CN110519597A - A kind of coding method based on HEVC, calculates equipment and medium at device - Google Patents
A kind of coding method based on HEVC, calculates equipment and medium at device Download PDFInfo
- Publication number
- CN110519597A CN110519597A CN201910837345.5A CN201910837345A CN110519597A CN 110519597 A CN110519597 A CN 110519597A CN 201910837345 A CN201910837345 A CN 201910837345A CN 110519597 A CN110519597 A CN 110519597A
- Authority
- CN
- China
- Prior art keywords
- video frame
- current video
- roi
- coding
- hevc
- 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
- H04N19/102—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
- H04N19/109—Selection of coding mode or of prediction mode among a plurality of temporal 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
- H04N19/142—Detection of scene cut or scene change
-
- 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
- H04N19/134—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
- H04N19/167—Position within a video image, e.g. region of interest [ROI]
-
- 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
- H04N19/17—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 an image region, e.g. an object
- H04N19/176—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 an image region, e.g. an object the region being a block, e.g. a macroblock
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 coding method based on HEVC, device, equipment and medium are calculated, this method comprises: obtaining current video frame to be encoded, and identifies the practical region of interest ROI for including in the current video frame；According to the positional relationship between the multiple HEVC coding units and the practical ROI of the current video frame, the coding depth of each HEVC coding unit is calculated；Using the matched inter-frame forecast mode of coding depth with each HEVC coding unit, Video coding is carried out to the current video frame.The technical solution of the embodiment of the present invention is when encoding whole frame video image, eliminate unnecessary inter-frame forecast mode, the calculation amount for reducing rate distortion costs ensure that the real-time and accuracy encoded under selected inter-frame forecast mode, improve the coding rate of HEVC.
Description
Technical field
The present embodiments relate to technology of video compressing encoding more particularly to a kind of coding method based on HEVC, device,
Calculate equipment and medium.
Background technique
Standard branch office, international telecommunication union telecommunication (ITU-T, ITU-T for ITU Telecommunication
Standardization Sector) formulate HEVC of new generation (High Efficiency Video Coding, high-performance view
Frequency coding standard).The main target of HEVC is to double the video image compression efficiency of high-resolution or high-fidelity.
CU (Coding Unit, coding unit) in HEVC is generally divided using quaternary tree recurrence, each coding depth
Under CU have its corresponding inter-frame forecast mode.The CU that HEVC encoder is directed under different depth can be top-downly to institute
There is prediction mode to carry out motion estimation and compensation, rate distortion costs calculated one by one to each of them inter-frame forecast mode,
Find best inter-frame forecast mode of the smallest inter-frame forecast mode of rate distortion costs as current CU.
This traversal calculating process makes the computation complexity of encoder very high, carries out coding consumed by video compress
Time is longer, is unable to satisfy real-time video compress demand.It can be seen that how to exclude unnecessary inter-frame forecast mode to
The operand for effectively reducing rate distortion costs becomes current urgent problem to be solved.
Summary of the invention
The embodiment of the present invention provides a kind of coding method based on HEVC, device, calculates equipment and medium, reduces rate mistake
The operand of true cost, ensure that the real-time and accuracy encoded under selected inter-frame forecast mode, improves the coding of HEVC
Speed.
In a first aspect, the embodiment of the invention provides a kind of coding methods based on HEVC, which comprises
Current video frame to be encoded is obtained, and identifies practical ROI (the region of for including in the current video frame
Interest, area-of-interest)；
According to the positional relationship between the multiple HEVC coding units and the practical ROI of the current video frame, calculate
The coding depth of each HEVC coding unit；
Using the matched inter-frame forecast mode of coding depth with each HEVC coding unit, the current video frame is carried out
Video coding.
Second aspect, the embodiment of the invention also provides a kind of code device based on HEVC, which includes:
Practical ROI identification module for obtaining current video frame to be encoded, and identifies in the current video frame and includes
Practical ROI；
The coding depth computing module of HEVC coding unit, it is single for being encoded according to multiple HEVC of the current video frame
First positional relationship between the practical ROI calculates the coding depth of each HEVC coding unit；
Current video frame coding module, for using the matched inter-prediction mould of coding depth with each HEVC coding unit
Formula carries out Video coding to the current video frame.
The third aspect, the embodiment of the invention also provides a kind of calculating equipment, which includes:
One or more processors；
Storage device, for storing one or more programs；
When one or more of programs are executed by one or more of processors, so that one or more of processing
Device realizes the coding method based on HEVC that any embodiment of that present invention provides.
Fourth aspect stores on the storage medium the embodiment of the invention also provides a kind of computer readable storage medium
There is computer program, the coding staff based on HEVC that any embodiment of that present invention provides is realized when which is executed by processor
Method.
The embodiment of the present invention passes through first obtains the position that current video frame to be encoded determines practical area-of-interest
Then parameter calculates the depth of corresponding coding unit, finally according to coding unit with the positional relationship of practical area-of-interest
Video coding is carried out to current video frame using the inter-frame forecast mode to match with coding unit depth.To whole frame video figure
When as coding, the rate distortion costs estimation to unnecessary inter-frame forecast mode is omitted in the embodiment of the present invention, solves rate mistake
The huge problem of true operand ensure that the real-time and accuracy encoded under selected inter-frame forecast mode, improve HEVC's
Coding rate.
Detailed description of the invention
Fig. 1 is the flow chart of coding method of one of the embodiment of the present invention one based on HEVC；
Fig. 2 is the flow chart of coding method of one of the embodiment of the present invention two based on HEVC；
Fig. 3 is coding method specific implementation flow chart of one of the embodiment of the present invention three based on HEVC；
Fig. 4 is code device structure chart of one of the embodiment of the present invention four based on HEVC；
Fig. 5 is that one of embodiment of the present invention five calculates device structure schematic diagram.
Specific embodiment
The present invention is described in further detail with reference to the accompanying drawings and examples.It is understood that this place is retouched
The specific embodiment stated is used only for explaining the present invention rather than limiting the invention.It also should be noted that in order to just
Only the parts related to the present invention are shown in description, attached drawing rather than entire infrastructure.
