CN108271020B - A kind of panoramic video quality evaluating method of view-based access control model attention model - Google Patents
A kind of panoramic video quality evaluating method of view-based access control model attention model Download PDFInfo
- Publication number
- CN108271020B CN108271020B CN201810372074.6A CN201810372074A CN108271020B CN 108271020 B CN108271020 B CN 108271020B CN 201810372074 A CN201810372074 A CN 201810372074A CN 108271020 B CN108271020 B CN 108271020B
- Authority
- CN
- China
- Prior art keywords
- video
- distortion
- panoramic video
- latitude
- model
- 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.)
- Active
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N17/00—Diagnosis, testing or measuring for television systems or their details
- H04N17/004—Diagnosis, testing or measuring for television systems or their details for digital television systems
Abstract
The present invention relates to a kind of panoramic video quality evaluating methods of view-based access control model attention model, by carrying out band division to original video, coding distortion is introduced to different bands respectively, generate distortion video, then subjective quality marking is carried out to distortion video, visual attention model is constructed according to the latitude of subjective quality scores and distortion band, the total quality evaluation index of panoramic video is obtained after weighting.The present invention solves attention distribution problem when user watches panoramic video, improves the accuracy of panoramic video quality evaluation.
Description
Technical field
The present invention relates to computer digit technical field of video image processing more particularly to a kind of view-based access control model attention moulds
The panoramic video quality evaluating method of type.
Background technique
Existing most of algorithms are all only applicable to the quality evaluation of ordinary video, and the quality evaluation of panoramic video is related to
Foot is very few, and with the rapid development of virtual reality technology, the quality evaluation of the panoramic video of strong applicability is only development trend.
Summary of the invention
Insufficient for the above-mentioned prior art, the panoramic video quality that the present invention provides a kind of view-based access control model attention model is commented
Valence method, mainly solves attention distribution problem when user watches panoramic video, and panoramic video quality is precisely predicted in realization
Purpose.
To achieve the above object, the technical scheme is that a kind of panoramic video matter of view-based access control model attention model
Measure evaluation method, comprising the following steps:
Step S1: by analyze original video enter encoder anterior-posterior projection transformation corresponding relationship, to original video into
Band of the row based on latitude divides, and guarantees that band has equal arc length;
Coding distortion: being introduced different bands by step S2 respectively, is generated distortion video, is collectively constituted entirely with original video
Scape video database;
Step S3: subjective assessment is carried out to distortion video, the subjective quality scores of distortion video are obtained, according to subjectivity
The latitude building distortion sensitivity model of mass fraction and distortion band;
Step S4: visual attention model of the building based on latitude;
Step S5: the visual attention model based on latitude is weighting on Y-PSNR, obtains the whole of panoramic video
Weight evaluation index.
Further, in the step S1,9 bands are divided into original video, latitude value is respectively+90 ,+70 ,+
50、+30、+10、-10、-30、-50、-70、-90。
Further, the step S2 specifically: using damaging video encoder with several quantization parameter values to original
Video is compressed, and quantization parameter value is distributed to different bands under equal matrix projection, coding distortion is introduced in band,
It is distorted video to generate, collectively constitutes panoramic video database with original video.
Further, sensitivity model is distorted in the step S3 are as follows:
Wherein, DsTo be distorted sensitivity, DMOS is the practical subjective matter that the full marks of subjective quality scores subtract distortion video
Difference obtained by score is measured, Δ Q is the average quantiser step size being distorted in band.
Further, the step S4 specifically:
Randomly choosed in panoramic video database using the method that reserves 60% training sample and other 40% samples into
Row test, it is as follows to obtain the visual attention model based on latitude:
Wherein,For the weight matrix form of visual attention model,For fitting parameter,It indicates
The resolution ratio of original video, j are the ordinate value for being distorted pixel (i, j) in video frame.
Further, the step S5 is specifically included:
It is as follows to calculate 360 degree of mean square error MSE:
It is as follows to calculate Y-PSNR PSNR:
Wherein,WithIt is the pixel value of the frame of original video and distortion video respectively,It is view
The maximum possible brightness degree of frequency frame, PSNR value is bigger, and the total quality for representing panoramic video is better.
Compared with prior art, the present invention has the utility model has the advantages that solving attention distribution when user watches panoramic video
Problem realizes the accurate prediction and evaluation to panoramic video quality.
Detailed description of the invention
Fig. 1 is a kind of flow diagram of the panoramic video quality evaluating method of view-based access control model attention model of the present invention;
Fig. 2 is the band dividing condition schematic diagram of original video in one embodiment of the invention;
Fig. 3 is the relation schematic diagram of subjective quality scores and latitude value in one embodiment of the invention;
Fig. 4 is the schematic diagram that sensitivity model is distorted in one embodiment of the invention.
Specific embodiment
The present invention will be further described with reference to the accompanying drawings and embodiments.
As shown in Figure 1, a kind of panoramic video quality evaluating method of view-based access control model attention model, comprising the following steps:
Step S1: by analyze original video enter encoder anterior-posterior projection transformation corresponding relationship, to original video into
Band of the row based on latitude divides, and guarantees that band has equal arc length;
Coding distortion: being introduced different bands by step S2 respectively, is generated distortion video, is collectively constituted entirely with original video
Scape video database;
Step S3: subjective assessment is carried out to distortion video, the subjective quality scores of distortion video are obtained, according to subjectivity
The latitude building distortion sensitivity model of mass fraction and distortion band;
Step S4: visual attention model of the building based on latitude;
Step S5: the visual attention model based on latitude is weighting on Y-PSNR, obtains the whole of panoramic video
Weight evaluation index.
In the present embodiment, as shown in Fig. 2, being divided into 9 bands to original video, latitude value is respectively+90 ,+70 ,+
50、+30、+10、-10、-30、-50、-70、-90。
It is compressed based on HEVC with three quantization parameter (QP) values, 35,40 and 45 pairs of videos, it will under equal matrix projection
These QP values distribute to 9 different bands.Fig. 2 shows the latitude variation of 9 different frequency ranges.It is introduced in these bands
Coding distortion to generate some distortion videos, while keeping wherein almost complete overall objective distortion.Track is dropped to keep away
Exempt from the influence of acoustic information.The original video of 5 different contents is chosen, therefore, a total of 5 original series and 135 are impaired
Sequence (5 sequence × 3 QP × 9 band), finally establishes the database of 140 panoramic video.
Subjective assessment is carried out to distortion video, for panoramic video random display on the same screen, experimenter will
Original video point is given a mark as reference pair remaining distortion video according to 0-5.The subjective quality of panoramic video point in database
Number is indicated by the Mean Opinion Score number of 23 experimenters.As shown in figure 3, being the latitude of average subjective quality scores and distortion
The relationship of value.
It is distorted sensitivity model are as follows:
Wherein, DsTo be distorted sensitivity, DMOS is that the full marks 5 of subjective quality scores divide the practical subjectivity for subtracting distortion video
Difference obtained by mass fraction, Δ Q are the average quantiser step size being distorted in band.Distortion sensitivity model based on latitude, it can
To determine the average quality of panoramic video.This distortion sensitivity table is shown as the objective distortion of different zones to whole master by us
The influence of appearance quality.
Distortion sensitivity model by its maximum value in all latitudes as shown in figure 4, be normalized, definition latitude
Degree is 1 for 0 sensitivity, other latitudes between 0~1.It is observed that above-mentioned trend is basic for different QP
It is similar, therefore the influence of QP can be eliminated.This has given full expression to influence of the latitude to perceptual distortion, helps to establish based on latitude
Visual attention model, it follows that formula speculate it close to cos function, then by fitting, obtain vision attention
Power model.The training sample and other 40% samples for being randomly choosed 60% in panoramic video database using the method that reserves are carried out
Test, above-mentioned processing have carried out five times, and it is as follows to obtain the visual attention model based on latitude:
Wherein,For the weight matrix form of visual attention model,For fitting parameter,It indicates
The resolution ratio of original video, j are the ordinate value for being distorted pixel (i, j) in video frame.Each pixel in frameIt can be by
It is calculated as value corresponding with weight matrix.In the present embodiment, the average value of the parameter a obtained from each training group is
0.4618。
Final quality evaluation algorithm is constructed by the weighting of visual attention model, the vision based on distortion sensitivity
Attention model can be used for being weighted local quality index to generate total quality index.By the vision attention based on latitude
Power model-weight calculates 360 degree of mean square errors in the whole measurement standard for obtaining panoramic video on Y-PSNR (PSNR)
Poor MSE is as follows:
It is as follows to calculate Y-PSNR PSNR:
Wherein,WithIt is the pixel value of the frame of original video and distortion video respectively,It is view
The maximum possible brightness degree of frequency frame, PSNR value is bigger, and the total quality for representing panoramic video is better.
Inventive method is verified:
We verify the present invention by calculating the related coefficient between the mass value that subjective quality scores and algorithm obtain
The performance of algorithm, wherein related coefficient includes Pearson's linearly dependent coefficient (Pearson Linear Correlation
Coefficient, PLCC), Spearman rank correlation coefficient (Spearman Rank-order Correlation
Coefficient, SRCC) and Ken Deer rank related coefficient (Kendall Rank-order Correlation
Coefficient, KRCC), for the value of related coefficient closer to 1, the accuracy of representative model is higher.Experimental result is as shown in table 1.
Table 1
The foregoing is only a preferred embodiment of the present invention, but scope of protection of the present invention is not limited thereto,
Anyone skilled in the art in the technical scope disclosed by the present invention, according to the technique and scheme of the present invention and its
Inventive concept is subject to equivalent substitution or change, should be covered by the protection scope of the present invention.
Claims (5)
1. a kind of panoramic video quality evaluating method of view-based access control model attention model, which comprises the following steps:
Step S1: entering the corresponding relationship of encoder anterior-posterior projection transformation by analyzing original video, carries out base to original video
It is divided in the band of latitude, guarantees that band has equal arc length;
Coding distortion: being introduced different bands by step S2 respectively, is generated distortion video, is collectively constituted aphorama with original video
Frequency database;
Step S3: subjective assessment is carried out to distortion video, the subjective quality scores of distortion video are obtained, according to subjective quality
The latitude building distortion sensitivity model of score and distortion band;
Step S4: visual attention model of the building based on latitude;The step S4 specifically:
60% training sample is randomly choosed in panoramic video database using the method that reserves and other 40% samples are surveyed
Examination, it is as follows to obtain the visual attention model based on latitude:
Wherein, Ds(i, j) is the weight matrix form of visual attention model, and a is fitting parameter, and M × N indicates original video
Resolution ratio, j are the ordinate value for being distorted pixel (i, j) in video frame;
Step S5: the visual attention model based on latitude is weighting on Y-PSNR, obtains the whole matter of panoramic video
Measure evaluation index.
2. panoramic video quality evaluating method according to claim 1, which is characterized in that in the step S1, to original
Video is divided into 9 bands, and latitude value is respectively+90 ,+70 ,+50 ,+30 ,+10, -10, -30, -50, -70, -90.
3. panoramic video quality evaluating method according to claim 1, which is characterized in that the step S2 specifically: make
Original video is compressed with several quantization parameter values with video encoder is damaged, by quantization parameter under equal matrix projection
Value distributes to different bands, and coding distortion is introduced in band, to generate distortion video, collectively constitutes panorama with original video
Video database.
4. panoramic video quality evaluating method according to claim 1, which is characterized in that be distorted in the step S3 sensitive
Spend model are as follows:
Wherein, DsTo be distorted sensitivity, DMOS is the practical subjective quality scores that the full marks of subjective quality scores subtract distortion video
Gained difference, Δ Q are the average quantiser step size being distorted in band.
5. panoramic video quality evaluating method according to claim 4, which is characterized in that the step S5 is specifically included:
It is as follows to calculate 360 degree of mean square error MSE:
It is as follows to calculate Y-PSNR PSNR:
Wherein, y (i, j) and y ' (i, j) is the pixel value of the frame of original video and distortion video, MAX respectivelyIBe video frame most
Brightness degree possible greatly, PSNR value is bigger, and the total quality for representing panoramic video is better.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810372074.6A CN108271020B (en) | 2018-04-24 | 2018-04-24 | A kind of panoramic video quality evaluating method of view-based access control model attention model |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810372074.6A CN108271020B (en) | 2018-04-24 | 2018-04-24 | A kind of panoramic video quality evaluating method of view-based access control model attention model |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108271020A CN108271020A (en) | 2018-07-10 |
CN108271020B true CN108271020B (en) | 2019-08-09 |
Family
ID=62777740
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810372074.6A Active CN108271020B (en) | 2018-04-24 | 2018-04-24 | A kind of panoramic video quality evaluating method of view-based access control model attention model |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108271020B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110930370B (en) * | 2019-11-04 | 2023-03-31 | 北京工业大学 | Panoramic image quality evaluation method |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106657980A (en) * | 2016-10-21 | 2017-05-10 | 乐视控股(北京)有限公司 | Testing method and apparatus for the quality of panorama video |
CN107347153A (en) * | 2016-10-20 | 2017-11-14 | 中国科学技术大学 | The quality evaluating method of longitude and latitude figure panoramic video |
CN107454418A (en) * | 2017-03-03 | 2017-12-08 | 叠境数字科技(上海)有限公司 | 360 degree of panorama video code methods based on motion attention model |
CN107483920A (en) * | 2017-08-11 | 2017-12-15 | 北京理工大学 | A kind of panoramic video appraisal procedure and system based on multi-layer quality factor |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US10339688B2 (en) * | 2016-07-28 | 2019-07-02 | Cyberlink Corp. | Systems and methods for rendering effects in 360 video |
-
2018
- 2018-04-24 CN CN201810372074.6A patent/CN108271020B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107347153A (en) * | 2016-10-20 | 2017-11-14 | 中国科学技术大学 | The quality evaluating method of longitude and latitude figure panoramic video |
CN106657980A (en) * | 2016-10-21 | 2017-05-10 | 乐视控股(北京)有限公司 | Testing method and apparatus for the quality of panorama video |
CN107454418A (en) * | 2017-03-03 | 2017-12-08 | 叠境数字科技(上海)有限公司 | 360 degree of panorama video code methods based on motion attention model |
CN107483920A (en) * | 2017-08-11 | 2017-12-15 | 北京理工大学 | A kind of panoramic video appraisal procedure and system based on multi-layer quality factor |
Non-Patent Citations (2)
Title |
---|
A SUBJECTIVE VISUAL QUALITY ASSESSMENT METHOD OF PANORAMIC VIDEOS;Mai Xu ET Al;《Proceedings of the IEEE International Conference on Multimedia and Expo (ICME) 2017》;20170714;第517-522页 * |
Modeling Attention in Panoramic Video:A Deep Reinforcement Learning Approach;Yuhang Song ET AL;《JOURNAL OF LATEX CLASS FILES》;20150831;第14卷(第8期);第1-14页 * |
Also Published As
Publication number | Publication date |
---|---|
CN108271020A (en) | 2018-07-10 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
KR100731358B1 (en) | Method and system for measuring the video quality | |
CN105100789B (en) | A kind of method for evaluating video quality | |
CN109389591B (en) | Color descriptor-based color image quality evaluation method | |
CN102663745B (en) | Color fusion image quality evaluation method based on vision task. | |
Watson et al. | Measurement of visual impairment scales for digital video | |
CN104572538B (en) | A kind of Chinese medicine tongue image color correction method based on K PLS regression models | |
CN101562675B (en) | No-reference image quality evaluation method based on Contourlet transform | |
CN109191460B (en) | Quality evaluation method for tone mapping image | |
Saviane et al. | The red giant branches of Galactic globular clusters in the [(VI) o, Mv] plane: metallicity indices and morphology | |
CN107464222B (en) | Based on tensor space without reference high dynamic range images method for evaluating objective quality | |
CN106548472A (en) | Non-reference picture quality appraisement method based on Walsh Hadamard transform | |
Pulsifer et al. | An objective fingerprint quality-grading system | |
CN101976444A (en) | Pixel type based objective assessment method of image quality by utilizing structural similarity | |
Jiang et al. | Tone-mapped image quality assessment for electronics displays by combining luminance partition and colorfulness index | |
CN108271020B (en) | A kind of panoramic video quality evaluating method of view-based access control model attention model | |
CN106412571A (en) | Video quality evaluation method based on gradient similarity standard deviation | |
CN109191386A (en) | A kind of quick Gamma bearing calibration and device based on BPNN | |
CN107018410A (en) | A kind of non-reference picture quality appraisement method based on pre- attention mechanism and spatial dependence | |
CN107945154A (en) | Color image quality evaluation method based on quaternary number discrete cosine transform | |
WO2016123815A1 (en) | Evaluation method and system based on pet/ct image quality objective algorithm | |
CN101895787B (en) | Method and system for subjectively evaluating video coding performance | |
CN113222902B (en) | No-reference image quality evaluation method and system | |
Bernardo et al. | A study on the user perception to color variations | |
CN104915959A (en) | Aerial photography image quality evaluation method and system | |
CN107845089A (en) | A kind of measuring method based on colouring information, apparatus and system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |