CN104602001A - Low-bit-rate video coding method based on frame difference image - Google Patents

Low-bit-rate video coding method based on frame difference image Download PDF

Info

Publication number
CN104602001A
CN104602001A CN201510023402.8A CN201510023402A CN104602001A CN 104602001 A CN104602001 A CN 104602001A CN 201510023402 A CN201510023402 A CN 201510023402A CN 104602001 A CN104602001 A CN 104602001A
Authority
CN
China
Prior art keywords
frame
difference image
video
frame difference
component
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.)
Pending
Application number
CN201510023402.8A
Other languages
Chinese (zh)
Inventor
刘鹏宇
贾克斌
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing University of Technology
Original Assignee
Beijing University of Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Beijing University of Technology filed Critical Beijing University of Technology
Priority to CN201510023402.8A priority Critical patent/CN104602001A/en
Publication of CN104602001A publication Critical patent/CN104602001A/en
Pending legal-status Critical Current

Links

Abstract

The invention discloses a low-bit-rate video coding method based on a frame difference image. The low-bit-rate video coding method based on the frame difference image includes that inputting a video sequence, if the current input frame fn is the I frame image and not the last frame image, respectively reading components Y, Cb and Cr of the current frame fn and reference frame fn+1; acquiring the luminance component DY and chromaticity component DCb and DCr of the frame difference image; performing scale change to obtain the converted frame difference components RY, RCb and RCr; storing the components RY, RCb and RCr in sequence to synthesize into a new frame difference image DR; using the new frame difference image DR to replace the I frame image in an original video sequence, jointly forming a new video sequence with a P frame image and a B frame image, and serving as the coding object of an HEVC video encoder to perform normal video coding; otherwise, directly entering the HEVC video encoder to perform normal video coding; outputting code stream. The low-bit-rate video coding method based on the frame difference image realizes low-bit-rate video coding based on keeping the quality of the video reconstructed image, improves the video coding compression ratio, saves bit resource and alleviates the contradictions between the reconstructed image quality of the video coding and the coding resource occupancy.

Description

A kind of Low Bit-Rate Video Coding method based on frame difference image
Technical field
The invention belongs to technical field of video processing, relate to a kind of Low Bit-Rate Video Coding method based on frame difference image.
Background technology
In existing video encoding standard, comprise three kinds of encoding frame type, be respectively: I frame, P frame and B frame.In actual coding, usually select the code flow structures such as IPPP, IBBB, IPBP, namely most of frame of video carries out P frame or the predictive coding of B frame.Although P frame and B frame coding respectively by forward prediction, bi-directional predictedly can obtain higher compression ratio, the video reconstruction picture quality of P frame and B frame coding often relies on the encoding precision of corresponding I frame.Therefore, in order to ensure decoded video reconstruction picture quality, usually can, in the mode suitably inserting I frame, avoid the prediction error of P frame and B frame coding to propagate.In addition, when video scene changes, insert I frame and also can improve precision of prediction to subsequent video images and coding quality.Meanwhile, in scalable video and video monitoring coding, also I frame can be added in appropriate amount, for improving Video coding precision.
Therefore, I frame Chang Zuowei key frame records the main information of whole video sequence, is pith indispensable in Video coding.At present, with International video coding standard HEVC of new generation (HighEfficiency Video Coding) in the video encoding standard of representative, normally according to the spatial coherence between coding unit inside, coding unit, compressed encoding is carried out for I frame, compared with encoding with P frame, B frame, I frame coding can consume more bit resources.And the higher bit occupation rate of I frame coding becomes the overall Main Bottleneck improving Video coding compression ratio, save Video coding bit resources further.
Summary of the invention
The object of this invention is to provide a kind of Low Bit-Rate Video Coding method based on frame difference image, under the prerequisite keeping video reconstruction picture quality, realize Low Bit-Rate Video Coding, improve video compression ratio, save bit resources, the contradiction between the reconstructed image quality of alleviation Video coding and coding resource take.
The technical solution adopted in the present invention is, a kind of Low Bit-Rate Video Coding method based on frame difference image, carries out according to following steps:
The first step: input video sequence, reads configuration file, inserts the cycle, judge present incoming frame f according to coding structure GOP and I frame nwhether be I two field picture, if I two field picture, then carry out second step; If not I two field picture, then present incoming frame f nnamely be P two field picture or B two field picture, carry out the 6th step;
Second step: judge present frame f nwhether be last frame, if so, then carry out the 7th step; If not, then choose present frame f na rear two field picture as with reference to frame f n+1, respectively from present frame f nwith reference frame f n+1middle reading Y, Cb, Cr tri-components, obtain present frame f nthree component: Y n, Cb n, Cr nwith reference frame f n+1three component: Y n+1, Cb n+1, Cr n+1;
Wherein, n presentation code frame of video position on a timeline, Y represents luminance component, Cb and Cr represents chromatic component;
3rd step: with reference to frame f n+1each component and present frame f neach component does difference, obtains the luminance component D of frame difference image D (x, y) y(x, y), the chromatic component D of frame difference image D (x, y) cb(x, y) and D cr(x, y):
D ( x , y ) = D Y ( x , y ) = Y n + 1 ( x , y ) - Y n ( x , y ) D Cb ( x , y ) = Cb n + 1 ( x , y ) - Cb n ( x , y ) D Cr ( x , y ) = Cr n + 1 ( x , y ) - Cr n ( x , y ) - - - ( 1 )
Wherein, x and y is illustrated respectively in abscissa and the ordinate of pixel in present frame;
4th step: in order to solve the negative value problem of each component of three-wave-length (1) frame difference image, respectively change of scale is carried out to each component of frame difference image D (x, y):
Wherein, R y(x, y), R cb(x, y) and R cr(x, y) is respectively luminance component and the chromatic component of the frame difference image after change of scale; Max (D (x, y)) and min (D (x, y)) represents maximum and the minimum value of the interior pixel value of D (x, y) respectively; represent lower rounding operation;
5th step: by R y, R cband R crcomponent sequential storage synthesizes new frame difference image D r;
6th step: by new frame difference image D rsubstitute the I two field picture of correspondence position in former video sequence, form new video sequence with P two field picture, B two field picture according to coding structure GOP and the I frame cycle of insertion;
7th step: input HEVC video encoder carries out convention video coding, comprises predictive coding, change quantization coding and entropy code standard code process;
8th step: output encoder code stream.
The invention has the beneficial effects as follows, before employing International video coding standard HEVC carries out Video coding, first preliminary treatment is carried out to input video frame, by carrying out the process of frame difference to I two field picture, obtain rational frame difference image, replace former I two field picture, new video sequence is reformulated again, as the input video source of HEVC video encoder, compared with International video coding standard HEVC with conventional P two field picture, B two field picture, effectively can save number of coded bits, improve Video coding compression ratio on the whole.
Accompanying drawing explanation
Fig. 1 is method for video coding flow chart of the present invention.
Embodiment
Below in conjunction with the drawings and specific embodiments, the present invention is described in detail.
In order to frame difference image reasonably being substituted I two field picture and being applied to Video coding, accurately obtaining frame difference image is an important technology.The frame difference image directly adjacent two two field pictures being done difference acquisition must comprise negative value, the such as luminance component of a certain video image, after doing difference operation through consecutive frame, its distribution is [-255,255], beyond conventional gray value distribution [0,255], therefore, the frame difference image obtained by directly doing difference directly can not be applied to I two field picture coding and rebuild.In order to solve the negative value problem directly done the process of frame difference and cause, the invention discloses a kind of Low Bit-Rate Video Coding method based on frame difference image, flow process as shown in Figure 1, is specifically carried out according to following steps:
The first step: input video sequence, reads configuration file, inserts the cycle, judge present incoming frame f according to coding structure GOP and I frame nwhether be I two field picture, if I two field picture, then carry out second step; If not I two field picture, then present incoming frame f nnamely be P two field picture or B two field picture, carry out the 6th step;
Second step: judge present frame f nwhether be last frame, if so, then carry out the 7th step; If not, then choose present frame f na rear two field picture as with reference to frame f n+1, respectively from present frame f nwith reference frame f n+1middle reading Y, Cb, Cr tri-components, obtain present frame f nthree component: Y n, Cb n, Cr nwith reference frame f n+1three component: Y n+1, Cb n+1, Cr n+1;
Wherein, n presentation code frame of video position on a timeline, Y represents luminance component, Cb and Cr represents chromatic component;
3rd step: with reference to frame f n+1each component and present frame f neach component does difference, obtains the luminance component D of frame difference image D (x, y) y(x, y), the chromatic component D of frame difference image D (x, y) cb(x, y) and D cr(x, y):
D ( x , y ) = D Y ( x , y ) = Y n + 1 ( x , y ) - Y n ( x , y ) D Cb ( x , y ) = Cb n + 1 ( x , y ) - Cb n ( x , y ) D Cr ( x , y ) = Cr n + 1 ( x , y ) - Cr n ( x , y ) - - - ( 1 )
Wherein, x and y is illustrated respectively in abscissa and the ordinate of pixel in present frame;
4th step: in order to solve the negative value problem of each component of three-wave-length (1) frame difference image, respectively change of scale is carried out to each component of frame difference image D (x, y):
Wherein, R y(x, y), R cb(x, y) and R cr(x, y) is respectively luminance component and the chromatic component of the frame difference image after change of scale; Max (D (x, y)) and min (D (x, y)) represents maximum and the minimum value of the interior pixel value of D (x, y) respectively; represent lower rounding operation;
5th step: by R y, R cband R crcomponent sequential storage synthesizes new frame difference image D r;
6th step: by new frame difference image D rsubstitute the I two field picture of correspondence position in former video sequence, form new video sequence with P two field picture, B two field picture according to coding structure GOP and the I frame cycle of insertion;
7th step: input HEVC video encoder carries out convention video coding, comprises predictive coding, change quantization coding and entropy code standard code process;
8th step: output encoder code stream.
Frame difference image is introduced Video coding and mainly make use of following characteristics:
1) frequency characteristic: the frequency of image is the index of grey scale change severe degree in token image.Video sequence has stronger temporal correlation, namely there is larger inter-frame redundancy, and do the poor frame difference image obtained by adjacent two frames, most region is smooth, and its gray-value variation is slow, and corresponding picture frequency value is lower; The a few regions of frame difference image has reacted the change between adjacent two two field pictures, and this kind of have quick spatial variations, region that gray-value variation is violent, and corresponding picture frequency value is higher.Generally, at frame difference image medium and low frequency composition proportion much larger than radio-frequency component proportion.
2) energy response: image energy characteristic can be obtained by the spectrogram analysis after two-dimensional Fourier transform.Because the flat site of frame difference image is more, frame difference image have accumulated more energy in low-frequency range compared with normal picture; And relatively, the energy assembled at high band is less.
3) correlation properties: be different from original video sequence, do through consecutive frame the frame difference image obtained after difference and decrease inter-frame redundancy, temporal correlation weakens, and is unsuitable for and adopts inter-frame encoding to predict; And a large amount of flat sites of frame difference image often have stronger spatial coherence, adopt infra-frame prediction can obtain better predictive coding performance.
4) entropy characteristic: image entropy is the estimated value of image complexity, the entropy having the image of more details and gray-value variation corresponding is higher; The entropy that more smooth image is corresponding is lower.Under normal conditions, compared with original image, the flat site proportion of frame difference image is larger, therefore, makes its entropy be less than original image.
5) prediction characteristic: in predictive mode selection course, because frame difference image comprises more flat site, so its probability selecting large scale pattern to carry out predicting is higher, is conducive to the coding difficulty reducing optimal prediction modes selection.
6) bit characteristic: frame difference image, when carrying out Video coding, after predictive coding, has more residual signals to be tending towards null value.Therefore, compared with original image, frame difference image only needs less bit resources just can presentation code result;
According to above-mentioned feature, can infer that frame difference image reflects the notable feature of original image.Frame difference image is encoded, can realize predicting a large amount of low frequency regions and a small amount of high-frequency region with less coding resource, the computational complexity selecting optimal prediction modes can be reduced simultaneously, can under high-quality reconstruction image, by less bit resources presentation code result, reach the object of Low Bit-rate Coding.
Principle of the present invention is as follows: research finds, using frame difference image as coded object, although do after difference operation through consecutive frame, its grey level distribution scope has expanded [-255,255] of frame difference image to from [0,255] of original image.But normally, the number of the non-zero gray level of frame difference image is the reduction of, there is the characteristic that standard deviation distribution is less and concentrated, frame difference image is replaced original I frame image and carry out Video coding, there is lower information source information entropy, according to Shannon coding theory, Low Bit-Rate Video Coding can be realized to frame difference image coding.For this reason, the present invention proposes a reasonably getting frame difference image method, achieve the Low Bit-Rate Video Coding based on frame difference image.
Method for video coding based on frame difference image can be applicable to based in the scalable video of HEVC and video monitoring due to its low bit-rate advantage.
The Low Bit-Rate Video Coding method and the international standard algorithm HEVC (HM10.0) that the present invention is based on frame difference image compare, and coding efficiency statistics is in table 1.
Coding environment is configured to:
HEVC international standard code algorithm versions HM10.0,
Quantization step QP: choose 22 respectively, 27,32,37;
Coding structure GOP is IBBB;
It is 4 that I frame inserts the cycle;
Code length: 52 frames;
Standard test sequences: choose the international standard cycle tests (Traffic, Cactus, SlideShow, ChinaSpeed, PartyScene, BlowingBubbles) under 6 kinds of different resolutions;
Coding efficiency statistical method: employing general international standard BDBR ( delta bitrate, BDBR).
The Performance comparision of table 1 the inventive method and HEVC international standard code method
Table 1 statistics shows, compared with international video standard HEVC (HM10.0) of new generation, the Low Bit-Rate Video Coding method that the present invention is based on frame difference image presents more outstanding distortion performance.For the cycle tests of polymorphic type, different resolution, under equivalent video reconstructed image quality, on average save the code check of 16.1%, improve video compression ratio on the whole.And along with the increase of quantization step QP, the code check saving effect of the inventive method is more remarkable.Above Exemplary statistical data demonstrates the reasonability of the getting frame difference image that the present invention proposes and realizes the high efficiency of Low Bit-Rate Video Coding based on frame difference image.

Claims (1)

1. based on a Low Bit-Rate Video Coding method for frame difference image, it is characterized in that, carry out according to following steps:
The first step: input video sequence, reads configuration file, inserts the cycle, judge present incoming frame f according to coding structure GOP and I frame nwhether be I two field picture, if I two field picture, then carry out second step; If not I two field picture, then present incoming frame f nnamely be P two field picture or B two field picture, carry out the 6th step;
Second step: judge present frame f nwhether be last frame, if so, then carry out the 7th step; If not, then choose present frame f na rear two field picture as with reference to frame f n+1, respectively from present frame f nwith reference frame f n+1middle reading Y, Cb, Cr tri-components, obtain present frame f nthree component: Y n, Cbn, Cr nwith reference frame f n+1three component: Y n+1, Cb n+1, Cr n+1;
Wherein, n presentation code frame of video position on a timeline, Y represents luminance component, Cb and Cr represents chromatic component;
3rd step: with reference to frame f n+1each component and present frame f neach component does difference, obtains the luminance component D of frame difference image D (x, y) y(x, y), the chromatic component D of frame difference image D (x, y) cb(x, y) and D cr(x, y):
D ( x , y ) = D Y ( x , y ) = Y n + 1 ( x , y ) - Y n ( x , y ) D Cb ( x , y ) = Cb n + 1 ( x , y ) - Cb n ( x , y ) D Cr ( x , y ) = Cr n + 1 ( x , y ) - Cr n ( x , y ) - - - ( 1 )
Wherein, x and y is illustrated respectively in abscissa and the ordinate of pixel in present frame;
4th step: in order to solve the negative value problem of each component of three-wave-length (1) frame difference image, respectively change of scale is carried out to each component of frame difference image D (x, y):
Wherein, R y(x, y), R cb(x, y) and R cr(x, y) is respectively luminance component and the chromatic component of the frame difference image after change of scale; Max (D (x, y)) and min (D (x, y)) represents maximum and the minimum value of the interior pixel value of D (x, y) respectively; represent lower rounding operation;
5th step: by R y, R cband R crcomponent sequential storage synthesizes new frame difference image D r;
6th step: by new frame difference image D rsubstitute the I two field picture of correspondence position in former video sequence, form new video sequence with P two field picture, B two field picture according to coding structure GOP and the I frame cycle of insertion;
7th step: input HEVC video encoder carries out convention video coding, comprises predictive coding, change quantization coding and entropy code standard code process;
8th step: output encoder code stream.
CN201510023402.8A 2015-01-19 2015-01-19 Low-bit-rate video coding method based on frame difference image Pending CN104602001A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201510023402.8A CN104602001A (en) 2015-01-19 2015-01-19 Low-bit-rate video coding method based on frame difference image

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510023402.8A CN104602001A (en) 2015-01-19 2015-01-19 Low-bit-rate video coding method based on frame difference image

Publications (1)

Publication Number Publication Date
CN104602001A true CN104602001A (en) 2015-05-06

Family

ID=53127451

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510023402.8A Pending CN104602001A (en) 2015-01-19 2015-01-19 Low-bit-rate video coding method based on frame difference image

Country Status (1)

Country Link
CN (1) CN104602001A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107749994A (en) * 2017-09-26 2018-03-02 中国科学院电子学研究所 Deep space Coherent Noise in GPR Record compression method based on inter prediction
US10499065B2 (en) 2016-07-21 2019-12-03 Samsung Display Co. Ltd. System and method for sending video data over a wireless channel

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2007047755A1 (en) * 2005-10-17 2007-04-26 Qualcomm Incorporated Adaptive gop structure in video streaming
CN103037219A (en) * 2012-10-22 2013-04-10 北京航空航天大学 Video compression and decompression method based on fractal and H.264
CN103957410A (en) * 2013-12-30 2014-07-30 南京邮电大学 I-frame code rate control method based on residual frequency domain complexity

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2007047755A1 (en) * 2005-10-17 2007-04-26 Qualcomm Incorporated Adaptive gop structure in video streaming
CN103037219A (en) * 2012-10-22 2013-04-10 北京航空航天大学 Video compression and decompression method based on fractal and H.264
CN103957410A (en) * 2013-12-30 2014-07-30 南京邮电大学 I-frame code rate control method based on residual frequency domain complexity

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
YUAN GAO ET AL: "Optimization for I-frame based on residual image", 《2014 TENTH INTERNATIONAL CONFERENCE ON INTELLIGENT INFORMATION HIDING AND MULTIMEDIA SIGNAL PROCESSING》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10499065B2 (en) 2016-07-21 2019-12-03 Samsung Display Co. Ltd. System and method for sending video data over a wireless channel
CN107749994A (en) * 2017-09-26 2018-03-02 中国科学院电子学研究所 Deep space Coherent Noise in GPR Record compression method based on inter prediction

Similar Documents

Publication Publication Date Title
CN102474614B (en) Video encoding method and apparatus and video decoding method and apparatus, based on hierarchical coded block pattern information
CN103220528B (en) Method and apparatus by using large-scale converter unit coding and decoding image
CN101252686B (en) Undamaged encoding and decoding method and system based on interweave forecast
TWI656786B (en) Sampling adaptive offset device
CN102065298B (en) High-performance macroblock coding implementation method
JP2015164358A (en) Video decoding method and device
CN104380741A (en) Reference pixel reduction for intra lm prediction
CN102137263A (en) Distributed video coding and decoding methods based on classification of key frames of correlation noise model (CNM)
CN104255033A (en) Reduced look-up table for lm mode calculation
CN103916675B (en) A kind of low latency inner frame coding method divided based on band
CN101742288B (en) Video noise reduction encoding method and video noise reduction encoding device
CN101742290B (en) Methods and devices for denoising in video coding and decoding,
CN105306957A (en) Adaptive loop filtering method and device
CN100502507C (en) Video data compaction coding method
CN108063947B (en) Lossless reference frame compression method based on pixel texture
CN100493191C (en) Video encoder with low complexity noise reduction and video encoding method
CN117480778A (en) Residual coding and video coding methods, devices, equipment and systems
CN108347602A (en) Method and apparatus for lossless compression video data
CN102026002B (en) Downsampling transcoding method and device of frame rate, vector reconstruction method and device thereof
CN1194544C (en) Video encoding method based on prediction time and space domain conerent movement vectors
CN101115200A (en) Effective motion vector telescopic encoding method
CN112422989B (en) Video coding method
CN104602001A (en) Low-bit-rate video coding method based on frame difference image
CN111757126A (en) VR-oriented real-time image compression method, system and storage medium
CN113163199B (en) H265-based video rapid prediction method, rapid coding method and system

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
CB03 Change of inventor or designer information

Inventor after: Liu Pengyu

Inventor after: Gao Yuan

Inventor after: Jia Kebin

Inventor before: Liu Pengyu

Inventor before: Jia Kebin

COR Change of bibliographic data
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20150506