Embodiment one
Fig. 1 is a kind of coding method flow chart based on HEVC that the embodiment of the present invention one provides, and the present embodiment is applicable
In multimedia communication the case where video compression coding, this method can be by being executed based on the code device of HEVC, the device
It can be realized, and generally be can integrate in terminal or service with Video compression function by software and/or hardware
In device, typically, it is integrated in the video encoder of terminal or server, specifically comprises the following steps:
Step 110 obtains current video frame to be encoded, and identifies the practical ROI for including in the current video frame.
In the present embodiment, obtain current video frame to be encoded mode can for video encoder from generate video counts
According to equipment or computer directly acquire.ROI refers to the approximate region for needing to handle and analyze in current video frame, referred to as interested
Region is usually sketched the contours of in a manner of box, circle, ellipse and irregular polygon etc., and practical ROI, which refers to, has accurate position ginseng
Several area-of-interests.
Step 120 is closed according to the position between the multiple HEVC coding units and the practical ROI of the current video frame
System calculates the coding depth of each HEVC coding unit.
In this step, HEVC coding unit is the encoding block to current video frame ready-portioned larger size in advance,
Abbreviation CU block.The CU block size that depth is 0 is 64*64, and depth is that the corresponding size of CU block of 1-3 is respectively 32*32,16*16
And 8*8.
Positional relationship between CU block and the practical ROI includes three kinds: be located at practical ROI is internal, outside practical ROI with
And the practical edge ROI, the relative position of CU block and practical ROI determine the coding depth of CU block.
Specifically, being compiled when it is regions of non-interest that CU block, which is in the practical outside ROI, using the CU block of larger size
Code；When CU block is in the practical edge ROI, then needs to carry out micronization processes to image, be compiled using the CU block of minimum dimension
Code；When CU block is in inside practical ROI, encoded using the CU block of smaller size.Wherein, the size of CU block is smaller, compiles
Code depth is bigger.
Step 130, using the matched inter-frame forecast mode of coding depth with each HEVC coding unit, work as forward sight to described
Frequency frame carries out Video coding.
Wherein, in HEVC cataloged procedure, the coding unit under each depth has its corresponding inter-frame forecast mode,
To carry out the motion estimation and compensation in video compression coding.
Specifically, existing inter-frame forecast mode mainly include the following types: skip (skipping) or merge (merging) mode,
Square (square) Fractionation regimen, SMP (Symmetrical Motion Partitions, symmetric motion divide) mode,
AMP (Asymmetric Motion Partitions, asymmetrical movement divide) mode and intra mode.
The technical solution of the embodiment of the present invention obtains current video frame to be encoded, then by video encoder first
The practical area-of-interest for including in video frame before identifying, then according to the multiple HEVC coding units and reality of current video frame
The positional relationship of area-of-interest has determined the coding depth of coding unit, and the last depth according to coding unit uses and its phase
Matched inter-frame forecast mode, to carry out Video coding to current video frame.The technical solution of the present embodiment is guaranteeing HEVC
Under the premise of encoding compression efficiency and area-of-interest picture quality, the meter for finding best inter-frame forecast mode is reduced significantly
Calculation amount accelerates the speed of video compression coding, meets the demand of real-time video compress.
Embodiment two
The present embodiment on the basis of the above embodiment 1, provides practical ROI in specific identification current video frame
Embodiment.Same as the previously described embodiments or corresponding term explains that this embodiment is not repeated.It is obtained in the embodiment of the present invention
Current video to be encoded, and identify the practical ROI for including in current video frame, specific step are as follows: first with interframe figure
As difference algorithm, the alternative interested pixel point for including in the current video frame is identified, then obtains and works as forward sight with described
The corresponding photographed scene of frequency frame, finally selection and the matched recognition strategy of the photographed scene, according to the alternative picture interested
Vegetarian refreshments identifies the practical ROI for including in the current video frame.
Fig. 2 is the coding method flow chart provided by Embodiment 2 of the present invention based on HEVC, as shown in Fig. 2, this method packet
Include following steps:
Step 210, using inter frame image difference algorithm, identify the alternative picture interested for including in the current video frame
Vegetarian refreshments.
In this step, the embodiment of the present invention obtains a upper video frame corresponding with current video frame and next view first
Frequency frame, then according to the gray value of pixel each in the current video frame and a upper video frame or next video
Difference in frame between the gray value of the pixel of corresponding position identifies described alternative interested in the current video frame
Pixel.
Wherein, the embodiment of the present invention first determines whether each corresponding position of the current video frame Yu a upper video frame
The absolute value of the difference of gray value of pixel whether be greater than or equal to preset threshold, if so, by the current video frame
First gray value of each pixel is set as 255, and otherwise the first gray value of each pixel of the current video frame is arranged
It is 0；Then judge the difference of the gray value of the pixel of each corresponding position of the current video frame and next video frame
Absolute value whether be greater than or equal to preset threshold, if so, by the second gray value of each pixel of the current video frame
255 are set as, otherwise sets 0 for second gray value of each pixel of the current video frame；Finally work as by described in
In preceding video frame the gray value of each pixel be set as first gray value and second gray value and operation result, with
Bianry image corresponding with the current video frame is obtained, each pixel that gray value is 255 in the bianry image is obtained, is made
For the alternative interested pixel point.
Specifically, two-value difference image corresponding position picture of first gray value between current video frame and a upper video frame
The gray value of vegetarian refreshments, two-value difference image corresponding position pixel of second gray value between current video frame and next video frame
The gray value of point.
Preferably, above-mentioned preset threshold is set as 15 in this step.
Step 220 obtains photographed scene corresponding with the current video frame.
Wherein, the corresponding photographed scene of current video frame is divided into two classes, and the first kind is object of which movement and camera lens is static, and second
Class is object of which movement and camera lens rotates.
In this step, the current video frame is divided by the embodiment of the present invention as unit of the first presetted pixel point first
Multiple level-one blocks；Then each level-one block is divided into as unit of the second presetted pixel point multiple second-order blocks, calculated every
The alternative interested pixel point number for including in a second-order block, and it is small to calculate the wherein alternative interested pixel point number
In the second-order block number of preset value and the ratio of the second-order block total number；It is pre- finally to judge whether the ratio is greater than or equal to
If threshold value, if so, determining that the photographed scene is object of which movement and camera lens static scene, otherwise, it determines the photographed scene
For object of which movement and camera lens rotates scene.
Preferably, the first presetted pixel point is set as 128*128 in this step, and the second presetted pixel point is set as 32*32,
The embodiment of the present invention calculates the alternative interested pixel point number for including in each second-order block, and calculates wherein described alternative
The ratio of interested pixel point number the second-order block number less than 5 and the second-order block total number；Finally judging the ratio is
It is no to be greater than or equal to 0.73, if so, determining that the photographed scene is object of which movement and camera lens static scene, otherwise, it determines institute
Stating photographed scene is object of which movement and camera lens rotation scene.
Step 230 is chosen and is known with the matched recognition strategy of the photographed scene according to the alternative interested pixel point
The practical ROI for including in the not described current video frame.
If photographed scene is first kind scene, that is, object of which movement and camera lens is static, the embodiment of the present invention calculates each institute first
The alternative interested pixel point number for including in level-one block is stated, and wherein the alternative interested pixel point number will be greater than or wait
It is determined as candidate ROI in the level-one block of preset threshold；Then each second-order block is divided as unit of third presetted pixel point
At multiple three-level blocks, according to the field of search of the picture position of each candidate ROI construction and the images match of the candidate ROI
Domain will be searched if not including any alternative interested pixel point in each three-level block in described search region with described
The corresponding candidate ROI in rope region is determined as independent candidate ROI；Finally calculate center of gravity, width and the height of each independent candidate ROI
Degree, and calculated and the independent candidate according to the independently alternative interested pixel point number of candidate ROI, width and height
The corresponding motion intense degree of ROI determines the practical ROI in the independent candidate ROI according to the motion intense degree.
If photographed scene is that the second class scene, that is, object of which movement and camera lens rotate, the embodiment of the present invention first will be described on
The gray value of the pixel of the practical position ROI in one video frame is as reference value, and according to the reality in a upper video frame
The border position ROI, determines search range；Then the picture according to the reference value and current video frame within the scope of described search
The gray value of vegetarian refreshments carries out estimation, obtains multiple match points；Finally work as forward sight according to the determination of the multiple match point is described
Center of gravity, width and the height of the practical ROI of frequency frame is worked as according to center of gravity, width and the height of the practical ROI described
The practical ROI is determined in preceding video frame.
Wherein, if photographed scene is first kind scene, that is, object of which movement and camera lens is static, the embodiment of the present invention calculates each
The alternative interested pixel point number for including in the level-one block, and will wherein the alternative interested pixel point number be greater than or
Level-one block equal to 56 is determined as candidate ROI.Above-mentioned third presetted pixel point is preferably 16*16, as long as candidate make a circle for ROI weeks
Each three-level block in without alternative interested pixel point, this candidate ROI is just independent candidate ROI.By by independent time
It selects the coordinate of the alternative interested pixel point of each of ROI to be superimposed and is averaging the center of gravity that can calculate independent candidate ROI
Coordinate, so that the width and height of independent candidate ROI is calculated, according to the alternative interested pixel point of the independent candidate ROI
Number, width and height calculate motion intense degree corresponding with the independently candidate ROI.
Preferably, if above-mentioned motion intense intensity is greater than or equal to 1.75, which is actual ROI.
If photographed scene is the second class scene, that is, object of which movement and camera lens rotates, by the practical ROI in a upper video frame
The gray value of the pixel of position is taken exercises in current video frame estimation as reference value, the reality of the above video frame
The center of gravity of ROI is origin, and 9 pixels are that radius determines search range, and motion estimated values in search range are less than or equal to 3
Point as optimal match point, center of gravity, width and the height of the practical ROI of the current video frame are determined according to match point,
According to center of gravity, width and the height of the practical ROI, the practical ROI is determined in the current video frame.
Step 240 is closed according to the position between the multiple HEVC coding units and the practical ROI of the current video frame
System calculates the coding depth of each HEVC coding unit.
In this step, according to the photographed scene, the HEVC inside any practical ROI will be not at and encode list
Member is set as the first coding depth or the second coding depth, object of which movement and corresponding first coding depth of camera lens static scene,
Object of which movement and corresponding second coding depth of camera lens rotation scene.It will be partially in inside at least one described practical ROI
HEVC coding unit is set as third coding depth.It will be completely in the HEVC coding unit inside any practical ROI,
It is set as the 4th coding depth.Wherein, the first coding depth is encoded less than the second coding depth, the second coding depth less than third
Depth, third coding depth are greater than the 4th coding depth.Coding depth is bigger, unit block size used in HEVC coding unit
It is smaller.
Wherein, for first kind scene, that is, object of which movement and camera lens it is static, feel emerging when CU block is in outside practical ROI i.e. non-
When interesting region, the first coding depth is set as 0, CU block size and is set as 64*64；When CU block is in the practical edge ROI, third coding
Depth is set as 3, CU block size and is set as 8*8；When CU block is in inside practical ROI, the 4th coding depth is set as 2, CU block size
It is set as 16*16.
For the second class scene, that is, object of which movement and camera lens rotates, and is non-region of interest outside practical ROI when CU block is in
When domain, the second coding depth is set as 1, CU block size and is set as 32*32；When CU block is in the practical edge ROI, third coding depth
It is set as 3, CU block size and is set as 8*8；When CU block is in inside practical ROI, the 4th coding depth is set as 2, CU block size and is set as
16*16。
Step 250, using the matched inter-frame forecast mode of coding depth with each HEVC coding unit, work as forward sight to described
Frequency frame carries out Video coding.
In this step, matched inter-frame forecast mode is selected for the depth of coding unit, about the first kind
Scene, that is, object of which movement and camera lens is static, when i.e. regions of non-interest outside CU block is in practical ROI, prediction mode SKIP
Or merge；When CU block is in the practical edge ROI, prediction mode is AMP and intra；When CU block is in inside practical ROI,
Prediction mode is Square and SMP.
About the second class scene, that is, object of which movement and camera lens rotates, and is non-region of interest outside practical ROI when CU block is in
When domain, inter-frame forecast mode can be SKIP, merge and Square；When CU block is in the practical edge ROI, inter-prediction mould
Formula is AMP and intra；When CU block is in inside practical ROI, inter-frame forecast mode is Square and SMP.
The technical solution of the present embodiment passes through inter frame image difference algorithm first and identifies include in current video frame standby
Interested pixel point is selected, simply and effectively identifies rough area-of-interest；Then bat corresponding with current video frame is obtained
Scene is taken the photograph, chooses and is identified described current with the matched recognition strategy of the photographed scene according to the alternative interested pixel point
The practical ROI for including in video frame, provides condition for Fast video coding；Finally compiled according to multiple HEVC of current video frame
Positional relationship between code unit and practical ROI, calculates the coding depth of each HEVC coding unit, single using encoding with each HEVC
The matched inter-frame forecast mode of coding depth of member carries out Video coding to current video frame.The present embodiment is interested by determination
Region eliminates unnecessary inter-frame forecast mode, reduces the calculation amount of rate distortion costs, is guaranteeing area-of-interest video
Coding rate is greatly improved under the premise of quality, realizes the real-time coding in video transmitting procedure.
Embodiment three
Fig. 3 is that flow chart, this implementation is embodied in a kind of coding method based on HEVC that the embodiment of the present invention three provides
Example provides a specific implementation step of the coding method based on HEVC on the basis of the various embodiments described above.Such as Fig. 3 institute
Show, this method comprises the following steps:
Using inter frame image difference algorithm, the alternative interested pixel point for including in the current video frame is identified；
In this step, the embodiment of the present invention obtains a upper video frame corresponding with current video frame and next view first
Frequency frame, then according to the gray value of pixel each in the current video frame and a upper video frame or next video
Difference in frame between the gray value of the pixel of corresponding position identifies described alternative interested in the current video frame
Pixel.
Obtain photographed scene corresponding with the current video frame；Wherein, the corresponding photographed scene of current video frame is divided into
Two classes, the first kind is object of which movement and camera lens is static, and the second class is object of which movement and camera lens rotates.
In this step, the current video frame is divided by the embodiment of the present invention as unit of the first presetted pixel point first
Multiple level-one blocks；Then each level-one block is divided into as unit of the second presetted pixel point multiple second-order blocks, calculated every
The alternative interested pixel point number for including in a second-order block, and it is small to calculate the wherein alternative interested pixel point number
In the second-order block number of preset value and the ratio of the second-order block total number.
Judge whether above-mentioned ratio is greater than or equal to preset threshold, if so, determining that the photographed scene is object of which movement
And camera lens static scene, otherwise, it determines the photographed scene is object of which movement and camera lens rotates scene.
The scene that will acquire carries out classification processing；
It chooses and is worked as described in identification with the matched recognition strategy of the photographed scene according to the alternative interested pixel point
The practical ROI for including in preceding video frame；
In this step, different recognition strategies is chosen for different photographed scenes, if photographed scene is the first kind
Scene, that is, object of which movement and camera lens is static, the embodiment of the present invention calculates the alternative picture interested for including in each level-one block first
Vegetarian refreshments number, and the level-one block that the wherein alternative interested pixel point number is greater than or equal to preset threshold is determined as candidate
ROI；Then each second-order block is divided into multiple three-level blocks as unit of third presetted pixel point, according to each candidate
The region of search of the picture position construction and the images match of the candidate ROI of ROI, if each of described search region three
Do not include any alternative interested pixel point in grade block, then candidate ROI corresponding with described search region is determined as independence
Candidate ROI；Center of gravity, width and the height of each independent candidate ROI are finally calculated, and according to the standby of the independent candidate ROI
Interested pixel point number, width and height is selected to calculate motion intense degree corresponding with the independently candidate ROI, according to the fortune
Dynamic acutely degree determines the practical ROI in the independent candidate ROI.
If photographed scene is that the second class scene, that is, object of which movement and camera lens rotate, the embodiment of the present invention first will be described on
The gray value of the pixel of the practical position ROI in one video frame is as reference value, and according to the reality in a upper video frame
The border position ROI, determines search range；Then the picture according to the reference value and current video frame within the scope of described search
The gray value of vegetarian refreshments carries out estimation, obtains multiple match points；Finally work as forward sight according to the determination of the multiple match point is described
Center of gravity, width and the height of the practical ROI of frequency frame is worked as according to center of gravity, width and the height of the practical ROI described
The practical ROI is determined in preceding video frame.
By the positional relationship classification between CU block and the practical ROI；
Wherein, CU block is the encoding block to current video frame ready-portioned larger size in advance, the position of CU block and practical ROI
Setting relationship includes inside external, the practical edge ROI three kinds: practical ROI and practical ROI.
According to the positional relationship between the multiple CU blocks and the practical ROI of the current video frame, each CU block is calculated
Depth；
Wherein, for first kind scene, that is, object of which movement and camera lens it is static, feel emerging when CU block is in outside practical ROI i.e. non-
When interesting region, CU block depth is set as 0, and size is set as 64*64；When CU block is in the practical edge ROI, CU block depth is set as 3, ruler
It is very little to be set as 8*8；When CU block is in inside practical ROI, CU block depth is set as 2, and size is set as 16*16.
For the second class scene, that is, object of which movement and camera lens rotates, and is non-region of interest outside practical ROI when CU block is in
When domain, CU block depth is set as 1, and size is set as 32*32；When CU block is in the practical edge ROI, CU block depth is set as 3, and size is set
For 8*8；When CU block is in inside practical ROI, CU block depth is set as 2, and size is set as 16*16.
Using with each matched inter-frame forecast mode of CU block depth, to the current video frame carry out Video coding；
In this step, matched inter-frame forecast mode is selected for the depth of CU block, about first kind scene
That is object of which movement and camera lens is static, when being regions of non-interest outside CU block is in practical ROI, prediction mode for SKIP or
merge；When CU block is in the practical edge ROI, prediction mode is AMP and intra；When CU block is in inside practical ROI, in advance
Survey mode is Square and SMP.
About the second class scene, that is, object of which movement and camera lens rotates, and is non-region of interest outside practical ROI when CU block is in
When domain, inter-frame forecast mode can be SKIP, merge and Square；When CU block is in the practical edge ROI, inter-prediction mould
Formula is AMP and intra；When CU block is in inside practical ROI, inter-frame forecast mode is Square and SMP.
The technical solution of the present embodiment passes through inter frame image difference algorithm first and identifies include in current video frame standby
Interested pixel point is selected, simply and effectively identifies rough area-of-interest；Then bat corresponding with current video frame is obtained
Scene is taken the photograph, the scene that will acquire carries out classification processing, selection and the matched recognition strategy of the photographed scene, according to described alternative
Interested pixel point identifies the practical ROI for including in the current video frame, provides condition for Fast video coding；Finally
According to the positional relationship between multiple CU blocks of current video frame and practical ROI, calculate the depth of each CU block, using with each CU block
The matched inter-frame forecast mode of depth carries out Video coding to current video frame.The present embodiment is by determining practical area-of-interest
Reduce the calculation amount for finding best inter-frame forecast mode, is greatlyd improve under the premise of guaranteeing area-of-interest video quality
Coding rate, realizes the real-time coding in video transmitting procedure.
Example IV
Fig. 4 is a kind of structure chart for code device based on HEVC that the embodiment of the present invention four provides, which includes: reality
The coding depth computing module 420 and current video frame coding module 430 of border ROI identification module 410, HEVC coding unit.
Wherein, practical ROI identification module 410 for obtaining current video frame to be encoded, and identifies the current video
The practical ROI for including in frame；The coding depth computing module 420 of HEVC coding unit, for according to the current video frame
Positional relationship between multiple HEVC coding units and the practical ROI calculates the coding depth of each HEVC coding unit；
Current video frame coding module 430, it is right for using the matched inter-frame forecast mode of coding depth with each HEVC coding unit
The current video frame carries out Video coding.
The technical solution of the present embodiment determines the reality for including in current video frame by practical ROI identification module first
ROI provides condition for subsequent video coding；Then according to the positional relationship of HEVC coding unit and practical ROI, pass through HEVC
The coding depth computing module of coding unit calculates the corresponding coding depth of current video frame different location；Finally according to each volume
Coding depth corresponding to code unit, is used and the matched inter-frame forecast mode of coding depth using current video frame coding module
Current video frame is encoded.One kind is present embodiments provided effectively based on the code device of HEVC, HEVC is solved and seeks
The problem for looking for best inter-frame forecast mode operand huge, reduces the complexity of HEVC cataloged procedure, meets real-time view
The demand of frequency compression.
On the basis of the various embodiments described above, practical ROI identification module 410 may include:
Alternative interested pixel point identification module identifies the current video for utilizing inter frame image difference algorithm
The alternative interested pixel point for including in frame；
Photographed scene obtains module, for obtaining photographed scene corresponding with the current video frame；
Practical ROI determining module alternatively feels emerging according to described for selection and the matched recognition strategy of the photographed scene
Interesting pixel determines the practical ROI for including in the current video frame.
Wherein, alternative interested pixel point identification module may include:
Video frame obtains module, for obtaining a upper video frame corresponding with the current video frame and next video
Frame；
Alternative interested pixel point determining module, for according to pixel each in the current video frame gray value and institute
The difference in a video frame or next video frame between the gray value of the pixel of corresponding position is stated, described current
The alternative interested pixel point is identified in video frame.
Alternatively interested pixel point determining module may include:
First gray value setup module, for judging each corresponding position of the current video frame with a upper video frame
Whether the absolute value of the difference of the gray value for the pixel set is greater than or equal to preset threshold, if so, by the current video frame
The first gray value of each pixel be set as 255, otherwise the first gray value of each pixel of the current video frame is set
It is set to 0；
Second gray value setup module, for judging each corresponding position of the current video frame with next video frame
Whether the absolute value of the difference of the gray value for the pixel set is greater than or equal to preset threshold, if so, by the current video frame
The second gray value of each pixel be set as 255, otherwise by second gray scale of each pixel of the current video frame
Value is set as 0；
Current video frame gray scale setup module, for setting institute for the gray value of pixel each in the current video frame
State the first gray value and second gray value and operation result, to obtain binary map corresponding with the current video frame
Picture；
Wherein, each pixel that gray value is 255 in the bianry image is obtained, as the alternative interested pixel
Point.
Photographed scene obtains module
Level-one block division module, for the current video frame to be divided into multiple level-ones as unit of the first presetted pixel point
Block；
Second-order block division module, for each level-one block to be divided into multiple two as unit of the second presetted pixel point
Grade block；
Photographed scene determining module, for calculating the alternative interested pixel point number for including in each second-order block,
And calculate second-order block number and the second-order block total number that the wherein alternative interested pixel point number is less than preset value
Ratio, then judges whether the ratio is greater than or equal to preset threshold, if so, determining that the photographed scene is object of which movement
And camera lens static scene, otherwise, it determines the photographed scene is object of which movement and camera lens rotates scene.
Practical ROI determining module may include:
Candidate ROI determining module is applied to object of which movement and camera lens static scene, wraps for calculating in each level-one block
The alternative interested pixel point number included, and the wherein alternative interested pixel point number is greater than or equal to preset threshold
Level-one block is determined as candidate ROI；
Three-level block division module is applied to object of which movement and camera lens static scene, is used for each second-order block with third
Presetted pixel point is that unit is divided into multiple three-level blocks；
Region of search constructing module is applied to object of which movement and camera lens static scene, for according to each candidate ROI's
The region of search of picture position construction and the images match of the candidate ROI；
Independent candidate's ROI determining module, is applied to object of which movement and camera lens static scene, if in described search region
It does not include any alternative interested pixel point in each three-level block, then this module is for will time corresponding with described search region
ROI is selected to be determined as independent candidate ROI；
Motion intense degree computing module is applied to object of which movement and camera lens static scene, for calculating each independent time
Center of gravity, width and the height of ROI are selected, and according to alternative interested pixel point number, width and the height of the independent candidate ROI
Motion intense degree corresponding with the independently candidate ROI is calculated, according to the motion intense degree, in the independent candidate ROI
Determine the practical ROI；
Search range determining module is applied to for object of which movement and camera lens rotates scene, is used for a upper video
The gray value of the pixel of the practical position ROI in frame is as reference value, and according to the practical ROI institute in a upper video frame
In position, search range is determined；
Match point obtains module, is applied to for object of which movement and camera lens rotates scene, for according to the reference value and
The gray value of pixel of the current video frame within the scope of described search carries out estimation, obtains multiple match points；
Practical ROI parameter calculating module is applied to for object of which movement and camera lens rotates scene, for according to the multiple
Match point determines center of gravity, width and the height of the practical ROI of the current video frame, according to the center of gravity of the practical ROI, width
Degree and height, determine the practical ROI in the current video frame.
The coding depth computing module 420 of HEVC coding unit may include:
First coding depth or the second coding depth setup module, for that will be not at any according to the photographed scene
HEVC coding unit inside the practical ROI is set as the first coding depth or the second coding depth, object of which movement and mirror
Corresponding first coding depth of head static scene, object of which movement and corresponding second coding depth of camera lens rotation scene；
Third coding depth setup module, the HEVC coding for that will be partially in inside at least one described practical ROI
Unit is set as third coding depth；
4th coding depth setup module, the HEVC coding unit for that will be completely in inside any practical ROI,
It is set as the 4th coding depth；
Wherein, the first coding depth is less than third coding depth, third volume less than the second coding depth, the second coding depth
Code depth is greater than the 4th coding depth；Coding depth is bigger, and unit block size used in HEVC coding unit is smaller.
Embodiment five
Fig. 5 is a kind of structural schematic diagram for calculating equipment that the embodiment of the present invention five provides, as shown in figure 5, the calculating is set
Standby includes processor 510, memory 520, input unit 530 and output device 540；Calculate the quantity of processor 510 in equipment
It can be one or more, in Fig. 5 by taking a processor 510 as an example；Calculate processor 510 in equipment, memory 520, defeated
Entering device 530 can be connected with output device 540 by bus or other modes, in Fig. 5 for being connected by bus.
Memory 520 is used as a kind of computer readable storage medium, can be used for storing software program, journey can be performed in computer
Sequence and module, if the corresponding program instruction/module of the coding method based on HEVC in the embodiment of the present invention is (for example, be based on
Practical ROI identification module 410 in the code device of HEVC, the coding depth computing module 420 of HEVC coding unit and current
Coding video frames module 430).Software program, instruction and the module that processor 510 is stored in memory 520 by operation,
Thereby executing the various function application and data processing for calculating equipment, that is, realize the above-mentioned coding method based on HEVC.
That is, realization when the program is executed by processor:
Current video frame to be encoded is obtained, and identifies the practical ROI for including in the current video frame；
According to the positional relationship between the multiple HEVC coding units and the practical ROI of the current video frame, calculate
The coding depth of each HEVC coding unit；
Using the matched inter-frame forecast mode of coding depth with each HEVC coding unit, the current video frame is carried out
Video coding.
Memory 520 can mainly include storing program area and storage data area, wherein storing program area can store operation system
Application program needed for system, at least one function；Storage data area, which can be stored, uses created data etc. according to terminal.This
Outside, memory 520 may include high-speed random access memory, can also include nonvolatile memory, for example, at least one
Disk memory, flush memory device or other non-volatile solid state memory parts.In some instances, memory 520 can be into one
Step includes the memory remotely located relative to processor 510, these remote memories can be set by network connection to calculating
It is standby.The example of above-mentioned network includes but is not limited to internet, intranet, local area network, mobile radio communication and combinations thereof.
Input unit 530 can be used for receiving the number or character information of input, and generates and set with the user for calculating equipment
It sets and the related key signals of function control inputs, may include keyboard and mouse etc..Output device 540 may include display screen etc.
Show equipment.
Embodiment six
The embodiment of the present invention six also provides a kind of computer readable storage medium, is stored thereon with computer program, the journey
A kind of coding method based on HEVC described in any embodiment of that present invention is realized when sequence is executed by processor.Certainly, of the invention
A kind of computer readable storage medium provided by embodiment, can execute provided by any embodiment of the invention be based on
Relevant operation in the coding method of HEVC.That is, realization when the program is executed by processor:
Current video frame to be encoded is obtained, and identifies the practical ROI for including in the current video frame；
According to the positional relationship between the multiple HEVC coding units and the practical ROI of the current video frame, calculate
The coding depth of each HEVC coding unit；
Using the matched inter-frame forecast mode of coding depth with each HEVC coding unit, the current video frame is carried out
Video coding.
By the description above with respect to embodiment, it is apparent to those skilled in the art that, the present invention
It can be realized by software and required common hardware, naturally it is also possible to which by hardware realization, but in many cases, the former is more
Good embodiment.Based on this understanding, technical solution of the present invention substantially in other words contributes to the prior art
Part can be embodied in the form of software products, which can store in computer readable storage medium
In, floppy disk, read-only memory (Read-Only Memory, ROM), random access memory (Random such as computer
Access Memory, RAM), flash memory (FLASH), hard disk or CD etc., including some instructions are with so that a computer is set
Standby (can be personal computer, server or the network equipment etc.) executes method described in each embodiment of the present invention.
It is worth noting that, in the embodiment of the above-mentioned code device based on HEVC, included each unit and module
It is only divided according to the functional logic, but is not limited to the above division, as long as corresponding functions can be realized；
In addition, the specific name of each functional unit is also only for convenience of distinguishing each other, the protection scope being not intended to restrict the invention.
Note that the above is only a better embodiment of the present invention and the applied technical principle.It will be appreciated by those skilled in the art that
The invention is not limited to the specific embodiments described herein, be able to carry out for a person skilled in the art it is various it is apparent variation,
It readjusts and substitutes without departing from protection scope of the present invention.Therefore, although being carried out by above embodiments to the present invention
It is described in further detail, but the present invention is not limited to the above embodiments only, without departing from the inventive concept, also
It may include more other equivalent embodiments, and the scope of the invention is determined by the scope of the appended claims.
Claims (11)
1. a kind of coding method based on HEVC characterized by comprising
Current video frame to be encoded is obtained, and identifies the practical region of interest ROI for including in the current video frame；
According to the positional relationship between the multiple HEVC coding units and the practical ROI of the current video frame, each institute is calculated
State the coding depth of HEVC coding unit；
Using the matched inter-frame forecast mode of coding depth with each HEVC coding unit, video is carried out to the current video frame
Coding.
2. the method according to claim 1, wherein obtaining current video frame to be encoded, and working as described in identification
The practical ROI for including in preceding video frame, comprising:
Using inter frame image difference algorithm, the alternative interested pixel point for including in the current video frame is identified；
Obtain photographed scene corresponding with the current video frame；
Selection and the matched recognition strategy of the photographed scene, according to the alternative interested pixel point, determination is described to work as forward sight
The practical ROI for including in frequency frame.
3. according to the method described in claim 2, it is characterized in that, being identified described current using inter frame image difference algorithm
The alternative interested pixel point for including in video frame, comprising:
Obtain a upper video frame corresponding with the current video frame and next video frame；
According in the gray value of pixel each in the current video frame and a upper video frame or next video frame
Difference between the gray value of the pixel of corresponding position identifies the alternative interested pixel in the current video frame
Point.
4. according to the method described in claim 3, it is characterized in that, according to the gray value of pixel each in the current video frame
With the difference in a upper video frame or next video frame between the gray value of the pixel of corresponding position, described
The alternative interested pixel point is identified in current video frame, comprising:
Judge the exhausted of the current video frame and the difference of the gray value of the pixel of each corresponding position of a upper video frame
Whether preset threshold is greater than or equal to value；If so, the first gray value of each pixel of the current video frame is arranged
It is 255, otherwise sets 0 for the first gray value of each pixel of the current video frame；
Judge the exhausted of the current video frame and the difference of the gray value of the pixel of each corresponding position of next video frame
Whether preset threshold is greater than or equal to value；If so, the second gray value of each pixel of the current video frame is arranged
It is 255, otherwise sets 0 for second gray value of each pixel of the current video frame；
First gray value and second gray value are set by the gray value of pixel each in the current video frame
With operation result, to obtain bianry image corresponding with the current video frame；
Each pixel that gray value is 255 in the bianry image is obtained, as the alternative interested pixel point.
5. according to the method described in claim 2, it is characterized in that, obtain corresponding with current video frame photographed scene,
Include:
The current video frame is divided into multiple level-one blocks as unit of the first presetted pixel point；
Each level-one block is divided into multiple second-order blocks as unit of the second presetted pixel point；
The alternative interested pixel point number for including in each second-order block is calculated, and calculates the wherein alternative picture interested
Vegetarian refreshments number is less than the second-order block number of preset value and the ratio of the second-order block total number；
Judge whether the ratio is greater than or equal to preset threshold；
If so, determining that the photographed scene is object of which movement and camera lens static scene；Otherwise, it determines the photographed scene is object
Body movement and camera lens rotation scene.
6. according to the method described in claim 5, it is characterized in that, choosing and the matched recognition strategy of the photographed scene, root
According to the alternative interested pixel point, the practical ROI for including in the current video frame is determined, comprising:
If it is determined that the photographed scene is object of which movement and camera lens static scene, then include in each level-one block standby is calculated
Interested pixel point number is selected, and the wherein alternative interested pixel point number is greater than or equal to the level-one block of preset threshold
It is determined as candidate ROI；
Each second-order block is divided into multiple three-level blocks as unit of third presetted pixel point；
According to the region of search of the picture position of each candidate ROI construction and the images match of the candidate ROI；
If not including any alternative interested pixel point in each three-level block in described search region, will be searched with described
The corresponding candidate ROI in rope region is determined as independent candidate ROI；
Center of gravity, width and the height of each independent candidate ROI are calculated, and according to the alternative interested of the independent candidate ROI
Pixel number, width and height calculate motion intense degree corresponding with the independently candidate ROI；
According to the motion intense degree, the practical ROI is determined in the independent candidate ROI.
7. according to the method described in claim 2, it is characterized in that, choosing and the matched recognition strategy of the photographed scene, root
According to the alternative interested pixel point, the practical ROI for including in the current video frame is determined, comprising:
If it is determined that the photographed scene is object of which movement and camera lens rotates scene, then by the reality in a upper video frame
The gray value of the pixel of the position ROI is as reference value, and according to the practical position ROI in a upper video frame, really
Determine search range；
According to the gray value of the pixel of the reference value and current video frame within the scope of described search, estimation is carried out,
Obtain multiple match points；
Center of gravity, width and the height of the practical ROI of the current video frame are determined according to the multiple match point；
According to center of gravity, width and the height of the practical ROI, the practical ROI is determined in the current video frame.
8. according to the method described in claim 2, it is characterized in that, according to multiple HEVC coding units of the current video frame
With the positional relationship between the practical ROI, the coding depth of each HEVC coding unit is calculated, comprising:
According to the photographed scene, it will be not at the HEVC coding unit inside any practical ROI, be set as the first coding
Corresponding first coding depth of depth or the second coding depth, object of which movement and camera lens static scene, object of which movement and camera lens turn
Dynamic corresponding second coding depth of scene；
It will be partially in the HEVC coding unit inside at least one described practical ROI, is set as third coding depth；
It will be completely in the HEVC coding unit inside any practical ROI, is set as the 4th coding depth；
Wherein, the first coding depth is less than third coding depth, third coding deeply less than the second coding depth, the second coding depth
Degree is greater than the 4th coding depth；Coding depth is bigger, and unit block size used in HEVC coding unit is smaller.
9. a kind of code device based on HEVC characterized by comprising
Practical region of interest ROI identification module for obtaining current video frame to be encoded, and identifies the current video frame
In include practical ROI；
The coding depth computing module of HEVC coding unit, for according to multiple HEVC coding units of the current video frame with
Positional relationship between the practical ROI calculates the coding depth of each HEVC coding unit；
Current video frame coding module, for using the matched inter-frame forecast mode of coding depth with each HEVC coding unit,
Video coding is carried out to the current video frame.
10. a kind of calculating equipment, which is characterized in that the calculating equipment includes:
One or more processors；
Storage device, for storing one or more programs；
When one or more of programs are executed by one or more of processors, so that one or more of processors are real
Now such as a kind of coding method based on HEVC described in any one of claims 1-8.
11. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the program is by processor
A kind of such as coding method based on HEVC described in any one of claims 1-8 is realized when execution.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910837345.5A CN110519597B (en) | 2019-09-05 | 2019-09-05 | HEVC-based encoding method and device, computing equipment and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910837345.5A CN110519597B (en) | 2019-09-05 | 2019-09-05 | HEVC-based encoding method and device, computing equipment and medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110519597A true CN110519597A (en) | 2019-11-29 |
| CN110519597B CN110519597B (en) | 2022-05-10 |
Family
ID=68631103
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910837345.5A Active CN110519597B (en) | 2019-09-05 | 2019-09-05 | HEVC-based encoding method and device, computing equipment and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110519597B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113163199A (en) * | 2021-05-10 | 2021-07-23 | 浙江裕瀚科技有限公司 | H265-based video rapid prediction method, rapid coding method and system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104463910A (en) * | 2014-12-08 | 2015-03-25 | 中国人民解放军国防科学技术大学 | High-speed motion target extraction method based on motion vector |
| CN109063659A (en) * | 2018-08-08 | 2018-12-21 | 北京佳讯飞鸿电气股份有限公司 | The detection and tracking and system of moving target |
| CN109379594A (en) * | 2018-10-31 | 2019-02-22 | 北京佳讯飞鸿电气股份有限公司 | Video coding compression method, device, equipment and medium |
| CN109889838A (en) * | 2018-12-29 | 2019-06-14 | 北京佳讯飞鸿电气股份有限公司 | A kind of HEVC fast encoding method based on ROI region |
-
2019
- 2019-09-05 CN CN201910837345.5A patent/CN110519597B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104463910A (en) * | 2014-12-08 | 2015-03-25 | 中国人民解放军国防科学技术大学 | High-speed motion target extraction method based on motion vector |
| CN109063659A (en) * | 2018-08-08 | 2018-12-21 | 北京佳讯飞鸿电气股份有限公司 | The detection and tracking and system of moving target |
| CN109379594A (en) * | 2018-10-31 | 2019-02-22 | 北京佳讯飞鸿电气股份有限公司 | Video coding compression method, device, equipment and medium |
| CN109889838A (en) * | 2018-12-29 | 2019-06-14 | 北京佳讯飞鸿电气股份有限公司 | A kind of HEVC fast encoding method based on ROI region |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113163199A (en) * | 2021-05-10 | 2021-07-23 | 浙江裕瀚科技有限公司 | H265-based video rapid prediction method, rapid coding method and system |
| CN113163199B (en) * | 2021-05-10 | 2023-06-30 | 浙江裕瀚科技有限公司 | H265-based video rapid prediction method, rapid coding method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110519597B (en) | 2022-05-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Li et al. | PDR-Net: Perception-inspired single image dehazing network with refinement | |
| CN108921782B (en) | Image processing method, device and storage medium | |
| CN111583097A (en) | Image processing method, image processing device, electronic equipment and computer readable storage medium | |
| CN110232418B (en) | Semantic recognition method, terminal and computer readable storage medium | |
| CN109903315B (en) | Method, apparatus, device and readable storage medium for optical flow prediction | |
| Wang et al. | Reference-free DIBR-synthesized video quality metric in spatial and temporal domains | |
| Zuo et al. | Minimum spanning forest with embedded edge inconsistency measurement model for guided depth map enhancement | |
| CN117218246A (en) | Training method and device for image generation model, electronic equipment and storage medium | |
| CN115272565A (en) | Head three-dimensional model reconstruction method and electronic equipment | |
| CN114972016A (en) | Image processing method, image processing apparatus, computer device, storage medium, and program product | |
| CN114972010A (en) | Image processing method, image processing apparatus, computer device, storage medium, and program product | |
| CN110519597A (en) | A kind of coding method based on HEVC, calculates equipment and medium at device | |
| CN113822849A (en) | Identification code detection method and device, electronic equipment and storage medium | |
| CN117132744A (en) | Virtual scene construction method, device, medium and electronic equipment | |
| CN113298707A (en) | Image frame splicing method, video inspection method, device, equipment and storage medium | |
| EP2309452A1 (en) | Method and arrangement for distance parameter calculation between images | |
| CN114842287B (en) | Monocular three-dimensional target detection model training method and device of depth-guided deformer | |
| CN114663810B (en) | Object image augmentation method, device, equipment and storage medium based on multiple modes | |
| CN117372464A (en) | Image processing method, device, computer equipment and storage medium | |
| CN113808157B (en) | Image processing method and device and computer equipment | |
| CN110245553B (en) | Road surface distance measuring method and device | |
| Wang et al. | Coarse-and fine-grained fusion hierarchical network for hole filling in view synthesis | |
| CN116228895B (en) | Video generation method, deep learning model training method, device and equipment | |
| Zhang et al. | Depth Guided Fast Rendering Neural Radiance Fields | |
| Song et al. | A Robust Attention-based Convolutional Neural Network for Monocular Depth Estimation |
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 |