <mn>4</mn> </mrow> </msubsup> <msubsup> <mi>&Sigma;</mi> <mrow> <mi>j</mi> <mo>=</mo> <mi>l</mi> <mo>&times;</mo> <mn>4</mn> <mo>+</mo> <mn>1</mn> </mrow> <mrow> <mi>l</mi> <mo>&times;</mo> <mn>4</mn> <mo>+</mo> <mn>4</mn> </mrow> </msubsup> <mi>S</mi> <mrow> <mo>(</mo> <mi>i</mi> <mo>,</mo> <mi>j</mi> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>1</mn> <mo>)</mo> </mrow> </mrow>Afterwards, by present frame with 4 × 4 pieces of notable angle value S ' for unit4×4(k, l) is normalized between [0,1], obtains S4×4 (k,l);Step 4:Notable angle value S after each 4 × 4 pieces of normalization obtained according to step 34×4(k, l) is calculated in current encoded frame All maximum coding unit LCU notable angle value, and make normalized, first, calculated in frame and owned according to formula (2) LCU notable angle value S 'LCU(p, q), (p, q) are LCU ranks coordinates, and LCU is that size is 64 × 64 block of pixels, each LCU bags The block of pixels of 16 × 16 4 × 4 sizes, therefore p=[1,2 ..., P], q=[1,2 ..., Q], wherein P=K/16, Q=are contained L/16, i.e., whole frame have P × Q=(K × L)/(16 × 16)=(M × N)/(64 × 64) individual using 64 × 64 block of pixels as unit LCU notable angle value,<mrow> <msub> <msup> <mi>S</mi> <mo>&prime;</mo> </msup> <mrow> <mi>L</mi> <mi>C</mi> <mi>U</mi> </mrow> </msub> <mrow> <mo>(</mo> <mi>p</mi> <mo>,</mo> <mi>q</mi> <mo>)</mo> </mrow> <mo>=</mo> <mfrac> <mn>1</mn> <mn>256</mn> </mfrac> <msubsup> <mi>&Sigma;</mi> <mrow> <mi>k</mi> <mo>=</mo> <mi>p</mi> <mo>&times;</mo> <mn>16</mn> <mo>+</mo> <mn>1</mn> </mrow> <mrow> <mi>p</mi> <mo>&times;</mo> <mn>16</mn> <mo>+</mo> <mn>16</mn> </mrow> </msubsup> <msubsup> <mi>&Sigma;</mi> <mrow> <mi>l</mi> <mo>=</mo> <mi>q</mi> <mo>&times;</mo> <mn>16</mn> <mo>+</mo> <mn>1</mn> </mrow> <mrow> <mi>q</mi> <mo>&times;</mo> <mn>16</mn> <mo>+</mo> <mn>16</mn> </mrow> </msubsup> <msub> <mi>S</mi> <mrow> <mn>4</mn> <mo>&times;</mo> <mn>4</mn> </mrow> </msub> <mrow> <mo>(</mo> <mi>k</mi> <mo>,</mo> <mi>l</mi> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> </mrow>Afterwards, by all LCU of present frame notable angle value S 'LCU(p, q) is normalized between [0,1], obtains SLCU(p,q);Step 5:Notable angle value after current frame to be encoded is normalized by LCU layers according to ratio is divided into three regions, i.e., low Significance region SL, middle significance region SMWith high significance region SH;Step 6:The significance region belonging to present frame each LCU obtained according to step 5, HEVC is set for each LCU Different depth division scope during intraframe predictive coding;Step 7:Depth bounds is divided in current encoded frame each the LCU frame set according to step 6, using the frame of HEVC standard Inner estimation mode system of selection is encoded to all LCU in frame by order from left to right, from top to bottom, first, percent of pass Distortion cost selection LCU divides the optimal intra prediction mode under depth in each frame, and then passes through rate distortion costs Divided in the frame of setting and select to divide depth in optimal frame in depth bounds, in the optimal frames for thus obtaining each LCU Partition mode;Step 8:In current encoded frame after the completion of all LCU codings, then present frame coding is completed, and reads in next frame figure afterwards Picture, each LCU of 1 to 7 pair of repeat step divided in frame setting and the coding of depth bounds.
