CN107194927A - The measuring method of stereo-picture comfort level chromaticity range based on salient region - Google Patents
The measuring method of stereo-picture comfort level chromaticity range based on salient region Download PDFInfo
- Publication number
- CN107194927A CN107194927A CN201710442862.3A CN201710442862A CN107194927A CN 107194927 A CN107194927 A CN 107194927A CN 201710442862 A CN201710442862 A CN 201710442862A CN 107194927 A CN107194927 A CN 107194927A
- Authority
- CN
- China
- Prior art keywords
- stereo
- picture
- notable
- image
- dimensional
- 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
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/12—Edge-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
- G06T2207/10021—Stereoscopic video; Stereoscopic image sequence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30168—Image quality inspection
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
- Processing Or Creating Images (AREA)
Abstract
The invention belongs to image processing field, to provide appropriate method, the comfort level of stereo-picture can be reflected well, is made for stereo content and stronger technical support is provided.For this, the present invention, the measuring method of stereo-picture comfort level chromaticity range based on salient region, the method being first combined using disparity map and two dimensional image notable figure obtains three-dimensional saliency map, recycle fuzzy membership and mask to optimize it and draw final notable stereo-picture, and utilize the correctness of the three-dimensional significantly algorithm extraction marking area of eye tracker experimental verification；Then subjective experiment is carried out using approximatioss step by step from coarse to fine, experimental data is fitted by least square method, the comfortable chromaticity match figure and comfortable colourity disparity map of notable stereo-picture are obtained, and the subjective feeling that image is tested under different scenes is analyzed.Present invention is mainly applied to image procossing occasion.
Description
Technical field
The invention belongs to image processing field, it is related to the Elements research that influence stereo-picture watches comfort level, especially relates to
And a kind of comfortable colourity Quantitative research method based on stereo-picture marking area.Concretely relate to based on salient region
The measuring method of stereo-picture comfort level chromaticity range.
Background technology
In recent years, three-dimensional stereo display technique develops rapidly in scientific research and business activity, particularly in business stereoscopic electric
Immense success has been obtained on shadow, but some spectators are when watching stereoscopic picture plane, it may appear that different degrees of vision is uncomfortable existing
As[1-5].It is the major defect of current parallax type 3D films and TV that vision is uncomfortable, if the problem cannot be solved, three-dimensional
Display Technique is difficult to expanded[6].Uncomfortable vision is a kind of subjective feeling, is generally weighed with visual comfort[5].Vision
Uncomfortable main source has：Display device in itself, four aspects such as stereo content, the Physiological Psychology of people and viewing environment.
It is the key factor for restricting stereoscopic imaging technology popularization that stereo content, which lacks,[7], therefore, make abundant stereo content and seem outstanding
To be important.It is well known that the influence uncomfortable factor of stereo content has a lot, such as brightness, colourity, saturation degree, contrast, string
Disturb and the factor such as parallax.If the quantitative criteria on each influence factor can be obtained, it will be provided by force for the making of stereo content
Favourable technical support.Experiment herein is exactly the vision noticing mechanism with reference to human eye, and quantitative study is three-dimensional interior as caused by colourity
The problem of whether comfortable holding.
So far, existing many experts and scholars both at home and abroad are ground to the influence factor of stereoscopic image content comfort level
Study carefully.Abroad, in document [8], Frank L.K are by testing systematically qualitative research brightness, colourity, contrast, string
Disturb, the influence of the various factors to binocular stereo image comfort level such as parallax, test result indicates that, if binocular view mismatches journey
Degree, which exceedes certain thresholding, will seriously reduce stereo image vision comfort level.Document [9] points out that big parallax be able to can strengthen
Third dimension, but it is uncomfortable to also often result in vision.A kind of algorithm is further devised in the document to allow the viewer to enough adjust automatically
Whole parallax size obtains more comfortable stereo-picture.At home, document [10] qualitatively have studied colourity, brightness, contrast
Degree and 4 parameters of resolution ratio influence to double vision point stereo imaging system display effect, and analytic explanation Parameters variation pair
The reason for imaging effect is impacted.Document [11] systematically elaborate to watch stereo-picture cause the physiology of visual fatigue because
Element, it is indicated that one of the reason for causing visual fatigue is color and the excessive left and right view of luminance difference not to be occurred not in beholder's eye
With anaglyph pair.Document [12] points out that International Organization for standardization once issued stereo-picture, Video security file IWA3：
Image safety, kopiopia can be caused or induce illness in eye by being expressly recited the excessive stereo-picture of binocular disparity.
At present, mostly it is qualitatively, and without unified evaluation criteria to the research of stereo-picture comfort level both at home and abroad.For
More directly, more accurately judge the given stereo image vision comfort level whether, it is necessary to carry out quantitative study.Part document pair
The various factors of influence viewing comfort level has carried out quantitative study, and document [13] [14] [15] [16] [17] is to influence stereo-picture
The factors such as brightness, saturation degree, colourity, contrast and the parallax of comfort level are quantitatively studied.But, such method is all
Studied based on view picture stereo-picture.However, document [18] shows that the key property of human visual system is vision note
Meaning, and human eye vision often only focuses on the area-of-interest of scene.Document [19] shows that human eye often more inclines when observing image
To in observing pith therein rather than entire image, saliency region is an important spy of human eye vision
Property, the characteristic has been supplied in many aspects, and achieves good result.
The content of the invention
To overcome the deficiencies in the prior art, the present invention is intended to provide appropriate method, stereo-picture can be reflected well
Comfort level, for stereo content make stronger technical support is provided.Therefore, the technical solution adopted by the present invention is, it is based on
The measuring method of the stereo-picture comfort level chromaticity range of salient region, is first mutually tied using disparity map with two dimensional image notable figure
The method of conjunction obtains three-dimensional saliency map, recycles fuzzy membership and mask to optimize it and draws final notable stereogram
Picture, and utilize the correctness of the three-dimensional significantly algorithm extraction marking area of eye tracker experimental verification；Then using it is from coarse to fine by
Level approximatioss carries out subjective experiment, and experimental data is fitted by least square method, the comfortable of notable stereo-picture is obtained
Chromaticity match figure and comfortable colourity disparity map, and the subjective feeling that image is tested under different scenes is analyzed.
Obtain comprising the concrete steps that for three-dimensional saliency map, two dimensional image notable figure passes through image brightness in itself, colourity, right
Calculated than degree and spatial frequency, specific calculated using GBVS algorithms obtains right viewpoint plane notable figure, is designated as SMR(x,y)；
Use Fast Stereo Matching Algorithm to obtain the disparity map on the basis of right viewpoint, be designated as dR(x,y)；Using linear mode by plane
Visual saliency map SMR(x, y) and right anaglyph dR(x, y), which is weighted, obtains three-dimensional saliency map I (x, y), such as formula (1)：
I (x, y)=w1dR(x,y)+w2SMR(x,y) (1)
In formula：I (x, y) is three-dimensional saliency map；w1And w2To weight proportion, and w1+w2=1.
Comprising the concrete steps that for the acquisition of notable stereo-picture, the feature letter of image is described using the theory of fuzzy mathematics
Breath, three-dimensional saliency map is optimized, and the notable figure after optimize is bianry image, referred to as mask images, pixel
Pixel value shows that the point in former stereo-picture belongs to marking area for 1, otherwise belongs to non-significant region.
The further detailed process of the acquisition of notable stereo-picture is as follows：
If domain X is three-dimensional saliency map, the element in domain is divided into two classes, marking area A and non-significant region B, A
It is an X division with B, i.e.,：
A ∪ B=X
A ∩ B=φ
The gray value of three-dimensional saliency map represents that the pixel of this in original image belongs to the degree of marking area, therefore directly
Membership function, such as formula (2) are provided, A (x, y) represents the degrees of membership of the gray value L to marking area A at pixel (x, y) place.
Using the method deblurring of Threshold segmentation, segmentation threshold T is determined by maximum between-cluster variance method, is gone by the threshold value
It is fuzzy to obtain mask images M (x, y)[28], such as formula (3)：
For any pixel point (x, y) in three-dimensional saliency map I (x, y), if A (x, y) is subordinate to marking area
Degree is more than threshold value T, then the white portion that the pixel belongs in vision significance region, correspondence mask images M (x, y), otherwise
Belong to black region.In order to remove the burr and cavitation at mask image edge, using morphologic opening and closing operation to mask
Image M (x, y), which is optimized, to be obtained optimizing mask image M'(x, y)；Finally will optimization mask images and original visual point image phase
It is multiplied to arrive notable stereo-picture.
Subjective measurement is carried out to the marking area of source stereo-picture using eye tracker, calibration experiments are first carried out before experiment every time
And reference measurement, tested when subject is less than 1 degree to the observation error for calibrating datum mark, per pictures random presentation and
The presentation time is 5s.The thermal map for the interesting image regions measured according to eye tracker, with notable stereo-picture regional correlation.
Data processing is carried out using approximatioss step by step from coarse to fine, step-length is approached into series and is divided into 3 grades, first order step-length
It is respectively 75,15 and 3 to third level step-length, linear transformation is carried out to three-dimensional image chroma using three-level step-length, wherein, only to a left side
View carries out the 2nd grade of step-length segmentation, and the 1st grade, the 2nd grade, the segmentation of 3rd level step-length are carried out to right view；Experiment is used
Acquisitions and at different levels colourity change of the MATLAB2014b to notable stereo-picture are handled, and process step is as follows：
A. source stereo-picture is chosen, three-dimensional saliency map is obtained and finally gives notable stereo-picture；
B. notable stereo-picture is subjected to color notation conversion space, hsv color space is transformed into by rgb space, extract image
Colourity H.Notable stereo-picture in a determines the notable and non-significant part of image.Marking area is in the stereo-picture of source
Corresponding pixel, non-significant part is covered with gray scale for 255 image；
C. line translation is entered with step-length 15 to notable stereo-picture left view colourity, obtains 24 width left views, to significantly three-dimensional
Image right view colourity enters line translation with step-length 72, obtains 5 width right views；Can be obtained after combination of two 120 it is to be measured significantly
Stereo-picture；
D. subjective scoring is carried out to the notable stereo-picture of 120 width in step c, chosenThe image divided is comfortable to meet
Desired notable stereo-picture is spent, is recorded respectivelyRight view minimax chromatic value n1rup、n1rdownAnd recordRight view minimax chromatic value n1rup'、n1rdown'；
E. to chromatic value range n1 in step drup、n1rup' and n1rdown、n1rdown' carry out the second level step change, obtain
240 notable stereo-pictures to be measured can be obtained to 10 two grades of right views, therefore after the view combination of two of left and right；
F. subjective scoring is carried out to the notable stereo-picture of 240 width in step e, chosenThe image divided is comfortable to meet
Desired notable stereo-picture is spent, is recorded respectivelyRight view minimax chromatic value n2rup、n2rdownAnd recordRight view minimax chromatic value n2rup'、n2rdown'；
G. similarly, to chromatic value range n2 in step frup、n2rup' and n2rdown、n2rdown' carry out the third level step-length become
Change, obtain 12 width three-level right views, therefore 288 notable stereo-pictures to be measured can be obtained after the view combination of two of left and right；
H. subjective scoring is carried out to the notable stereo-picture of 288 width in step g, chosenThe image divided relaxes to meet
The notable stereo-picture that appropriateness is required, is recorded respectivelyRight view minimax chromatic value n3rup、n3rdown。
The features of the present invention and beneficial effect are：
Document [14] shadow of quantitative study colourity factor to visual comfort in the case where not accounting for vision significance
Ring, such as figure (10) as can be seen that broken line surround chromaticity range it is basically identical, illustrate marking area comfort level can represent it is whole
The comfort level of width stereo-picture, it is consistent with theory.Document [14] is given original image, is allowed using double experimental methods stimulated
Whether subject is comfortably evaluated test image, can there is certain limitation to the subjectivity of subject.Due to being to have reference
Subjective experiment, so for some scenes, its left view colourity interval (0 °, 360 °), right view should not have comfortably
Colourity interval is matching.Herein using single experimental method stimulated, double stimulating methods compared with document [14] more meet human eye and regarded
The subjectivity of feel, its left view colourity interval (0 °, 360 °) right view has certain comfortable colourity interval, and scene is different
The comfortable interval size of colourity is also different.The conclusion drawn herein has authenticity, can be more provided with for the making of stereo content
The technical support of profit, rational scheme is provided for the making of stereo content comfort standard.
Brief description of the drawings：
The width source stereo-pictures of Fig. 14.In figure, a schemes for table tennis, and b is billboard, child's c figure, d flower figures.
Fig. 2 algorithm flow charts.
The notable stereogram extraction process figures of Fig. 3.A is right view in figure, and b disparity maps, c is that plane shows figure, and d is three-dimensional aobvious
Diagram, e mask images, f optimization mask images, the notable stereo-pictures of g.
Fig. 4 eye trackers test notable thermal map.
The notable stereo-pictures of Fig. 5.
Tetra- groups of stereo image vision comfort level matching figures of Fig. 6.A schemes for table tennis in figure, and b is billboard, child's c figure, d flowers
Figure.
Tetra- groups of stereo image vision comfort level disparity maps of Fig. 7.A schemes for table tennis in figure, and b is billboard, child's c figure, d flowers
Figure.
The comfortable chromaticity match figure of Fig. 8 stereo-pictures.
The comfortable colourity disparity map of Fig. 9 stereo-pictures.
Figure 10 this paper experimental results and document [14] Comparative result.
The width source stereo-pictures of Figure 11 2.A schemes for table tennis in figure, and b is billboard.
Embodiment
Combine shadow of the human eye vision attention mechanism quantitative study colourity factor to three-dimensional image-watching comfort level first herein
Ring.The method that the work is combined using disparity map and two dimensional image notable figure first obtains three-dimensional saliency map, aobvious to solid
Work degree figure, which is optimized, obtains notable stereo-picture, and utilizes the three-dimensional notable algorithm of eye tracker experimental verification to extract marking area
Correctness；Then subjective experiment is carried out using approximatioss step by step from coarse to fine, experimental data entered by least square method
Row fitting, obtains the comfortable chromaticity match figure and comfortable colourity disparity map of notable stereo-picture, and to being attempted under different scenes
The subjective feeling of picture is analyzed.Test result indicates that, the quantizing range of the comfortable colourity obtained according to context of methods is more accorded with
Close the result of human eye vision subjective observation.
Vision noticing mechanism is combined first herein, by a large amount of subjective experiments quantitatively to influence stereo-picture euphorosia
The colourity factor of degree is studied.First, three-dimensional saliency map is obtained with reference to disparity map and plane notable figure, recycles fuzzy be subordinate to
Category degree and mask optimize to it draws final notable stereo-picture.Using reasonability of the eye tracker to the notable stereo-picture of gained
Verified；Then, experimental data is obtained using approximatioss from coarse to fine and carries out subjective experiment, obtained different scenes and significantly stand
The comfortable chromaticity match figure and disparity map of body image.Test result indicates that, the comfortable colourity interval of left and right view can be with different fields
Scape and it is different, the average maximum of the comfortable color difference value of binocular view can be 122.5o, i.e. left and right view colourity difference can not be excessive.
The comfortable chromaticity range of gained reflects the comfort level of stereo-picture well herein, is provided for stereo content making stronger
Technical support.
Vision noticing mechanism quantitative study colourity factor is combined to three-dimensional image-watching comfort level the invention provides one kind
Method.First, three-dimensional saliency map is obtained with reference to disparity map and plane notable figure, recycles fuzzy membership and mask to it
Optimization draws final notable stereo-picture.The reasonability of the notable stereo-picture of gained is verified using eye tracker；Then,
Experimental data is obtained using approximatioss from coarse to fine and carries out subjective experiment, the comfortable color of the notable stereo-picture of different scenes is obtained
Degree matching figure and disparity map.
Experiment flow of the present invention is the extraction of 1) three-dimensional saliency map.Two dimensional image notable figure and stereo image parallax figure knot
Conjunction obtains three-dimensional saliency map.2) the significantly acquisition of stereo-picture.Processing, three-dimensional significance are optimized to three-dimensional saliency map
Figure obtains notable stereo-picture by fuzzy theory processing and mask optimization.3) subjective experiment.Notable stereo-picture is subjected to color
Degree classification step-length converts and carries out subjective experiment.4) comfortable chromaticity match figure and comfortable colourity disparity map are obtained.Experiment flow figure
As shown in Figure 2.
Detailed analysis will be carried out to each step below：
1. the acquisition of three-dimensional saliency map
Stereo-picture has more information content than one-view image, and human eye can not possibly match all in a short time
Edge feature, most people only focuses on those " important areas ", then extracts the border of object in these regions, and finally match
These borders form stereoscopic vision.According to the stereoscopic vision attention characteristic of human eye, observer can be in marking area in image
Appearance is more paid close attention to[25], it is therefore possible to use the comfort level of significantly stereo-picture reflects the comfort level feelings of view picture stereo-picture
Condition, to improve test accuracy rate and reduce computation complexity.The parallax characteristic and spatial frequency of stereo-picture can all influence vision
Comfort level[26].Therefore, three-dimensional saliency map is finally obtained by stereo image parallax figure and two dimensional image notable figure.
The factors such as brightness, colourity, contrast and spatial frequency of the two dimensional image notable figure by image in itself are calculated.
GBVS algorithms are used herein[27]Calculating obtains right viewpoint plane notable figure, is designated as SMR(x,y).Matched and calculated using quick stereo
Method[28]The disparity map on the basis of right viewpoint is obtained, d is designated asR(x,y).Using linear mode by plane visual notable figure SMR(x,
Y) with right anaglyph dR(x, y), which is weighted, obtains three-dimensional saliency map I (x, y), such as formula (1)：
I (x, y)=w1dR(x,y)+w2SMR(x,y) (1)
In formula：I (x, y) is three-dimensional saliency map；w1And w2To weight proportion, and w1+w2=1；Take w herein1=w2=
0.5[29]。
2. the acquisition of notable stereo-picture
Because the obtained gray value L of three-dimensional saliency map is between 0-255, i.e. L ∈ (0,255), the gray scale point of image
The information such as cloth and marking area edge all has ambiguity, it is impossible to be directly clearly divided into significantly according to this three-dimensional saliency map
Region and non-significant region, it is therefore desirable to determine a threshold value, enable to be divided gray scale saliency map.Paper is utilized
The theory of fuzzy mathematics describes the characteristic information of image[30], three-dimensional saliency map is optimized, it is notable after being optimized
Figure is bianry image, referred to as mask images[26], the pixel value of pixel shows that the point in former stereo-picture belongs to notable for 1
Region, on the contrary belong to non-significant region.Detailed process is as follows：
Assuming that domain X is three-dimensional saliency map, the element in domain is divided into two classes, marking area A and non-significant region
B, A and B are an X divisions, i.e.,：
A ∪ B=X
A ∩ B=φ
The gray value of three-dimensional saliency map represents that the pixel of this in original image belongs to the degree of marking area, therefore can be with
Membership function, such as formula (2) are directly given, A (x, y) represents that the gray value L at pixel (x, y) place is subordinate to marking area A
Degree.
Because final goal is to obtain marking area A, it is therefore desirable to deblurring[30].The method for using Threshold segmentation herein,
Segmentation threshold T is determined by maximum between-cluster variance method[31].Mask images M (x, y) is obtained by the threshold value deblurring[28], it is such as public
Formula (3)：
For any pixel point (x, y) in three-dimensional saliency map I (x, y), if A (x, y) is subordinate to marking area
Degree is more than threshold value T, then the white portion that the pixel belongs in vision significance region, correspondence mask images M (x, y), otherwise
Belong to black region.In order to remove the burr and cavitation at mask image edge, using morphologic opening and closing operation to mask
Image M (x, y), which is optimized, to be obtained optimizing mask image M'(x, y).
Finally optimization mask images are multiplied with original visual point image and obtain notable stereo-picture, shown in such as Fig. 3 (a-g).
Fig. 3 sets forth 4 groups of right viewpoints of stereo-picture, disparity map, plane notable figure, three-dimensional saliency map, mask image, optimization and cover
Film image and notable stereo-picture.
3. eye tracker experimental verification
In order to verify whether the salient region of significantly stereo-picture obtained by this paper algorithms is true value, herein using eye tracker
Subjective measurement is carried out to the marking area of source stereo-picture.Calibration experiments and reference measurement are first carried out before experiment every time, works as subject
Person tests when being less than 1 degree to the observation error for calibrating datum mark.Presented at random per pictures and the time is presented for 5s.According to
The thermal map for the interesting image regions that eye tracker is measured, finds that the two is carried with this paper notable stereo-picture regional correlation
The marking area gone out is basically identical, and Fig. 4 is the thermal map for the interesting image regions that eye tracker is measured.
Eye tracker experiment shows the time of blinkpunkt and the dynamic change of position with color warm colour degree, in Fig. 4 mark for
It is red, the region of yellow shows that fixation time is most long, is also part the most significant.Green Marker part represents more important,
Subject is more interested in this part in an experiment；Blue markings part represents that significance level is relatively low, but is tested in an experiment
Person is also paid close attention to blue markings part；Unmarked part represents that subject pays close attention to seldom within the experimental stage to this part
Or be not concerned with, represent non-significant region.
Discovery is contrasted by the three-dimensional notable figure that Fig. 4 eye trackers are tested to notable thermal map and Fig. 3, it is vertical that this paper algorithms are obtained
The result of body marking area and subjective experiment is basically identical, shows that this method can be good at extracting the notable area of stereo-picture
Domain, and have very strong uniformity with the conclusion of subjective experiment.
4. subjective experiment
Hsv color model[32]It is a kind of color space created according to the intuitive nature of color, than RGB color more
Close to the experience of people and to colored perception.Therefore, this experiment uses the chromatic component of hsv color model extraction image.
In hsv color model, the about distinguishable 128 kinds of colourity of human eye, hue change range is from 0 ° to 360 °, and old friend's eye is distinguishable
Minimal color changing value is about 3 °.If carrying out the colourity change process with 3 for step-length respectively to stereo-picture or so viewpoint, that
, left and right view respectively obtains the image after the change of 120 width colourities, 14400 width bulky colors is obtained after the view combination of two of left and right
Spend image.If allowing subject to carry out subjective scoring one by one these stereo-pictures, stereo-picture is excessive, carries out subjective experiment and takes
When it is laborious.
Therefore approximatioss step by step from coarse to fine is used herein[15]Data processing is carried out, the central idea of this method is first
Two borders of the stereo-picture comfortably with uncomfortable colourity are found out with the classification of big step-length；Then with small step-length by two brightness borders
Middle chromatic value refinement, find out again small step-length comfortably with uncomfortable colourity border；Repeatedly refinement and experiment are carried out, finally
The minimal color resolution ratio of human eye is reached, comfortable and uncomfortable colourity chromatic value is found out.Using approximatioss step by step, step-length is forced
Nearly series is divided into 3 grades, and first order step-length to third level step-length is respectively 75,15 and 3[15].Using three-level step-length to stereo-picture
Colourity carries out linear transformation.If carrying out three-level step-length processing, subjective experiment test image respectively to stereo-picture or so view
Quantity is still a lot, and first order colourity step change is larger, it is meant that colourity changes it is obvious that human eye is bad during subjective experiment
Judge comfortable chromaticity range.Subjective assessment shows, when view (left or right) color in notable stereo-picture or so view
When angle value changes in the range of [0 °, 360 °], another view (right or left) can find corresponding chromatic value range, make their groups
The comfort level of the stereo-picture closed is good.Therefore, this experiment only carries out the 2nd grade of step-length segmentation to left view, to right view
Carry out the 1st grade, the 2nd grade, the segmentation of 3rd level step-length.
Test the acquisition using MATLAB2014b to notable stereo-picture and colourity at different levels change is handled, processing step
It is rapid as follows：
A. source stereo-picture is chosen, three-dimensional saliency map is obtained and finally gives notable stereo-picture.
B. notable stereo-picture is subjected to color notation conversion space, hsv color space is transformed into by rgb space, extract image
Colourity H.Notable stereo-picture in a determines the notable and non-significant part of image.Marking area is in the stereo-picture of source
Corresponding pixel, non-significant part is covered with gray scale for 255 image, as shown in Figure 5.
C. line translation is entered with step-length 15 to notable stereo-picture left view colourity, obtains 24 width left views, to significantly three-dimensional
Image right view colourity enters line translation with step-length 72, obtains 5 width right views；Can be obtained after combination of two 120 it is to be measured significantly
Stereo-picture.
D. subjective scoring is carried out to the notable stereo-picture of 120 width in step c, chosenThe image divided is comfortable to meet
Spend desired notable stereo-picture.Record respectivelyRight view minimax chromatic value n1rup、n1rdownAnd recordRight view minimax chromatic value n1rup'、n1rdown'。
E. because first order step-length is larger, the thin of second level step-length is carried out to meeting the view of boundary of comfort level requirement
Change.To chromatic value range n1 in step drup、n1rup' and n1rdown、n1rdown' carry out the second level step change, obtain 10 width
240 notable stereo-pictures to be measured can be obtained after two grades of right views, therefore left and right view combination of two.
F. subjective scoring is carried out to the notable stereo-picture of 240 width in step e, chosenThe image divided is comfortable to meet
Spend desired notable stereo-picture.Record respectivelyRight view minimax chromatic value n2rup、n2rdownAnd recordRight view minimax chromatic value n2rup'、n2rdown'。
G. similarly, to chromatic value range n2 in step frup、n2rup' and n2rdown、n2rdown' carry out the third level step-length become
Change, obtain 12 width three-level right views, therefore 288 notable stereo-pictures to be measured can be obtained after the view combination of two of left and right.
H. subjective scoring is carried out to the notable stereo-picture of 288 width in step g, chosenThe image divided is comfortable to meet
Spend desired notable stereo-picture.Record respectivelyRight view minimax chromatic value n3rup、n3rdown。
648 width testing images are obtained using three-level step-length approximatioss, it is to be measured much smaller than 14400 obtained in theory vertical
Body amount of images, greatly reduces the quantity of experiment, and can obtain good experimental result.
Experimental facilities and material
Experimental facilities three-dimensional display uses 22 inches of polarization type 3D three-dimensional displays of AOC (AOC) company.Eye tracker
Using the iView X RED of German SMI companies, eye tracker is used for verifying that paper algorithm obtains the correct of three-dimensional marking area
Property.
The stereoscopic image data storehouse that experimental material is provided from advanced science and technology institutes of South Korea is used as stereo-picture sample
Data acquisition system[20].The stereo-picture storehouse includes polytype image (mankind, trees, building etc.), and totally 120 pairs of resolution ratio are
1920x1080 stereo-picture composition.This experiment choose the stereo-picture storehouse in four width source stereo-picture pingpang.bmp,
Signboard.bmp, kid.bmp and flower.bmp, as shown in figure 1, the third dimension of this 4 width source stereo-picture is strong, with non-
Often good visual comfort.
Experimental situation and subject
To avoid the subjective evaluation result of stray light subject, experiment need to be carried out in darkroom[21].For 22 inches
Full HD display, viewing distance should be the three times distance of screen height[22]。
To ensure the correctness of experimental result, the personnel of all participation experiments are that Tianjin Eye Hospital carries out binocular vision
Functional test it is qualified by envoy.The normal subject of binocular vision physiology totally 28 people, the wherein people of male 13, women are chosen in experiment
15 people；There are 12 people of stereo technology research background, 16 people without stereo technology research background, and subject's dominant eye is all right
Eye.
Subjective assessment
Subjective assessment is the important evidence that the paper tests quantitative study.Before experiment, first experimental evaluation is introduced to subject
Method and grading system；Then carrying out Experiment Training to subject makes it be more clearly understood that code of points, it is ensured that experiment appraisal result
Universality.Wherein, training image is a few width images comprising various comfort levels, enables subject's correct understanding image institute
Belong to grade.To avoid subject from influenceing experimental evaluation result because of visual fatigue, so, each test phase is no more than 30 points
Clock[23]。
Experiment test is using single stimulating method, and in experiment, same width testing image is shown 2 times, if same subject is to same
When one stereo-picture provides the scoring for differing 2 grades or the above, this score value is considered as invalid scoring.It is vertical according to ITU-R BT.1438
Criterion is recommended in the subjective assessment of body image, the comfort level of stereo-picture is divided into 5 grades, as shown in table 1.Subject is allowed to provide
The score value of half point, the good stereo-picture of comfort level should reach 4 points (containing) more than[24], i.e., what comfortable stereo-picture effectively scored
Assembly averagePoint.
With reference to technical scheme process in detail：
First, the acquisition of notable stereo-picture
Stereo-picture has more information content than one-view image, and human eye can not possibly match all in a short time
Edge feature, most people only focuses on those " important areas ", then extracts the border of object in these regions, and finally match
These borders form stereoscopic vision.According to the stereoscopic vision attention characteristic of human eye, observer can be in marking area in image
Appearance is more paid close attention to[25], it is therefore possible to use the comfort level of significantly stereo-picture reflects the comfort level feelings of view picture stereo-picture
Condition, to improve test accuracy rate and reduce computation complexity.The parallax characteristic and spatial frequency of stereo-picture can all influence vision
Comfort level[26].Therefore, three-dimensional saliency map is finally obtained by stereo image parallax figure and two dimensional image notable figure.
Because the obtained gray value L of three-dimensional saliency map is between 0-255, i.e. L ∈ (0,255), the gray scale point of image
The information such as cloth and marking area edge all has ambiguity, it is impossible to be directly clearly divided into significantly according to this three-dimensional saliency map
Region and non-significant region, it is therefore desirable to determine a threshold value, enable to be divided gray scale saliency map.Paper is utilized
The theory of fuzzy mathematics describes the characteristic information of image[30], three-dimensional saliency map is optimized, it is notable after being optimized
Figure is bianry image, referred to as mask images[26], in order to remove the burr and cavitation at mask image edge, using morphology
Opening and closing operation mask image M (x, y) is optimized obtain optimize mask image M'(x, y).Mask images will finally be optimized
It is multiplied with original visual point image and obtains notable stereo-picture.
In order to verify whether gained significantly stereo-picture is true value, and the present invention is using eye tracker to the notable of source stereo-picture
Region carries out subjective measurement.It is consistent with subjective value to the marking area obtained by verification algorithm, test result indicates that the present invention is calculated
The three-dimensional marking area and the result of subjective experiment that method is obtained are basically identical, show that this method can be good at extracting stereo-picture
Marking area, and have very strong uniformity with the conclusion of subjective experiment.
2nd, subjective experiment
Therefore approximatioss step by step from coarse to fine is used herein[15]Data processing is carried out, the central idea of this method is first
Two borders of the stereo-picture comfortably with uncomfortable colourity are found out with the classification of big step-length；Then with small step-length by two brightness borders
Middle chromatic value refinement, find out again small step-length comfortably with uncomfortable colourity border；Repeatedly refinement and experiment are carried out, finally
The minimal color resolution ratio of human eye is reached, comfortable and uncomfortable colourity chromatic value is found out.Using approximatioss step by step, step-length is forced
Nearly series is divided into 3 grades, and first order step-length to third level step-length is respectively 75,15 and 3[15].Using three-level step-length to stereo-picture
Colourity carries out linear transformation.If carrying out three-level step-length processing, subjective experiment test image respectively to stereo-picture or so view
Quantity is still a lot, and first order colourity step change is larger, it is meant that colourity changes it is obvious that human eye is bad during subjective experiment
Judge comfortable chromaticity range.Subjective assessment shows, when view (left or right) color in notable stereo-picture or so view
When angle value changes in the range of [0 °, 360 °], another view (right or left) can find corresponding chromatic value range, make their groups
The comfort level of the stereo-picture closed is good.Therefore, this experiment only carries out the 2nd grade of step-length segmentation to left view, to right view
Carry out the 1st grade, the 2nd grade, the segmentation of 3rd level step-length.
648 width testing images are obtained using three-level step-length approximatioss, it is to be measured much smaller than 14400 obtained in theory vertical
Body amount of images, greatly reduces the quantity of experiment, and can obtain good experimental result.
3rd, analysis of experimental data
Experimental data is handled using least square sectional linear fitting[33], its basic thought is by surveyed number
Approached according to the straight line with segmentation to obtain corresponding mathematical modeling, and the fitting of straightway still uses least square method.Fortune
4 groups of experimental datas are handled with least square method, shown in such as figure (6) and figure (7), figure (6) abscissa is left view colourity
Value, ordinate is right view chromatic value；Figure (7) abscissa is left view chromatic value, and ordinate is the color of left view point and right viewpoint
Spend difference；Scheme in (7) on the basis of left view, on the occasion of representing that right viewpoint chromatic value is bigger than left view point chromatic value, negative value represents right
Viewpoint chromatic value is smaller than left view point chromatic value.
1) different scenes matches interval size to stereo-picture comfort level and had a certain impact.Scheme (6) in flower and
Kind formation sharp contrasts, the interval substantially colourity interval more comfortable than flower of the comfortable colourities of kind is small.For character image, observation
What person observed at first is the colour of skin problem of people, there is certain limitation for color-match.There are several discrimination face in the eyes of people
The taper photosensory cell of color, most sensitive (wavelength is respectively the light to yellow green, green and bluish violet (or pansy) respectively
564th, 534 and 420 nanometers), if distinguishing cell of the stimulation slightly larger than discrimination green that the cell of yellow green is subject to, the sensation of people
It is yellow；If distinguishing that the stimulation that the cell of yellow green is subject to is much higher than the cell for distinguishing green, feeling for people is red.Area
Between (50o, 220o), corresponding color is changed into yellow to blueness from red, and the color of this interval left and right colourity synthesis more meets people couple
In the subjective consciousness of the colour of skin, so this interval range is relatively large；And the color that other intervals can be matched for left view colourity
Degree is interval relatively small.For this kind of images of flower, because the hobby of people is different, assorted Huadu is popular, so its
Left view is larger in the matched right view chromaticity range of any chroma areas.
2) figure (6) is the comfortable chromaticity match figure of four groups of stereo-pictures, the region representation that segmented linear is surrounded or so view
The stereo-picture that chromatic value is formed under this interval combinations feels it is comfortable to beholder's.It is four groups of stereo-pictures to scheme (7)
Comfortable colourity disparity map, the region representation that segmented linear is surrounded or so view chromatic value difference is in which kind of scope, stereo-picture
Visual comfort is good.The positive maximum different value scope of the comfortable colourity of the different scenes of table 1 and negative maximum difference are drawn according to figure (7)
It is worth scope.The colourity that this data is demonstrated corresponding to different scenes is interval different.According to figure (6), figure (7) and table 1, it can be seen that
For different scenes, its comfortable chromaticity range is different.Therefore, in order that scheme herein has universality, to figure (6)
It is averaged, is respectively obtained vertical with the comfortable chromaticity match figure of four groups of stereo-pictures and the experimental data of disparity map shown in figure (7)
Shown in the comfortable chromaticity match figure of body image and the comfortable colourity disparity map of stereo-picture, such as figure (8) and figure (9).Will figure (8) and figure
(9) cathetus fitting enclosing region is divided into R1、R2、R3、R4、R5Deng five parts：
a.R1、R2、R3、R4、R5Scored for stereo-picture five partsColourity zone of comfort.From figure (8) and figure (9)
In as can be seen that the comfortable chromaticity match figure of stereo-picture is symmetrical on positive diagonal axis, the comfortable colourity difference of stereo-picture
Figure is axisymmetric on level.Therefore, influence of the major-minor eye to comfortable chroma areas and little.As can be seen from the figure for
Different left view chromatic values is interval, and the comfortable chromaticity range size of right view is also different, it can be deduced that human eye is for different
The susceptibility of color of light is different.In R1、R2、R3、R4、R5Chromatic value outside region, is the uncomfortable chromaticity range of human eye, therefore can
Know, the colourity of right and left eyes only in certain scope, feeling of giving people be only it is comfortable, if right and left eyes chroma difference is excessive,
Human eye will be unable to merge left and right visual point image, causes retinal rivalry phenomenon, causes uncomfortable.
B. in R1And R5Region, left view hue change range is about (320.0 °, 9.8 °), and corresponding color is by purple transition
To pinkish red to red.Because the photon energy of different colours is different, the stimulated degree of human eye is also different.In hsv color model,
When tone is from 300.0 ° to 360.0 °, purple component decay, chromatogram becomes red.Red wavelength energy is smaller, to the thorn of human eye
Swash also smaller；It is wider in this zone comfort chroma areas for the picture of flower type because selected picture scene is different, still
Interval comfortable colourity diminishes personage's picture herein, and scheme (8) and scheme the statistical average that (9) are four groups of stereoscopic image datas, institute
With as can be seen from the figure R1And R5The chromaticity range in region can compare R2It is small.
C. in R2Region, left view hue change range is about (9.8 °, 200 °), and corresponding color is changed into yellow extremely from red
Blueness, as can be seen from the figure good wider range of the comfortable colourity of stereo-picture.One is because of chromatogram photon interval herein
Energy is of moderate size, and the stimulation to human eye is gentleer, than the viewing of convenient human eye；Two be because this interval is for polytype
Picture color is convenient；In summary, when the tone value of left and right view is interval herein, stereo image vision comfort level scope compared with
It is wide.
D. in R3Region, left view tone variations interval is about (200 °, 220 °), and corresponding color is blue, and blueness etc. is short
Wave energy is larger, and the stimulation to human eye is excessively strong, causes vision uncomfortable, but is due to that this experiment is subjective experiment, human brain
There is certain preference for blue light and green light.So, comfortable chromaticity range is compared with R2Region narrows but compares R4Comfortable colourity model
Enclose width.
E. in R4Region, left view tone variations interval is about (220 °, 320 °), and corresponding color has blueness to be changed into purple,
Because blueness and purple are shortwave, the photon energy of purple light is larger, and the stimulation to human eye is also larger, causes human eye not relax
It is suitable；Moreover, for the picture of a lot of scenes, purple is not appropriate for as its picture mass-tone, under priori conditions, gives people a kind of not relax
Suitable sensation, therefore from figure (8) and figure (9) it can be seen that R4Interval chromaticity range is most narrow in whole zone of comfort.
4th, experimental verification
1. contrast verification
Document [14] shadow of quantitative study colourity factor to visual comfort in the case where not accounting for vision significance
Ring, such as figure (10) as can be seen that broken line surround chromaticity range it is basically identical, illustrate marking area comfort level can represent it is whole
The comfort level of width stereo-picture, it is consistent with theory.Document [14] is given original image, is allowed using double experimental methods stimulated
Whether subject is comfortably evaluated test image, can there is certain limitation to the subjectivity of subject.Due to being to have reference
Subjective experiment, so for some scenes, its left view colourity interval (0 °, 360 °), right view should not have comfortably
Colourity interval is matching.Herein using single experimental method stimulated, double stimulating methods compared with document [14] more meet human eye and regarded
The subjectivity of feel, its left view colourity interval (0 °, 360 °) right view has certain comfortable colourity interval, and scene is different
The comfortable interval size of colourity is also different.The conclusion drawn herein has authenticity, can be more provided with for the making of stereo content
The technical support of profit, rational scheme is provided for the making of stereo content comfort standard.
2. test checking
The obtained comfortable chromaticity match figure of stereo-picture and the comfortable colourity disparity map of stereo-picture are tested herein in order to verify
Whether there is universality, have chosen two width source stereo-pictures in addition again herein and carry out result verification.Selected image such as figure (11) institute
Show, this two width stereo-picture is subjected to colourity processing according to hereinbefore step, due to consideration that picture number problem, therefore it is real
Checking only carries out third level step change to right figure piece, and every width stereo-picture obtains 120 width colourity modified-images, chooses it
In 80 width images carry out subjective experiment.Then subject is allowed to carry out subjective scoring according to hereinbefore subjective experiment method.It is right
Experimental result is counted, wherein, statistical average point in this group of picture of flower oneThe qualified stereo-picture divided has 45
The stereo-picture that width, wherein left and right view colourity average value fall in the comfortable chromaticity match figure of stereo-picture has 43 width.Therefore, exist
Proportion is 95.5% in qualified stereo-picture；Similarly, accuracy is 96.9% in this group of picture of boy.Two groups of pictures are just
True rate has been above 95%, so as to demonstrate the comfortable chromaticity match figure of this paper experimental result stereo-pictures and the comfortable color of stereo-picture
Degree disparity map has universality.So, this paper conclusions can easily judge that the euphorosia of a width stereo-picture is asked
Topic, improves time-consuming, the laborious problem of subjective experiment to a certain extent.
Bibliography
[1]Pking W M.Viewing comfort with stereoscopic pictures:An
experimental study on the subjective effects of disparity magnitude and depth
Of focus.Journal of the Society for Information Display, 1995,3 (3):101
[2]Yano S,Emoto M,Mitsuhashi T.Two factors in visual fatigue caused
by stereoscopic HDTV images.Displays,2004,25(4):141
[3]Nojiri Y,Yamanoue H,Hanazato A,et al.Visual comfort/discomfort and
visual fatigue caused by stereoscopic HDTV viewing//Proceedings of SPIE.USA,
2004:303
[4] Speranza F, Tam W J, Renaud R, et al.Effect of disparity and motion
on visual comfort of stereoscopic images//Proceedings of SPIE.USA,2006:1
[5]Lambooij M,Ijsselsteijn W,Fortuin M,et al.Visual discomfort and
visual fatigue of stereoscopic displays:a review.Journal of Imaging Science
and Technology,2009,53(3):1
[6]Meesters L,Ijsselsteijn W,Seuntiens P.A survey of perceptual
evaluations and requirements of three-dimensional TV.IEEE Transactions on
Circuits and Systems for Video Technology,2004,14(3):381
[7]Zheng H D,Yu Y J,Cheng W M.Areview on three-dimensional display
techniques.Optical Technique,2008,34(3):426
(the bright three-dimensional stereo display techniques New research progress optical technologies of Zheng Huadong, Yu Yingjie, Cheng Wei, 2008,34 (3):
426)
[8]FRANK L K.Visual comfort of binocular and 3D displays.Display,
2004,25(2):99
[9]PAN H,YUAN CH,DALY S.3D video disparity scaling for preference and
prevention of discomfort//Proceedings of SPIE,Shanghai,2011:4
[10]Yang L.Theoretical analysis and experimental research of multi-
view stationary free stereo image processing[Dissertation].Tianjin:Tianjin
University,2007
(theory analysis of Yang Lei multiple views still stereoscopic pictures processing and experimental study [academic dissertation] Tianjin：
University Of Tianjin, 2007)
[11]LI X F,WANG Q H.Auto stereoscopic display for visual
fatigue.Liquid Crystals and Displays,2008,23(4):464
(Li little Fang, Wang Qionghua auto-stereoscopic displays watch visual fatigue liquid crystal and display, 2008,23 (4):464)
[12]ISO IWA 3 2005(E).Image safety:Reducing the incidence of
undesirable biomedical effects caused by visual image sequences.Switzerland:
ISO Copyright Office IHS, 2005
[13]ZANG Y J,LI S M,WEI J J.Measurement of the range of the
brightness parameter influencing the comfort of stereoscopic
images.Electronic Measurement Technology,2012,35(2):59
(brightness range of the beautiful influences stereo-picture comfort level of Zang Yanjun, Li Sumei, Wei Jinjin, Liu Wen determines electronic, horologicals
Amount technology, 2012,02:59)
[14]LI S M,FU B B,ZANG Y J,et al.Measurement of the range of
saturation parameter influencing the comfort of stereoscopic images.Journal
of Tianjin University,2012,45(8):669
(the measure University Of Tianjin of the saturation degree scope of the such as Li Sumei, Fu Beibei, Zang Yanjun influence stereo-picture comfort levels
Journal, 2012,45 (8):669)
[15]LIU C,LI S M,ZHU D,et al.Quantitative research of the hue
parameter influence on the visual comfort of stereoscopic images.Journal of
Optoelectronics Laser,2014,01:178
(quantitative study photoelectricity of Liu Chang, Li Sumei, Zhu Dan, Zang Yan army the colourities to three-dimensional image vision comfort degree
Sub- laser, 2014,01:178)
[16]LIU CH,LI S M.Measurement of contrast range affecting comfort of
stereoscopic images.Journal of Optoelectronics Laser,2014,25(4):748
(the measure photoelectron laser of the contrast range of Liu Chang, Li Su plum influence stereo-picture comfort levels, 2014,25
(4):748)
[17]ZHU D.Research of the Parallax Influencing the Visual Comfort of
Stereoscopic Images[Dissertation].Tianjin:Tianjin University,2014
(parallax research [academic dissertation] the Tianjin of Zhu Dan influence stereo-picture comfort levels：University Of Tianjin, 2014)
[18]QIN M P,YIU M CH,XIN G Y,et al.A Hybrid of Local and Global
Saliencies for Detecting Image Salient Region and Appearance.IEEE
Transactions on Systems Man&Cybernetics Systems.2017,47(1):86.
[19]CHENG M M,ZHANG G X,Mitra N J,et al.Global contrast based salient
region detection//Proceedings of Computer Vision and Pattern Recognition,USA,
2015,37(3):569
[20]IVY Lab stereoscopic image database[OL].[2013-03-12].http://
ivylab.kaist.ac.kr/demo/3DVCA/3DVCA.html
[21]Seuntiens P.Visual experience of 3D TV.Eindhoven University of
Technology,2006
[22]Rec.ITU-R BT.1438.Subjective assessment of stereoscopic
television pictures,2000
[23]The State Administration of Radio Film and Television.GB/T7401-
1987,Method of subjective assessment of quality of color TV pictures.Beijing:
China Standards Press,1987
(State Administration of Radio, Film and Television .GB/T7401-1987 colour TV picture quality subjective evaluation method Beijing:China
Standard publishing house, 1987)
[24]Beijing Technology Quality Control Office.DBII/T384.5-2009,Image
information management system technical:image quality requirements and
evaluation methods.Beijing:China Standards Press,2009
(Administration of Quality and Technology Supervision of Beijing .DBII/T384.5-2009 image information management systems technical specification the 5th
Point：Image quality requirements and evaluation method Beijing:China Standards Press, 2009)
[25]WANG F.Visual saliency detection based on context and background
[Dissertation].Dalian University of Technology,2013
(Wang Fei are based on context and the vision significance of background detection [academic dissertation] Dalian University of Technology, 2013)
[26]JIANG Q P,SHAO F,YU M,et al.An Objective Stereoscopic Image
Visual Comfort Assessment Metric Based on Visual Important Regions.Chinese
Journal of electronics and information technology,2014,36(04):875
(Jiang Qiuping, Shao Feng, Jiang Gangyi, Yu Mei, Peng Zong lift the stereo image vision comfort level of view-based access control model important areas
Method for objectively evaluating electronics and information journal, 2014,36 (04):875)
[27]Harel,Jonathan,Christof K,et al.Graph-based visual
saliency.Advances in neural information processing systems,2007,19:545
[28]Rhemann C,Hosni A,Bleyer M,et al.Fast cost-volume filtering for
visual correspondence and beyond//Proceedings of Computer Vision and Pattern
Recognition,USA,2011,35:3017
[29]Wang J,Da Silva M P,Le Callet P,et al.Computational model of
stereoscopic 3D visual saliency.IEEE Transactions on Image Processing,2013,22
(6):2151
[30]LI ZH M.Research on image segmentation algorithm based on Fuzzy
Clustering [Dissertation].Hunan University,2009
(Research on Algorithms for Image Segmentation [academic dissertation] Hunan University of the Li Zhi plums based on fuzzy clustering, 2009)
[31]YU W J.Study of Otsu image segmentation based on Improved Genetic
Algorithm[Dissertation].Hainan University,2012
(maximum between-cluster variance Research on Method of Image Segmentation [academic dissertation] the Hainan of the handsome of Yu Wen based on improved adaptive GA-IAGA
University, 2012)
[32]TONG L F,TU Y.Daniel den Engelsen.digital TV display
technology.Beijin:Electronics Industry Press,2007,112
(Tong Lin is long-standing, Tu Yan .Daniel den Engelsen. Digital Display TV Beijing：Electronics industry is published
Society, 2007,112)
[33]XIE Y B.Least square piecewise linear fitting.modern electric
power,1992,01:19
(thank friendly treasured Sectional Line Fitting by means of Least Sequence Method modern electrics, 1992,01:19).
Claims (6)
1. a kind of measuring method of the stereo-picture comfort level chromaticity range based on salient region, it is characterized in that, first use and regard
The method that difference figure and two dimensional image notable figure are combined obtains three-dimensional saliency map, recycles fuzzy membership and mask excellent to its
Change draws final notable stereo-picture, and extracts the correct of marking area using the three-dimensional significantly algorithm of eye tracker experimental verification
Property；Then subjective experiment is carried out using approximatioss step by step from coarse to fine, experimental data is fitted by least square method,
The comfortable chromaticity match figure and comfortable colourity disparity map of notable stereo-picture are obtained, and to being tested the subjectivity of image under different scenes
Experience and analyzed.
2. the measuring method of the stereo-picture comfort level chromaticity range as claimed in claim 1 based on salient region, it is special
Levying is, obtains comprising the concrete steps that for three-dimensional saliency map, brightness, colourity, contrast of the two dimensional image notable figure by image in itself
Degree and spatial frequency are calculated, and specific calculated using GBVS algorithms obtains right viewpoint plane notable figure, is designated as SMR(x,y)；Adopt
The disparity map on the basis of right viewpoint is obtained with Fast Stereo Matching Algorithm, d is designated asR(x,y)；Plane is regarded using linear mode
Feel notable figure SMR(x, y) and right anaglyph dR(x, y), which is weighted, obtains three-dimensional saliency map I (x, y), such as formula (1)：
I (x, y)=w1dR(x,y)+w2SMR(x,y) (1)
In formula：I (x, y) is three-dimensional saliency map；w1And w2To weight proportion, and w1+w2=1.
3. the measuring method of the stereo-picture comfort level chromaticity range as claimed in claim 1 based on salient region, it is special
Levying is, comprising the concrete steps that for the acquisition of notable stereo-picture, and the characteristic information of image is described using the theory of fuzzy mathematics, will
Three-dimensional saliency map is optimized, and the notable figure after being optimized is bianry image, referred to as mask images, the pixel value of pixel
Show that the point in former stereo-picture belongs to marking area for 1, otherwise belong to non-significant region.
4. the measuring method of the stereo-picture comfort level chromaticity range as claimed in claim 3 based on salient region, it is special
Levying is, the further detailed process of the acquisition of notable stereo-picture is as follows：
If domain X is three-dimensional saliency map, the element in domain is divided into two classes, marking area A and non-significant region B, A and B
It is an X division, i.e.,：
A ∪ B=X
A ∩ B=φ
The gray value of three-dimensional saliency map represents that the pixel of this in original image belongs to the degree of marking area, therefore directly gives
Membership function, such as formula (2), A (x, y) represent the degrees of membership of the gray value L to marking area A at pixel (x, y) place.
<mrow>
<mi>A</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>,</mo>
<mi>y</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mi>L</mi>
<mn>255</mn>
</mfrac>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>2</mn>
<mo>)</mo>
</mrow>
</mrow>
Using the method deblurring of Threshold segmentation, segmentation threshold T is determined by maximum between-cluster variance method, passes through the threshold value deblurring
Mask images M (x, y) [28] is obtained, such as formula (3)：
For any pixel point (x, y) in three-dimensional saliency map I (x, y), if A (x, y) is big to the degree of membership of marking area
In threshold value T, then the white portion that the pixel belongs in vision significance region, correspondence mask images M (x, y) otherwise belongs to
Black region, in order to remove the burr and cavitation at mask image edge, using morphologic opening and closing operation to mask image M
(x, y), which is optimized, to be obtained optimizing mask image M'(x, y)；Finally optimization mask images are multiplied with original visual point image and obtained
Notable stereo-picture.
5. the measuring method of the stereo-picture comfort level chromaticity range as claimed in claim 1 based on salient region, it is special
Levying is, subjective measurement is carried out to the marking area of source stereo-picture using eye tracker, first carried out before experiment every time calibration experiments with
Reference measurement, tests when subject is less than 1 degree to the observation error for calibrating datum mark, presents at random per pictures and be in
It is 5s between current.The thermal map for the interesting image regions measured according to eye tracker, with notable stereo-picture regional correlation.
6. the measuring method of the stereo-picture comfort level chromaticity range as claimed in claim 1 based on salient region, it is special
Levying is, carries out data processing using approximatioss step by step from coarse to fine, step-length is approached into series and is divided into 3 grades, first order step-length is arrived
Third level step-length is respectively 75,15 and 3, and linear transformation is carried out to three-dimensional image chroma using three-level step-length, wherein, only to left view
Figure carries out the 2nd grade of step-length segmentation, and the 1st grade, the 2nd grade, the segmentation of 3rd level step-length are carried out to right view；Experiment uses MATLAB2014b
Acquisition and colourity at different levels change to notable stereo-picture are handled, and process step is as follows：
A. source stereo-picture is chosen, three-dimensional saliency map is obtained and finally gives notable stereo-picture；
B. notable stereo-picture is subjected to color notation conversion space, hsv color space is transformed into by rgb space, extract image chroma
H.Notable stereo-picture in a determines the notable and non-significant part of image.Marking area is correspondence in the stereo-picture of source
Pixel, non-significant part is covered with gray scale for 255 image；
C. line translation is entered with step-length 15 to notable stereo-picture left view colourity, 24 width left views is obtained, to notable stereo-picture
Right view colourity enters line translation with step-length 72, obtains 5 width right views；120 notable solids to be measured can be obtained after combination of two
Image；
D. subjective scoring is carried out to the notable stereo-picture of 120 width in step c, chosenThe image divided will to meet comfort level
The notable stereo-picture asked, is recorded respectivelyRight view minimax chromatic value n1rup、n1rdownAnd record
Right view minimax chromatic value n1rup'、n1rdown'；
E. to chromatic value range n1 in step drup、n1rup' and n1rdown、n1rdown' carry out the second level step change, obtain 10
240 notable stereo-pictures to be measured can be obtained after two grades of right views, therefore left and right view combination of two；
F. subjective scoring is carried out to the notable stereo-picture of 240 width in step e, chosenThe image divided will to meet comfort level
The notable stereo-picture asked, is recorded respectivelyRight view minimax chromatic value n2rup、n2rdownAnd record
Right view minimax chromatic value n2rup'、n2rdown'；
G. similarly, to chromatic value range n2 in step frup、n2rup' and n2rdown、n2rdown' carry out the third level step change, obtain
288 notable stereo-pictures to be measured can be obtained to 12 width three-level right views, therefore after the view combination of two of left and right；
H. subjective scoring is carried out to the notable stereo-picture of 288 width in step g, chosenThe image divided will to meet comfort level
The notable stereo-picture asked, is recorded respectivelyRight view minimax chromatic value n3rup、n3rdown。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710442862.3A CN107194927A (en) | 2017-06-13 | 2017-06-13 | The measuring method of stereo-picture comfort level chromaticity range based on salient region |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710442862.3A CN107194927A (en) | 2017-06-13 | 2017-06-13 | The measuring method of stereo-picture comfort level chromaticity range based on salient region |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107194927A true CN107194927A (en) | 2017-09-22 |
Family
ID=59877529
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710442862.3A Pending CN107194927A (en) | 2017-06-13 | 2017-06-13 | The measuring method of stereo-picture comfort level chromaticity range based on salient region |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107194927A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108510453A (en) * | 2018-03-07 | 2018-09-07 | 西安工程大学 | The intelligent traffic monitoring image deblurring method of view-based access control model attention mechanism |
| CN108520506A (en) * | 2018-03-26 | 2018-09-11 | 天津大学 | The measurement method of stereo-picture saturation degree comfort standard based on marking area |
| CN109167988A (en) * | 2018-08-29 | 2019-01-08 | 长春理工大学 | A kind of stereo image vision comfort level evaluation method based on D+W model and contrast |
| CN110188766A (en) * | 2019-04-17 | 2019-08-30 | 平安科技（深圳）有限公司 | Image major heading detection method and device based on convolutional neural networks |
| WO2022111140A1 (en) * | 2020-11-25 | 2022-06-02 | International Business Machines Corporation | Video encoding through non-saliency compression for live streaming of high definition videos in low-bandwidth transmission |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120121138A1 (en) * | 2010-11-17 | 2012-05-17 | Fedorovskaya Elena A | Method of identifying motion sickness |
| CN103096125A (en) * | 2013-02-22 | 2013-05-08 | 吉林大学 | Stereoscopic video visual comfort evaluation method based on region segmentation |
| CN104243956A (en) * | 2014-09-12 | 2014-12-24 | 宁波大学 | Three-dimensional image visual saliency map extraction method |
-
2017
- 2017-06-13 CN CN201710442862.3A patent/CN107194927A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120121138A1 (en) * | 2010-11-17 | 2012-05-17 | Fedorovskaya Elena A | Method of identifying motion sickness |
| CN103096125A (en) * | 2013-02-22 | 2013-05-08 | 吉林大学 | Stereoscopic video visual comfort evaluation method based on region segmentation |
| CN104243956A (en) * | 2014-09-12 | 2014-12-24 | 宁波大学 | Three-dimensional image visual saliency map extraction method |
Non-Patent Citations (2)
| Title |
|---|
| 刘畅 等: "色度对立体图像视觉舒适度影响的定量研究", 《光电子.激光》 * |
| 徐姝宁 等: "基于视觉显著性的立体图像质量评价方法", 《信息技术》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108510453A (en) * | 2018-03-07 | 2018-09-07 | 西安工程大学 | The intelligent traffic monitoring image deblurring method of view-based access control model attention mechanism |
| CN108510453B (en) * | 2018-03-07 | 2021-06-29 | 西安工程大学 | Intelligent traffic monitoring image deblurring method based on visual attention mechanism |
| CN108520506A (en) * | 2018-03-26 | 2018-09-11 | 天津大学 | The measurement method of stereo-picture saturation degree comfort standard based on marking area |
| CN109167988A (en) * | 2018-08-29 | 2019-01-08 | 长春理工大学 | A kind of stereo image vision comfort level evaluation method based on D+W model and contrast |
| CN109167988B (en) * | 2018-08-29 | 2020-06-26 | 长春理工大学 | Stereo image visual comfort evaluation method based on D + W model and contrast |
| CN110188766A (en) * | 2019-04-17 | 2019-08-30 | 平安科技（深圳）有限公司 | Image major heading detection method and device based on convolutional neural networks |
| CN110188766B (en) * | 2019-04-17 | 2023-06-30 | 平安科技（深圳）有限公司 | Image main target detection method and device based on convolutional neural network |
| WO2022111140A1 (en) * | 2020-11-25 | 2022-06-02 | International Business Machines Corporation | Video encoding through non-saliency compression for live streaming of high definition videos in low-bandwidth transmission |
| US11758182B2 (en) | 2020-11-25 | 2023-09-12 | International Business Machines Corporation | Video encoding through non-saliency compression for live streaming of high definition videos in low-bandwidth transmission |
| GB2616998A (en) * | 2020-11-25 | 2023-09-27 | Ibm | Video encoding through non-saliency compression for live streaming of high definition videos in low-bandwidth transmission |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107194927A (en) | The measuring method of stereo-picture comfort level chromaticity range based on salient region | |
| CN103402117B (en) | Based on the video image color cast detection method of Lab chrominance space | |
| CN102663745B (en) | Color fusion image quality evaluation method based on vision task. | |
| CN108038456A (en) | A kind of anti-fraud method in face identification system | |
| CN102497565B (en) | Method for measuring brightness range influencing comfort degree of stereoscopic image | |
| CN102750731B (en) | Based on the remarkable computing method of stereoscopic vision of the simple eye receptive field in left and right and binocular fusion | |
| CN103096122B (en) | Stereoscopic vision comfort level evaluation method based on motion features inside area of interest | |
| Lambooij et al. | The impact of video characteristics and subtitles on visual comfort of 3D TV | |
| CN104853184B (en) | Detection method for parallax scope influencing comfort level of stereo image | |
| Marko et al. | Sterecon—three-dimensional reconstructions from stereoscopic contouring | |
| CN104346809A (en) | Image quality evaluation method for image quality dataset adopting high dynamic range | |
| CN106993183A (en) | Comfortable brightness quantitative approach based on stereo-picture marking area | |
| CN109741285A (en) | A kind of construction method and system of underwater picture data set | |
| Zhou et al. | Visual comfort assessment for stereoscopic image retargeting | |
| Zhang et al. | Perceived transparency in optical see-through augmented reality | |
| CN107680076A (en) | Stereo-picture comfort level contrast range measuring method based on salient region | |
| Kim et al. | Quality assessment of perceptual crosstalk on two-view auto-stereoscopic displays | |
| Shen et al. | Color enhancement algorithm based on Daltonization and image fusion for improving the color visibility to color vision deficiencies and normal trichromats | |
| CN105872516A (en) | Method and device for obtaining parallax parameters of three-dimensional film source | |
| Vangorp et al. | Depth from HDR: depth induction or increased realism? | |
| Chen et al. | A quantitative measurement of binocular color fusion limit for different disparities | |
| Kim et al. | Perceptual crosstalk prediction on autostereoscopic 3D display | |
| Watanabe et al. | Depth capture and transparency of regions bounded by illusory and chromatic contours | |
| Chen et al. | Visual discomfort caused by color asymmetry in 3D displays | |
| Gomes et al. | Near perfect visual compensation for atmospheric color distortions |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170922 |
|
| RJ01 | Rejection of invention patent application after publication |