- 2. a kind of HEVC intraframe codings depth division high-speed decision method based on significance according to claim 1, its It is characterised by, the notable angle value after current frame to be encoded is normalized by LCU layers according to ratio in the step 5 is divided into three The method in region specifically includes:A) sum for remembering current all LCU of frame to be encoded is NLCU=P × Q, (p, q) individual LCU are LCU (p, q);B) by the notable angle value S after current all LCU normalization of frame to be encodedLCU(p, q) is according to incremental order i.e. from small to large Sequence;C) N is definedLCUTwo-dimensional array S [the N that row 2 arrangesLCU] [2], first row storage label [1,2 ..., NLCU], secondary series storage LCU layers sort from small to large after notable angle value;D) the low significance threshold value of LCU layers is takenHigh significance threshold value WhereinExpression rounds downwards, i.e. threshold value SaWith SbIt is to be determined according to the quantitative proportion of whole frame LCU significances, rather than according to exhausted Notable angle value is determined;E) magnitude relationship of notable angle value and two threshold values after being normalized according to each LCU, there will be difference using formula (3) The LCU of notable angle value is divided into low significance region SL, middle significance region SMWith high significance region SH:<mrow> <msub> <mi>A</mi> <mrow> <mi>L</mi> <mi>C</mi> <mi>U</mi> </mrow> </msub> <mrow> <mo>(</mo> <mi>p</mi> <mo>,</mo> <mi>q</mi> <mo>)</mo> </mrow> <mo>&Element;</mo> <mfenced open = "{" close = ""> <mtable> <mtr> <mtd> <mrow> <msub> <mi>S</mi> <mi>L</mi> </msub> <mo>,</mo> </mrow> </mtd> <mtd> <mrow> <mi>i</mi> <mi>f</mi> <mi> </mi> <msub> <mi>S</mi> <mrow> <mi>L</mi> <mi>C</mi> <mi>U</mi> </mrow> </msub> <mrow> <mo>(</mo> <mi>p</mi> <mo>,</mo> <mi>q</mi> <mo>)</mo> </mrow> <mo>&le;</mo> <msub> <mi>S</mi> <mi>a</mi> </msub> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <msub> <mi>S</mi> <mi>M</mi> </msub> <mo>,</mo> </mrow> </mtd> <mtd> <mrow> <mi>i</mi> <mi>f</mi> <mi> </mi> <msub> <mi>S</mi> <mi>a</mi> </msub> <mo><</mo> <msub> <mi>S</mi> <mrow> <mi>L</mi> <mi>C</mi> <mi>U</mi> </mrow> </msub> <mrow> <mo>(</mo> <mi>p</mi> <mo>,</mo> <mi>q</mi> <mo>)</mo> </mrow> <mo>&le;</mo> <msub> <mi>S</mi> <mi>b</mi> </msub> </mrow> </mtd> </mtr> <mtr> <mtd> <msub> <mi>S</mi> <mi>H</mi> </msub> </mtd> <mtd> <mrow> <mi>i</mi> <mi>f</mi> <mi> </mi> <msub> <mi>S</mi> <mrow> <mi>L</mi> <mi>C</mi> <mi>U</mi> </mrow> </msub> <mrow> <mo>(</mo> <mi>p</mi> <mo>,</mo> <mi>q</mi> <mo>)</mo> </mrow> <mo>></mo> <msub> <mi>S</mi> <mi>b</mi> </msub> </mrow> </mtd> </mtr> </mtable> </mfenced> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>3</mn> <mo>)</mo> </mrow> </mrow>Wherein ALCU(p, q) represents the significance region that LCU (p, q) is divided into.
- 3. a kind of HEVC intraframe codings depth division high-speed decision method based on significance according to claim 1, its It is characterised by, each LCU that is divided into the step 6 according to present frame significance region sets HEVC intraframe predictive codings The method of Shi Butong depth division scope is as follows:A) when HEVC carries out intra mode decision in units of LCU, depth { 0,1,2,3 }, the depth of consideration are divided totally in 4 frames More, then what is divided in LCU frames is finer, and code efficiency is higher, but computation complexity is also higher, if can be divided in LCU frames When, its optimal division depth is predicted, then can reduce computation complexity while code efficiency is kept;First, in Low significance region SLIn LCU, the region that its quantity accounts for the 10% of whole frame LCU and less paid close attention in human eye, setting is in Low significance region SLIn LCU frame in division depth bounds be { 0,1,2 };B) similarly, in middle significance region SMLCU, its quantity accounts for the 70% of whole frame LCU and compares concern in human eye Region, setting are in middle significance region SMLCU frame in division depth bounds be { 1,2,3 };C) in high significance region SHLCU, the region that its quantity accounts for the 20% of whole frame LCU and paid special attention in human eye, Setting is in high significance region SHLCU frame in division depth bounds be { 2,3 };Division depth bounds sets and can be expressed as follows with formula (4) in the frame of LCU in different significance regions:<mrow> <msub> <mi>DR</mi> <mrow> <mi>L</mi> <mi>C</mi> <mi>U</mi> </mrow> </msub> <mrow> <mo>(</mo> <mi>p</mi> <mo>,</mo> <mi>q</mi> <mo>)</mo> </mrow> <mo>&Element;</mo> <mfenced open = "{" close = ""> <mtable> <mtr> <mtd> <mrow> <mo>{</mo> <mn>0</mn> <mo>,</mo> <mn>1</mn> <mo>,</mo> <mn>2</mn> <mo>}</mo> <mo>,</mo> </mrow> </mtd> <mtd> <mrow> <mi>i</mi> <mi>f</mi> <mi> </mi> <msub> <mi>A</mi> <mrow> <mi>L</mi> <mi>C</mi> <mi>U</mi> </mrow> </msub> <mrow> <mo>(</mo> <mi>p</mi> <mo>,</mo> <mi>q</mi> <mo>)</mo> </mrow> <mo>&Element;</mo> <msub> <mi>S</mi> <mi>L</mi> </msub> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mo>{</mo> <mn>1</mn> <mo>,</mo> <mn>2</mn> <mo>,</mo> <mn>3</mn> <mo>}</mo> <mo>,</mo> </mrow> </mtd> <mtd> <mrow> <mi>i</mi> <mi>f</mi> <mi> </mi> <msub> <mi>A</mi> <mrow> <mi>L</mi> <mi>C</mi> <mi>U</mi> </mrow> </msub> <mrow> <mo>(</mo> <mi>p</mi> <mo>,</mo> <mi>q</mi> <mo>)</mo> </mrow> <mo>&Element;</mo> <msub> <mi>S</mi> <mi>M</mi> </msub> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mo>{</mo> <mn>2</mn> <mo>,</mo> <mn>3</mn> <mo>}</mo> <mo>,</mo> </mrow> </mtd> <mtd> <mrow> <mi>i</mi> <mi>f</mi> <mi> </mi> <msub> <mi>A</mi> <mrow> <mi>L</mi> <mi>C</mi> <mi>U</mi> </mrow> </msub> <mrow> <mo>(</mo> <mi>p</mi> <mo>,</mo> <mi>q</mi> <mo>)</mo> </mrow> <mo>&Element;</mo> <msub> <mi>S</mi> <mi>H</mi> </msub> </mrow> </mtd> </mtr> </mtable> </mfenced> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>4</mn> <mo>)</mo> </mrow> </mrow>Wherein DRLCU(p, q) represents that the depth of division in LCU (p, q) frame sets scope.
- 4. a kind of HEVC intraframe codings depth division high-speed decision method based on significance according to claim 1, its It is characterised by, the specific method of the notable angle value of current frame pixel layer is calculated in the step 2 so that accurate pixel layer can be obtained Notable angle value is condition.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710646984.4A CN107483931B (en) | 2017-08-01 | 2017-08-01 | A kind of HEVC intraframe coding depth division high-speed decision method based on significance |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710646984.4A CN107483931B (en) | 2017-08-01 | 2017-08-01 | A kind of HEVC intraframe coding depth division high-speed decision method based on significance |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107483931A true CN107483931A (en) | 2017-12-15 |
CN107483931B CN107483931B (en) | 2019-10-08 |
Family
ID=60597539
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710646984.4A Active CN107483931B (en) | 2017-08-01 | 2017-08-01 | A kind of HEVC intraframe coding depth division high-speed decision method based on significance |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107483931B (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110662028A (en) * | 2018-06-30 | 2020-01-07 | 上海天荷电子信息有限公司 | Data compression method and device for encoding parameters by hierarchically defined subsets |
CN110662029A (en) * | 2018-06-30 | 2020-01-07 | 上海天荷电子信息有限公司 | Data compression method and device for dynamically and adaptively limiting value range |
CN111107359A (en) * | 2019-12-16 | 2020-05-05 | 暨南大学 | Intra-frame prediction coding unit dividing method suitable for HEVC standard |
WO2020227911A1 (en) * | 2019-05-10 | 2020-11-19 | 上海德衡数据科技有限公司 | Method for accelerating coding/decoding of hevc video sequence |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103686165A (en) * | 2012-09-05 | 2014-03-26 | 乐金电子(中国)研究开发中心有限公司 | Depth image intra-frame coding and decoding method, video encoder and video decoder |
CN104639940A (en) * | 2015-03-06 | 2015-05-20 | 宁波大学 | Quick HEVC (High Efficiency Video Coding) inter-frame prediction mode selection method |
CN105120295A (en) * | 2015-08-11 | 2015-12-02 | 北京航空航天大学 | HEVC complexity control method based on quadtree coding segmentation |
CN105491385A (en) * | 2015-12-02 | 2016-04-13 | 同济大学 | Quick deciding method for division of HEVC intraframe coding units |
-
2017
- 2017-08-01 CN CN201710646984.4A patent/CN107483931B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103686165A (en) * | 2012-09-05 | 2014-03-26 | 乐金电子(中国)研究开发中心有限公司 | Depth image intra-frame coding and decoding method, video encoder and video decoder |
CN104639940A (en) * | 2015-03-06 | 2015-05-20 | 宁波大学 | Quick HEVC (High Efficiency Video Coding) inter-frame prediction mode selection method |
CN105120295A (en) * | 2015-08-11 | 2015-12-02 | 北京航空航天大学 | HEVC complexity control method based on quadtree coding segmentation |
CN105491385A (en) * | 2015-12-02 | 2016-04-13 | 同济大学 | Quick deciding method for division of HEVC intraframe coding units |
Non-Patent Citations (4)
Title |
---|
HAN-CHIOU FANG等: "FAST INTRA PREDICTION ALGORITHM AND DESIGN FOR HIGH EFFICIENCY VIDEO CODING", 《IEEE EXPLORE》 * |
冯磊: "数字视频的HEVC编码快速算法的研究", 《中国博士学位论文全文数据库》 * |
姬瑞旭: "HEVC帧内模式决策和CU划分快速算法", 《中国优秀硕士学位论文全文数据库》 * |
鲁雯: "基于空时域特征的视觉显著图生成算法", 《电视技术》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110662028A (en) * | 2018-06-30 | 2020-01-07 | 上海天荷电子信息有限公司 | Data compression method and device for encoding parameters by hierarchically defined subsets |
CN110662029A (en) * | 2018-06-30 | 2020-01-07 | 上海天荷电子信息有限公司 | Data compression method and device for dynamically and adaptively limiting value range |
CN110662028B (en) * | 2018-06-30 | 2023-03-03 | 上海天荷电子信息有限公司 | Encoding method and device for compressing sequence of images, and decoding method and device |
WO2020227911A1 (en) * | 2019-05-10 | 2020-11-19 | 上海德衡数据科技有限公司 | Method for accelerating coding/decoding of hevc video sequence |
US11212518B2 (en) | 2019-05-10 | 2021-12-28 | Shanghai Data Center Science Co., Ltd | Method for accelerating coding and decoding of an HEVC video sequence |
CN111107359A (en) * | 2019-12-16 | 2020-05-05 | 暨南大学 | Intra-frame prediction coding unit dividing method suitable for HEVC standard |
Also Published As
Publication number | Publication date |
---|---|
CN107483931B (en) | 2019-10-08 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111147867B (en) | Multifunctional video coding CU partition rapid decision-making method and storage medium | |
CN107483931A (en) | A kind of HEVC intraframe codings depth division high-speed decision method based on significance | |
CN103281538B (en) | Based on the inner frame coding method of rolling Hash and block rank infra-frame prediction | |
CN105025293B (en) | Method and apparatus to Video coding and to the decoded method and apparatus of video | |
CN104967850B (en) | The method and apparatus that image is coded and decoded by using big converter unit | |
CN104602010B (en) | Method and apparatus for encoding and decoding coding unit of picture boundary | |
CN103650496B (en) | Intra prediction pixel-based for being encoded in HEVC | |
CN104980743B (en) | Video decoding apparatus | |
CN104378643B (en) | A kind of 3D video depths image method for choosing frame inner forecast mode and system | |
CN102158712B (en) | Multi-viewpoint video signal coding method based on vision | |
CN106105191A (en) | For the method and apparatus processing multiview video signal | |
CN106716997A (en) | Video encoding method using in-loop filter parameter prediction and apparatus therefor, and video decoding method and apparatus therefor | |
CN107623850A (en) | A kind of quick screen contents encoding method based on temporal correlation | |
CN106303521B (en) | A kind of HEVC Rate-distortion optimization method based on sensitivity of awareness | |
CN106791768A (en) | A kind of depth map frame per second method for improving that optimization is cut based on figure | |
CN102447925A (en) | Method and device for synthesizing virtual viewpoint image | |
CN103702122B (en) | Coding mode selection method, device and encoder | |
CN106375766A (en) | Light field image compression method | |
CN103974063A (en) | Encoding and decoding method of depth model and video coder decoder | |
CN110446052A (en) | The quick CU depth selection method of depth map in a kind of 3D-HEVC frame | |
CN101977313B (en) | Video signal coding device and method | |
CN103391439B (en) | A kind of H.264/AVC bit rate control method hidden based on active macro block | |
CN101710985A (en) | Image brightness compensation method for image coding | |
CN102547282B (en) | Extensible video coding error hiding method, decoder and system | |
CN103458238A (en) | Scalable video code rate controlling method and device combined with visual perception |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |