CN102333233B

CN102333233B - Stereo image quality objective evaluation method based on visual perception

Info

Publication number: CN102333233B
Application number: CN 201110284944
Authority: CN
Inventors: 邵枫; 蒋刚毅; 郁梅; 李福翠; 彭宗举
Original assignee: Ningbo University
Current assignee: NANTONG OUKE NC EQUIPMENT Co.,Ltd.
Priority date: 2011-09-23
Filing date: 2011-09-23
Publication date: 2013-11-06
Anticipated expiration: 2031-09-23
Also published as: CN102333233A

Abstract

The invention discloses a stereo image quality objective evaluation method based on visual perception. Firstly, a stereo image is divided into a strong edge block, a weak edge block , a flat block and a texture block, and characteristic information reflecting image quality and depth perception of different zone blocks is extracted through computation so as to obtain stereo image characteristic vectors; and then the characteristic vectors of distorted stereo images of the same distortion type in a distorted stereo image set are trained through support vector regression, and each distorted stereo image of the same distortion type is tested by a support vector regression training model to obtain the objective image quality evaluation forecast value of each distorted stereo image. The method has the advantages that the obtained characteristic vector information reflecting the image quality and the depth perception has stronger stability and can better reflect the quality change condition of the stereo images, and the relevance of an objective evaluation result and subjective perception is improved.

Description

A kind of objective evaluation method for quality of stereo images based on visually-perceptible

Technical field

The present invention relates to a kind of image quality evaluating method, especially relate to a kind of objective evaluation method for quality of stereo images based on visually-perceptible.

Background technology

Along with developing rapidly of image coding technique and stereo display technique, the stereo-picture technology has been subject to paying close attention to more and more widely and using, and has become a current study hotspot.The stereo-picture technology is utilized the binocular parallax principle of human eye, and binocular receives the left and right visual point image from Same Scene independently of one another, merges by brain and forms binocular parallax, thereby enjoy the stereo-picture with depth perception and realism.Impact due to acquisition system, store compressed and transmission equipment, stereo-picture can inevitably be introduced a series of distortion, and compare with the single channel image, stereo-picture need to guarantee the picture quality of two passages simultaneously, it is carried out quality evaluation have very important significance.Yet the effective method for objectively evaluating of stereoscopic image quality shortage is estimated at present.Therefore, setting up effective stereo image quality objective evaluation model tool is of great significance.

Objective evaluation method for quality of stereo images mainly can be divided into two classes: 1) based on the left and right channel image quality evaluation of picture quality, it directly applies to the evaluation stereo image quality with the plane picture quality evaluating method, yet the left and right visual point image of stereoscopic image merges the relief process of generation also to be difficult to represent with simple mathematical method, and also exist between the visual point image of left and right to influence each other, the left and right visual point image is carried out the simple linear weighting be difficult to effectively estimate stereo image quality; 2) based on the left and right channel image quality evaluation of three-dimensional perception, it reflects by parallax information or depth information, yet the limitation due to present disparity estimation and estimation of Depth technology, how effectively depth image or anaglyph quality are estimated to characterize truly third dimension and know characteristic, remain one of difficulties in the stereo image quality objective evaluation.Therefore, how picture quality and depth perception information being attached in evaluation method simultaneously, making evaluation result more feel to meet the human visual system, is all needing in the evaluating objective quality process to carry out in stereoscopic image the problem researched and solved.

Summary of the invention

Technical problem to be solved by this invention is to provide a kind of objective evaluation method for quality of stereo images that can effectively improve the correlation of objective evaluation result and subjective perception.

The present invention solves the problems of the technologies described above the technical scheme that adopts: a kind of objective evaluation method for quality of stereo images based on visually-perceptible is characterized in that comprising the following steps:

1. make S _orgUndistorted stereo-picture for original makes S _disFor the stereo-picture of distortion to be evaluated, with S _orgLeft visual point image be designated as { L _org(x, y) }, with S _orgRight visual point image be designated as { R _org(x, y) }, with S _disLeft visual point image be designated as { L _dis(x, y) }, with S _disRight visual point image be designated as { R _dis(x, y) }, wherein, the coordinate position of pixel in (x, y) left visual point image of expression and right visual point image, 1≤x≤W, 1≤y≤H, W represent the width of left visual point image and right visual point image, H represents the height of left visual point image and right visual point image, L _org(x, y) represents S _orgLeft visual point image { L _org(x, y) } in coordinate position be the pixel value of the pixel of (x, y), R _org(x, y) represents S _orgRight visual point image { R _org(x, y) } in coordinate position be the pixel value of the pixel of (x, y), L _dis(x, y) represents S _disLeft visual point image { L _dis(x, y) } in coordinate position be the pixel value of the pixel of (x, y), R _dis(x, y) represents S _disRight visual point image { R _dis(x, y) } in coordinate position be the pixel value of the pixel of (x, y);

2. utilize human vision to the visual masking effect of background illumination and texture, extract respectively undistorted left visual point image { L _org(x, y) } and undistorted right visual point image { R _org(x, y) } minimum discernable distorted image, with undistorted left visual point image { L _org(x, y) } minimum discernable distorted image be designated as { J _L(x, y) }, with undistorted right visual point image { R _org(x, y) } minimum discernable distorted image be designated as { J _R(x, y) }, wherein, J _L(x, y) represents { J _L(x, y) } in coordinate position be the pixel value of the pixel of (x, y), J _R(x, y) represents { J _R(x, y) } in coordinate position be the pixel value of the pixel of (x, y);

3. obtain respectively undistorted left visual point image { L by regional detection algorithm _org(x, y) } and the left visual point image { L of distortion _dis(x, y) }, undistorted right visual point image { R _org(x, y) } and the right visual point image { R of distortion _dis(x, y) } in the block type of each 8 * 8 sub-block, be designated as p, wherein, p ∈ 1,2,3,4}, p=1 represents Strong edge blocks, p=2 represents weak edge block, p=3 represents smooth block, p=4 represents texture block;

4. according to undistorted left visual point image { L _org(x, y) } minimum discernable distorted image { J _L(x, y) } and undistorted right visual point image { R _org(x, y) } minimum discernable distorted image { J _R(x, y) }, by the left visual point image { L of calculated distortion _dis(x, y) } in various block types 8 * 8 sub-blocks the spatial noise intensity that is used for reflection picture quality and be used for the space structure intensity of reflection picture quality, and the right visual point image { R of distortion _dis(x, y) } in various block types 8 * 8 sub-blocks the spatial noise intensity that is used for reflection picture quality and be used for the space structure intensity of reflection picture quality, obtain respectively the left visual point image { L of distortion _dis(x, y) } the characteristic vector that is used for reflection picture quality and the right visual point image { R of distortion _dis(x, y) } the characteristic vector that is used for reflection picture quality, then to the left visual point image { L of distortion _dis(x, y) } and the right visual point image { R of distortion _dis(x, y) } the characteristic vector that is used for reflection picture quality carry out linear weighted function, obtain S _disThe characteristic vector that is used for reflection picture quality, be designated as F _q

5. according to undistorted left visual point image { L _org(x, y) } minimum discernable distorted image { J _L(x, y) } and undistorted right visual point image { R _org(x, y) } minimum discernable distorted image { J _R(x, y) }, by the left visual point image { L of calculated distortion _dis(x, y) } and the right visual point image { R of distortion _dis(x, y) } absolute difference image in various block types 8 * 8 sub-blocks the spatial noise intensity that is used for the reflection depth perception and be used for the space structure intensity of reflection depth perception, obtain S _disThe characteristic vector that is used for the reflection depth perception, be designated as F _s

6. with S _disThe characteristic vector F that is used for reflection picture quality _qWith the characteristic vector F that is used for the reflection depth perception _sForm the New Characteristics vector, as S _disCharacteristic vector, be designated as X, X=[F _q, F _s], " [] " is the vector representation symbol, [F _q, F _s] represent characteristic vector F _qWith characteristic vector F _sCouple together and form a New Characteristics vector;

7. adopt n undistorted stereo-picture, set up its distortion stereo-picture set under the different distortion levels of different type of distortion, this distortion stereo-picture set comprises the stereo-picture of several distortions, utilizes the subjective quality assessment method to obtain respectively the average subjective scoring difference of the stereo-picture of every width distortion in the set of distortion stereo-picture, is designated as DMOS, DMOS=100-MOS, wherein, MOS represents the subjective scoring average, DMOS ∈ [0,100], n 〉=1;

8. adopt and calculate S _disThe identical method of characteristic vector X, the characteristic vector of the stereo-picture of every width distortion in the set of calculated distortion stereo-picture respectively, the characteristic vector for the stereo-picture of i width distortion in the set of distortion stereo-picture is designated as X with it _i, wherein, 1≤i≤n ', the width number of the stereo-picture of the distortion that comprises in the set of n ' expression distortion stereo-picture;

9. adopt support vector regression that the characteristic vector of the stereo-picture of all distortions of identical type of distortion in the set of distortion stereo-picture is trained, and the support vector regression training pattern of utilizing training to obtain is tested the stereo-picture of every width distortion of same type of distortion, calculate the evaluating objective quality predicted value of the stereo-picture of every width distortion of identical type of distortion in the set of distortion stereo-picture, evaluating objective quality predicted value for the stereo-picture of i width distortion in the set of distortion stereo-picture is designated as Q with it _i, Q _i=f (X _i), f () is the function representation form, Q _i=f (X _i) expression Q _iBe X _iFunction, wherein, 1≤i≤n ', the width number of the stereo-picture of the distortion that comprises in the set of n ' expression distortion stereo-picture.

Described step detailed process 2. is:

2.-1, calculate undistorted left visual point image { L _org(x, y) } the visual threshold value set of visual masking effect of background illumination, be designated as { T _l(x, y) },

Wherein, T _l(x, y) represents undistorted left visual point image { L _org(x, y) } in coordinate position be the visual threshold value of visual masking effect of background illumination of the pixel of (x, y),

Represent undistorted left visual point image { L _org(x, y) } in the average brightness of all pixels in 5 * 5 windows centered by pixel take coordinate position as (x, y);

2.-2, calculate undistorted left visual point image { L _org(x, y) } the visual threshold value set of visual masking effect of texture, be designated as { T _t(x, y) }, T _t(x, y)=η * G (x, y) * W _e(x, y), wherein, T _t(x, y) represents undistorted left visual point image { L _org(x, y) } in coordinate position be the visual threshold value of visual masking effect of texture of the pixel of (x, y), η is the controlling elements greater than 0, G (x, y) represents undistorted left visual point image { L _org(x, y) } in coordinate position be that the pixel of (x, y) carries out the maximum weighted mean value that directed high-pass filtering obtains, W _e(x, y) expression is to undistorted left visual point image { L _org(x, y) } edge image in coordinate position be that the pixel of (x, y) carries out the Weighted Edges value that Gassian low-pass filter obtains;

2.-3, to undistorted left visual point image { L _org(x, y) } the visual threshold value set { T of visual masking effect of background illumination _l(x, y) } and the visual threshold value set { T of the visual masking effect of texture _t(x, y) } merge, obtain undistorted left visual point image { L _org(x, y) } minimum discernable distorted image, be designated as { J _L(x, y) }, J _L(x, y)=T _l(x, y)+T _t(x, y)-C _{L, t}* min{T _l(x, y), T _t(x, y) }, wherein, C _{L, t}The parameter of the visual masking effect eclipse effect of background illumination and texture, 0＜C are controlled in expression _{L, t}＜1, min{} is for getting minimum value function;

2. undistorted right visual point image { R is obtained in-4, the employing operation identical with step 2.-1 to 2.-3 _org(x, y) } minimum discernable distorted image, be designated as { J _R(x, y) }.

The detailed process of the regional detection algorithm of described step in 3. is:

3.-1, respectively with undistorted left visual point image { L _org(x, y) } and the left visual point image { L of distortion _dis(x, y) } be divided into

8 * 8 sub-blocks of individual non-overlapping copies define undistorted left visual point image { L _org(x, y) } in l 8 * 8 sub-blocks be current the first sub-block, be designated as

Left visual point image { the L of definition distortion _dis(x, y) } in l 8 * 8 sub-blocks be current the second sub-block, be designated as

Wherein,

(x ₂, y ₂) current the first sub-block of expression

With current the second sub-block

The coordinate position of middle pixel, 1≤x ₂≤ 8,1≤y ₂≤ 8, Represent current the first sub-block

Middle coordinate position is (x ₂, y ₂) the pixel value of pixel,

Represent current the second sub-block

Middle coordinate position is (x ₂, y ₂) the pixel value of pixel;

3.-2, calculate respectively current the first sub-block With current the second sub-block

In the Grad of all pixels, for current the first sub-block

Middle coordinate position is (x ₂', y ₂') pixel, its Grad is designated as P _o(x ₂', y ₂'), P _o(x ₂', y ₂')=| G _ox(x ₂', y ₂') |+| G _oy(x ₂', y ₂') |, for current the second sub-block Middle coordinate position is (x ₂', y ₂') pixel, its Grad is designated as P _d(x ₂', y ₂'), P _d(x ₂', y ₂')=| G _dx(x ₂', y ₂') |+| G _dy(x ₂', y ₂') |, wherein, 1≤x ₂'≤8,1≤y ₂'≤8, G _ox(x ₂', y ₂') current the first sub-block of expression

Middle coordinate position is (x ₂', y ₂') the horizontal gradient value of pixel, G _oy(x ₂', y ₂') current the first sub-block of expression

Middle coordinate position is (x ₂', y ₂') the vertical gradient value of pixel, G _dx(x ₂', y ₂') current the second sub-block of expression

Middle coordinate position is (x ₂', y ₂') the horizontal gradient value of pixel, G _dy(x ₂', y ₂') current the second sub-block of expression

Middle coordinate position is (x ₂', y ₂') the vertical gradient value of pixel, " || " is for asking absolute value sign;

3.-3, find out current the first sub-block

In the maximum of Grad of all pixels, be designated as G _max, then according to G _maxCalculate the first Grads threshold and the second Grads threshold, be designated as respectively T ₁And T ₂, T ₁=0.12 * G _max, T ₂=0.06 * G _max

3.-4, for current the first sub-block

Middle coordinate position is (x ₂', y ₂') pixel and current the second sub-block

Middle coordinate position is (x ₂', y ₂') pixel, the judgement P _o(x ₂', y ₂')＞T ₁And P _d(x ₂', y ₂')＞T ₁Whether set up, if so, judge current the first sub-block

Middle coordinate position is (x ₂', y ₂') pixel be strong fringe region, Num ₁=Nun ₁+ 1, then execution in step 3.-8, otherwise, execution in step 3.-5, wherein, Num ₁Initial value be 0;

3.-5, judgement P _o(x ₂', y ₂')＞T ₁And P _d(x ₂', y ₂')＜=T ₁, perhaps P _d(x ₂', y ₂')＞T ₁And P _o(x ₂', y ₂')＜=T ₁Whether set up, if so, judge current the first sub-block Middle coordinate position is (x ₂', y ₂') pixel and current the second sub-block

Middle coordinate position is (x ₂', y ₂') pixel be weak fringe region, Num ₂=Num ₂+ 1, then execution in step 3.-8, otherwise, execution in step 3.-6, wherein, Num ₂Initial value be 0;

3.-6, judgement P _o(x ₂', y ₂')＜T ₂And P _d(x ₂', y ₂')＜T ₁Whether set up, if so, judge current the first sub-block

Middle coordinate position is (x ₂', y ₂') pixel be smooth region, Num ₃=Num ₃+ 1, then execution in step 3.-8, otherwise, execution in step 3.-7, wherein, Num ₃Initial value be 0;

3.-7, judge current the first sub-block

Middle coordinate position is (x ₂', y ₂') pixel be texture region, Num ₄=Num ₄+ 1, wherein, Num ₄Initial value be 0;

3.-8, returning to step 3.-4 continues current the first sub-block With current the second sub-block

In remaining pixel process, until current the first sub-block

With current the second sub-block In 8 * 8 pixels all be disposed;

3.-9, with Num ₁, Num ₂, Num ₃And Num ₄In the corresponding area type of maximum as current the first sub-block

With current the second sub-block

Block type, be designated as p, wherein, p ∈ 1,2,3,4}, p=1 represents Strong edge blocks, p=2 represents weak edge block, p=3 represents smooth block, p=4 represents texture block;

3.-10, make l "=l+1, l=l ", with undistorted left visual point image { L _org(x, y) } in the next one 8 * 8 sub-blocks as current the first sub-block, with the left visual point image { L of distortion _dis(x, y) } in the next one 8 * 8 sub-blocks as current the second sub-block, return to step 3.-2 and continue to carry out, until undistorted left visual point image { L _org(x, y) } and the left visual point image { L of distortion _dis(x, y) } in

8 * 8 sub-blocks of individual non-overlapping copies all are disposed, and obtain undistorted left visual point image { L _org(x, y) } and the left visual point image { L of distortion _dis(x, y) } in the block type of all 8 * 8 sub-blocks, wherein, l " initial value be 0;

3.-11, the employing operation identical with step 3.-1 to 3.-10 obtains undistorted right visual point image { R _org(x, y) } and the right visual point image { R of distortion _dis(x, y) } in the block type of all 8 * 8 sub-blocks.

Described step detailed process 4. is:

4.-1, the left visual point image { L of calculated distortion _dis(x, y) } in all block types be the spatial noise intensity that is used for reflection picture quality of 8 * 8 sub-blocks of k, be designated as { fq _k(x ₂, y ₂), for the left visual point image { L of distortion _dis(x, y) } in block type be that in 8 * 8 sub-blocks of k, coordinate position is (x ₂, y ₂) pixel, use it for the reflection picture quality spatial noise intensity be designated as fq _k(x ₂, y ₂),

{fq}_{k} (x_{2}, y_{2}) = \frac{1}{N_{k}} \underset{(x_{3}, y_{3})}{Σ} \min {(\max (| L_{org} (x_{3}, y_{3}) - L_{dis} (x_{3}, y_{3}) | - J_{L} (x_{3}, y_{3}), 0), {ST}_{k})}^{2},

Wherein, k ∈ { p|1≤p≤4}, fq _k(x ₂, y ₂) expression distortion left visual point image { L _dis(x, y) } in block type be that in 8 * 8 sub-blocks of k, coordinate position is (x ₂, y ₂) the spatial noise intensity that is used for reflection picture quality of pixel, 1≤x ₂≤ 8,1≤y ₂≤ 8, N _kLeft visual point image { the L of expression distortion _dis(x, y) } in block type be the number of 8 * 8 sub-blocks of k, ST _kFor describing the saturation threshold value of error perception, max () is for getting max function, and min () is for getting minimum value function, (x ₃, y ₃) expression distortion left visual point image { L _dis(x, y) } in block type be that in 8 * 8 sub-blocks of k, coordinate position is (x ₂, y ₂) pixel at undistorted left visual point image { L _org(x, y) } or undistorted left visual point image { L _org(x, y) } minimum discernable distorted image { J _L(x, y) } in coordinate position, 1≤x ₃≤ W, 1≤y ₃≤ H, L _org(x ₃, y ₃) expression { L _org(x, y) } in coordinate position be (x ₃, y ₃) the pixel value of pixel, L _dis(x ₃, y ₃) expression { L _dis(x, y) } in coordinate position be (x ₃, y ₃) the pixel value of pixel, J _L(x ₃, y ₃) expression { J _L(x, y) } in coordinate position be (x ₃, y ₃) the pixel value of pixel, " || " is for asking absolute value sign;

4.-2, with the left visual point image { L of distortion _dis(x, y) } in the spatial noise intensity set expression that is used for reflection picture quality of 8 * 8 sub-blocks of various block types be { fq _k(x ₂, y ₂) | 1≤k≤4}, then with { fq _k(x ₂, y ₂) | all elements in 1≤k≤4} is arranged in order and is obtained the First Characteristic vector, is designated as F ₁, wherein, F ₁Dimension be 256;

4.-3, to undistorted left visual point image { L _org(x, y) } and the left visual point image { L of distortion _dis(x, y) } in each 8 * 8 sub-block implement respectively singular value decomposition, obtain respectively undistorted left visual point image { L _org(x, y) } and the left visual point image { L of distortion _dis(x, y) } in each self-corresponding singular value vector of each 8 * 8 sub-block, with undistorted left visual point image { L _org(x, y) } in the singular value vector of l 8 * 8 sub-blocks be designated as

Left visual point image { L with distortion _dis(x, y) } in the singular value vector of l 8 * 8 sub-blocks be designated as

Wherein, the dimension of singular value vector is 8,

1 \leq l \leq \frac{W \times H}{8 \times 8};

4.-4, the left visual point image { L of calculated distortion _dis(x, y) } in all block types be the space structure intensity that is used for reflection picture quality of 8 * 8 sub-blocks of k, be designated as

Wherein, the left visual point image { L of l ' expression distortion _dis(x, y) } in block type be that 8 * 8 sub-blocks of k are at undistorted left visual point image { L _org(x, y) } or undistorted left visual point image { L _org(x, y) } minimum discernable distorted image { J _L(x, y) } in sequence number;

4.-5, with the left visual point image { L of distortion _dis(x, y) } in the space structure intensity set expression that is used for reflection picture quality of 8 * 8 sub-blocks of various block types be

Then will

In all elements arrange in order and obtain the Second Characteristic vector, be designated as F ₂, wherein, F ₂Dimension be 32;

4.-6, with the First Characteristic vector F ₁With the Second Characteristic vector F ₂Form the New Characteristics vector, as the left visual point image { L of distortion _dis(x, y) } the characteristic vector that is used for reflection picture quality, be designated as F _L, F _L=[F ₁, F ₂], wherein, F _LDimension be 288, " [] " is the vector representation symbol, [F ₁, F ₂] represent the First Characteristic vector F ₁With the Second Characteristic vector F ₂Couple together and form a New Characteristics vector;

4.-7, to the right visual point image { R of distortion _dis(x, y) } adopt the operation identical with step 4.-1 to 4.-6, obtain the right visual point image { R of distortion _dis(x, y) } the characteristic vector that is used for reflection picture quality, be designated as F _R, wherein, F _RDimension be 288;

4.-8, to the left visual point image { L of distortion _dis(x, y) } the characteristic vector F that is used for reflection picture quality _LRight visual point image { R with distortion _dis(x, y) } the characteristic vector F that is used for reflection picture quality _RCarry out linear weighted function, obtain S _disThe characteristic vector that is used for reflection picture quality, be designated as F _q, F _q=w _L* F _L+ w _R* F _R, wherein, w _LLeft visual point image { the L of expression distortion _dis(x, y) } weights proportion, w _RRight visual point image { the R of expression distortion _dis(x, y) } weights proportion, w _L+ w _R=1.

Described step detailed process 5. is:

5.-1, calculate respectively undistorted left visual point image { L _org(x, y) } and undistorted right visual point image { R _org(x, y) } absolute difference image, the left visual point image { L of distortion _dis(x, y) } and the right visual point image { R of distortion _dis(x, y) } absolute difference image and undistorted left visual point image { L _org(x, y) } minimum discernable distorted image { J _L(x, y) } and undistorted right visual point image { R _org(x, y) } minimum discernable distorted image { J _R(x, y) } absolute difference image, be designated as respectively { D _org(x, y) }, { D _dis(x, y) } and { Δ J (x, y) }, D _org(x, y)=| L _org(x, y)-R _org(x, y) |, D _dis(x, y)=| L _dis(x, y)-R _dis(x, y) |, Δ J (x, y)=| J _L(x, y)-J _R(x, y) |, wherein, D _org(x, y) represents { D _org(x, y) } in coordinate position be the pixel value of the pixel of (x, y), D _dis(x, y) represents { D _dis(x, y) } in coordinate position be the pixel value of the pixel of (x, y), the middle coordinate position of Δ J (x, y) expression { Δ J (x, y) } is the pixel value of the pixel of (x, y), " || " is for asking absolute value sign;

5. { D is obtained respectively in-2,3. identical operation of employing and step _org(x, y) } and { D _dis(x, y) } in the block type of each 8 * 8 sub-block;

5.-3, calculate { D _dis(x, y) } in all block types be the spatial noise intensity that is used for the reflection depth perception of 8 * 8 sub-blocks of k, be designated as { fd _k(x ₂, y ₂), for { D _dis(x, y) } in block type be that in 8 * 8 sub-blocks of k, coordinate position is (x ₂, y ₂) pixel, use it for the reflection depth perception spatial noise intensity be designated as fd _k(x ₂, y ₂),

{fd}_{k} (x_{2}, y_{2}) = \frac{1}{M_{k}} \underset{(x_{3}, y_{3})}{Σ} \min {(\max (| D_{org} (x_{4}, y_{4}) - D_{dis} (x_{4}, y_{4}) | - ΔJ (x_{4}, y_{4}), 0), {ST}_{k})}^{2},

Wherein, fd _k(x ₂, y ₂) expression { D _dis(x, y) } in block type be that in 8 * 8 sub-blocks of k, coordinate position is (x ₂, y ₂) the spatial noise intensity that is used for the reflection depth perception of pixel, 1≤x ₂≤ 8,1≤y ₂≤ 8, M _kExpression { D _dis(x, y) } in block type be the number of 8 * 8 sub-blocks of k, ST _kFor describing the saturation threshold value of error perception, (x ₄, y ₄) expression { D _dis(x, y) } in block type be that in 8 * 8 sub-blocks of k, coordinate position is (x ₂, y ₂) pixel at { D _org(x, y) } or { Δ J (x, y) } in coordinate position, 1≤x ₄≤ W, 1≤y ₄≤ H, D _org(x ₄, y ₄) expression { D _org(x, y) } in coordinate position be (x ₄, y ₄) the pixel value of pixel, D _dis(x ₄, y ₄) expression { D _dis(x, y) } in coordinate position be (x ₄, y ₄) the pixel value of pixel, Δ J (x ₄, y ₄) represent that in { Δ J (x, y) }, coordinate position is (x ₄, y ₄) the pixel value of pixel;

5.-4, with { D _dis(x, y) } in the spatial noise intensity that is used for the reflection depth perception of 8 * 8 sub-blocks of various block types with set expression { fd _k(x ₂, y ₂) | 1≤k≤4}, then with { fd _k(x ₂, y ₂) | all elements in 1≤k≤4} is arranged in order and is obtained the 3rd characteristic vector, is designated as F ₃, wherein, F ₃Dimension be 256;

5.-5, to { D _org(x, y) } and { D _dis(x, y) } in each 8 * 8 sub-block implement respectively singular value decomposition, obtain respectively { D _org(x, y) } and { D _dis(x, y) } in each self-corresponding singular value vector of each 8 * 8 sub-block, with { D _org(x, y) } in the singular value vector of l 8 * 8 sub-blocks be designated as

With { D _dis(x, y) } in the singular value vector of l 8 * 8 sub-blocks be designated as

Wherein, the dimension of singular value vector is 8,

5.-6, calculate { D _dis(x, y) } in all block types be the space structure intensity that is used for the reflection depth perception of 8 * 8 sub-blocks of k, be designated as Wherein, l " expression { D _dis(x, y) } in block type be that 8 * 8 sub-blocks of k are at { D _org(x, y) } or { Δ J (x, y) } in sequence number;

5.-7, with { D _dis(x, y) } in the space structure intensity set expression that is used for the reflection depth perception of 8 * 8 sub-blocks of various block types be

Then will

In all elements arrange in order and obtain the 4th characteristic vector, be designated as F ₄, wherein, F ₄Dimension be 32;

5.-8, with the 3rd characteristic vector F ₃With the 4th characteristic vector F ₄Form the New Characteristics vector, as S _disThe characteristic vector that is used for the reflection depth perception, be designated as F _s, F _s=[F ₃, F ₄], wherein, F _sDimension be 288, " [] " is the vector representation symbol, [F ₃, F ₄] represent the 3rd characteristic vector F ₃With the 4th characteristic vector F ₄Couple together and form a New Characteristics vector.

Described step detailed process 9. is:

9.-1, the stereo-picture with all distortions of same type of distortion in the set of distortion stereo-picture is divided into mutually disjoint 5 groups of subsets, selects arbitrarily 4 groups of subset composing training sample datas set wherein, is designated as Ω _q, { X _k, DMOS _k∈ Ω _q, wherein, q represents training sample data set omega _qIn the width number of stereo-picture of the distortion that comprises, X _kExpression training sample data set omega _qIn the characteristic vector of stereo-picture of k width distortion, DMOS _kExpression training sample data set omega _qIn the average subjective scoring difference of stereo-picture of k width distortion, 1≤k≤q;

9.-2, structure X _kRegression function f (X _k),

Wherein, f () is the function representation form, and w is weight vector, w ^TBe the transposed matrix of w, b is bias term,

Expression training sample data set omega _qIn the characteristic vector X of stereo-picture of k width distortion _kLinear function,

D(X _k, X _l) be the kernel function in support vector regression,

X _lBe training sample data set omega _qIn the characteristic vector of stereo-picture of l width distortion, γ is nuclear parameter, is used for the scope of reflection input sample value, the scope of sample value is larger, and the γ value is also just larger, the exponential function of exp () expression take e the end of as, e=2.71828183, " || || for asking the Euclidean distance symbol;

9.-3, adopt support vector regression to training sample data set omega _qIn the characteristic vector of stereo-picture of all distortion train, make the regression function value and the error between average subjective scoring difference that obtain through training minimum, match obtains optimum weight vector w ^optBias term b with optimum ^opt, with the weight vector w of optimum ^optBias term b with optimum ^optCombination be designated as (w ^opt, b ^opt),

(w^{opt}, b^{opt}) = \underset{(w, b) &Element; Ψ}{\arg \min} Σ_{k = 1}^{q} {(f (X_{k}) - {DMOS}_{k})}^{2},

The weight vector w of the optimum that utilization obtains ^optBias term b with optimum ^optStructure support vector regression training pattern is designated as

Wherein, ψ represents training sample data set omega _qIn the set of combination of the characteristic vector of stereo-picture of all distortion all weight vector of training and bias term,

Expression minimizes probability density function, X _inpExpress support for the input vector of vector regression training pattern, (w ^opt) ^TBe w ^optTransposed matrix, Express support for the input vector X of vector regression training pattern _inpLinear function;

9.-4, according to the support vector regression training pattern, the stereo-picture that remains the every width distortion in 1 group of subset is tested, prediction obtains the evaluating objective quality predicted value of the stereo-picture of every width distortion in this group subset, evaluating objective quality predicted value for the stereo-picture of j width distortion in this group subset is designated as Q with it _j, Q _j=f (X _j),

Wherein, X _jThe characteristic vector that represents the stereo-picture of j width distortion in this group subset,

The linear function that represents the stereo-picture of j width distortion in this group subset;

9.-5, according to the process of step 9.-1 to 9.-4, respectively the stereo-picture of all distortions of different type of distortion in the set of distortion stereo-picture is trained, obtain the evaluating objective quality predicted value of the stereo-picture of every width distortion in the set of distortion stereo-picture.

Described step 4. with step 8. in the characteristic vector process that is used for reflection picture quality of the stereo-picture that calculates the JPEG compression artefacts, get w _L=0.50, w _R=0.50; In the characteristic vector process that is used for reflection picture quality of the stereo-picture that calculates the JPEG2000 compression artefacts, get w _L=0.15, w _R=0.85; In the characteristic vector process that is used for reflection picture quality of the stereo-picture that calculates the Gaussian Blur distortion, get w _L=0.10, w _R=0.90; In the characteristic vector process that is used for reflection picture quality of the stereo-picture that calculates the white noise distortion, get w _L=0.20, w _R=0.80; Calculating the H.264 characteristic vector process that is used for reflection picture quality of the stereo-picture of coding distortion, get w _L=0.10, w _R=0.90.

Compared with prior art, the invention has the advantages that:

1) the inventive method considers that zones of different has different responses to three-dimensional perception, stereo-picture is divided into Strong edge blocks, weak edge block, flat block and texture block and estimates respectively, simultaneously picture quality and depth perception information are attached in evaluation procedure, make evaluation result more feel to meet the human visual system.

2) the inventive method obtains minimum discernable distorted image according to the visual characteristic of human eye, and extract respectively the characteristic information of zones of different piece and form the characteristic vector of stereo-picture by computer memory noise intensity and space structure intensity, the characteristic vector information of the stereo-picture that obtains has stronger stability and can reflect preferably the mass change situation of stereo-picture, has improved the correlation of objective evaluation result and subjective perception.

Description of drawings

Fig. 1 be the inventive method totally realize block diagram;

Fig. 2 a is the left visual point image of Akko (being of a size of 640 * 480) stereo-picture;

Fig. 2 b is the right visual point image of Akko (being of a size of 640 * 480) stereo-picture;

Fig. 3 a is the left visual point image of Altmoabit (being of a size of 1024 * 768) stereo-picture;

Fig. 3 b is the right visual point image of Altmoabit (being of a size of 1024 * 768) stereo-picture;

Fig. 4 a is the left visual point image of Balloons (being of a size of 1024 * 768) stereo-picture;

Fig. 4 b is the right visual point image of Balloons (being of a size of 1024 * 768) stereo-picture;

Fig. 5 a is the left visual point image of Doorflower (being of a size of 1024 * 768) stereo-picture;

Fig. 5 b is the right visual point image of Doorflower (being of a size of 1024 * 768) stereo-picture;

Fig. 6 a is the left visual point image of Kendo (being of a size of 1024 * 768) stereo-picture;

Fig. 6 b is the right visual point image of Kendo (being of a size of 1024 * 768) stereo-picture;

Fig. 7 a is the left visual point image of LeaveLaptop (being of a size of 1024 * 768) stereo-picture;

Fig. 7 b is the right visual point image of LeaveLaptop (being of a size of 1024 * 768) stereo-picture;

Fig. 8 a is the left visual point image of Lovebierd1 (being of a size of 1024 * 768) stereo-picture;

Fig. 8 b is the right visual point image of Lovebierd1 (being of a size of 1024 * 768) stereo-picture;

Fig. 9 a is the left visual point image of Newspaper (being of a size of 1024 * 768) stereo-picture;

Fig. 9 b is the right visual point image of Newspaper (being of a size of 1024 * 768) stereo-picture;

Figure 10 a is the left visual point image of Puppy (being of a size of 720 * 480) stereo-picture;

Figure 10 b is the right visual point image of Puppy (being of a size of 720 * 480) stereo-picture;

Figure 11 a is the left visual point image of Soccer2 (being of a size of 720 * 480) stereo-picture;

Figure 11 b is the right visual point image of Soccer2 (being of a size of 720 * 480) stereo-picture;

Figure 12 a is the left visual point image of Horse (being of a size of 720 * 480) stereo-picture;

Figure 12 b is the right visual point image of Horse (being of a size of 720 * 480) stereo-picture;

Figure 13 a is the left visual point image of Xmas (being of a size of 640 * 480) stereo-picture;

Figure 13 b is the right visual point image of Xmas (being of a size of 640 * 480) stereo-picture.

Embodiment

Embodiment is described in further detail the present invention below in conjunction with accompanying drawing.

A kind of objective evaluation method for quality of stereo images based on visually-perceptible that the present invention proposes, it totally realizes block diagram as shown in Figure 1, it mainly comprises the following steps:

1. make S _orgUndistorted stereo-picture for original makes S _disFor the stereo-picture of distortion to be evaluated, with S _orgLeft visual point image be designated as { L _org(x, y) }, with S _orgRight visual point image be designated as { R _org(x, y) }, with S _disLeft visual point image be designated as { L _dis(x, y) }, with S _disRight visual point image be designated as { R _dis(x, y) }, wherein, the coordinate position of pixel in (x, y) left visual point image of expression and right visual point image, 1≤x≤W, 1≤y≤H, W represent the width of left visual point image and right visual point image, H represents the height of left visual point image and right visual point image, L _org(x, y) represents S _orgLeft visual point image { L _org(x, y) } in coordinate position be the pixel value of the pixel of (x, y), R _org(x, y) represents S _orgRight visual point image { R _org(x, y) } in coordinate position be the pixel value of the pixel of (x, y), L _dis(x, y) represents S _disLeft visual point image { L _dis(x, y) } in coordinate position be the pixel value of the pixel of (x, y), R _dis(x, y) represents S _disRight visual point image { R _dis(x, y) } in coordinate position be the pixel value of the pixel of (x, y).

2. human visual system (HVS) characteristic shows, human eye is non to changing less attribute or noise in image, unless the change intensity of this attribute or noise surpasses a certain threshold value, this threshold value is exactly minimum discernable distortion (Just noticeable distortion, JND).And the visual masking effect of human eye is a kind of local effect, is subjected to the impact of the factors such as background illuminance, Texture complication, and background is brighter, and texture is more complicated, and boundary value is just higher.Therefore the present invention utilizes human vision to the visual masking effect of background illumination and texture, extracts respectively undistorted left visual point image { L _org(x, y) } and undistorted right visual point image { R _org(x, y) } minimum discernable distorted image, with undistorted left visual point image { L _org(x, y) } minimum discernable distorted image be designated as { J _L(x, y) }, with undistorted right visual point image { R _org(x, y) } minimum discernable distorted image be designated as { J _R(x, y) }, wherein, J _L(x, y) represents { J _L(x, y) } in coordinate position be the pixel value of the pixel of (x, y), J _R(x, y) represents { J _R(x, y) } in coordinate position be the pixel value of the pixel of (x, y).

In this specific embodiment, step detailed process 2. is:

2.-1, calculate undistorted left visual point image { L _org(x, y) } the visual threshold value set of visual masking effect of background illumination, be designated as { T _t(x, y) },

Represent undistorted left visual point image { L _org(x, y) } in take coordinate position as (x, the average brightness of all pixels in 5 * 5 windows centered by pixel y), in actual process, also can adopt other big or small window, but through great many of experiments, result can obtain best effect when showing the window that adopts 5 * 5 sizes.

2.-2, calculate undistorted left visual point image { L _org(x, y) } the visual threshold value set of visual masking effect of texture, be designated as { T _t(x, y) }, T _t(x, y)=η * G (x, y) * W _e(x, y), wherein, T _t(x, y) represents undistorted left visual point image { L _org(x, y) } in coordinate position be the visual threshold value of visual masking effect of texture of the pixel of (x, y), η is the controlling elements greater than 0, in the present embodiment, η=0.05, G (x, y) represents undistorted left visual point image { L _org(x, y) } in coordinate position be that the pixel of (x, y) carries out the maximum weighted mean value that directed high-pass filtering obtains, W _e(x, y) expression is to undistorted left visual point image { L _org(x, y) } edge image in coordinate position be that the pixel of (x, y) carries out the Weighted Edges value that Gassian low-pass filter obtains.

2.-3, to undistorted left visual point image { L _org(x, y) } the visual threshold value set { T of visual masking effect of background illumination _l(x, y) } and the visual threshold value set { T of the visual masking effect of texture _t(x, y) } merge, obtain undistorted left visual point image { L _org(x, y) } minimum discernable distorted image, be designated as { J _L(x, y) }, J _L(x, y)=T _l(x, y)+T _t(x, y)-C _{L, t}* min{T _l(x, y), T _t(x, y) }, wherein, C _{L, t}The parameter of the visual masking effect eclipse effect of background illumination and texture, 0＜C are controlled in expression _{L, t}＜1, in the present embodiment, C _{L, t}=0.5, min{} is for getting minimum value function.

3. because the human visual system has different susceptibilitys to edge, texture, the smooth region of image, zones of different also should be different to the responsiveness of three-dimensional perception, therefore, should consider respectively the contribution of zones of different to estimating in the stereo image quality evaluation.The present invention obtains respectively undistorted left visual point image { L by regional detection algorithm _org(x, y) } and the left visual point image { L of distortion _dis(x, y) }, undistorted right visual point image { R _org(x, y) } and the right visual point image { R of distortion _dis(x, y) } in the block type of each 8 * 8 sub-block, be designated as p, wherein, p ∈ 1,2,3,4}, p=1 represents Strong edge blocks, p=2 represents weak edge block, p=3 represents smooth block, p=4 represents texture block.

In this specific embodiment, the detailed process of the regional detection algorithm of step in 3. is:

Wherein,

(x ₂, y ₂) current the first sub-block of expression

With current the second sub-block

The coordinate position of middle pixel, 1≤x ₂≤ 8,1≤y ₂≤ 8,

Represent current the first sub-block

Middle coordinate position is (x ₂, y ₂) the pixel value of pixel,

Represent current the second sub-block

Middle coordinate position is (x ₂, y ₂) the pixel value of pixel.

In the Grad of all pixels, for current the first sub-block

Middle coordinate position is (x ₂', y ₂') pixel, its Grad is designated as P _o(x ₂', y ₂'), P _o(x ₂', y ₂')=| G _ox(x ₂', y ₂') |+| G _oy(x ₂', y ₂') |, for current the second sub-block

Middle coordinate position is (x ₂', y ₂') pixel, its Grad is designated as P _d(x ₂', y ₂'), P _d(x ₂', y ₂')=| G _dx(x ₂', y ₂') |+G _dy(x ₂', y ₂') |, wherein, 1≤x ₂'≤8,1≤y ₂'≤8, G _ox(x ₂', y ₂') current the first sub-block of expression Middle coordinate position is (x ₂', y ₂') the horizontal gradient value of pixel, G _oy(x ₂', y ₂') current the first sub-block of expression Middle coordinate position is (x ₂', y ₂') the vertical gradient value of pixel, G _dx(x ₂', y ₂') current the second sub-block of expression

Middle coordinate position is (x ₂', y ₂') the vertical gradient value of pixel, " || " is for asking absolute value sign.

3.-3, find out current the first sub-block In the maximum of Grad of all pixels, be designated as G _max, then according to G _maxCalculate the first Grads threshold and the second Grads threshold, be designated as respectively T ₁And T ₂, T ₁=0.12 * G _max, T ₂=0.06 * G _max

3.-4, for current the first sub-block

Middle coordinate position is (x ₂', y ₂') pixel be strong fringe region, Num ₁=Num ₁+ 1, then execution in step 3.-8, otherwise, execution in step 3.-5, wherein, Num ₁Initial value be 0.

Middle coordinate position is (x ₂', y ₂') pixel be weak fringe region, Num ₂=Num ₂+ 1, then execution in step 3.-8, otherwise, execution in step 3.-6, wherein, Num ₂Initial value be 0.

Middle coordinate position is (x ₂', y ₂') pixel be smooth region, Num ₃=Num ₃+ 1, then execution in step 3.-8, otherwise, execution in step 3.-7, wherein, Num ₃Initial value be 0.

3.-7, judge current the first sub-block

Middle coordinate position is (x ₂', y ₂') pixel be texture region, Num ₄=Num ₄+ 1, wherein, Num ₄Initial value be 0.

3.-8, returning to step 3.-4 continues current the first sub-block

With current the second sub-block

In remaining pixel process, until current the first sub-block With current the second sub-block

In 8 * 8 pixels all be disposed.

With current the second sub-block

Block type, be designated as p, wherein, p ∈ 1,2,3,4}, p=1 represents Strong edge blocks, p=2 represents weak edge block, p=3 represents smooth block, p=4 represents texture block.

8 * 8 sub-blocks of individual non-overlapping copies all are disposed, and obtain undistorted left visual point image { L _org(x, y) } and the left visual point image { L of distortion _dis(x, y) } in the block type of all 8 * 8 sub-blocks, wherein, l " initial value be 0.

4. because the quality of stereo image quality is directly relevant with the left and right view-point image quality, introduce the correlation that the vision perception characteristics such as visual sensitivity, multichannel characteristic, masking effect can improve evaluation model and subjective scoring in image quality evaluation, consider distortion perceptibility and perception saturation phenomenon, with the discernable distorted image of minimum as vision perception characteristic.The present invention is according to undistorted left visual point image { L _org(x, y) } minimum discernable distorted image { J _L(x, y) } and undistorted right visual point image { R _org(x, y) } minimum discernable distorted image { J _R(x, y) }, by the left visual point image { L of calculated distortion _dis(x, y) } in various block types 8 * 8 sub-blocks the spatial noise intensity that is used for reflection picture quality and be used for the space structure intensity of reflection picture quality, and the right visual point image { R of distortion _dis(x, y) } in various block types 8 * 8 sub-blocks the spatial noise intensity that is used for reflection picture quality and be used for the space structure intensity of reflection picture quality, obtain respectively the left visual point image { L of distortion _dis(x, y) } the characteristic vector that is used for reflection picture quality and the right visual point image { R of distortion _dos(x, y) } the characteristic vector that is used for reflection picture quality, then to the left visual point image { L of distortion _dis(x, y) } and the right visual point image { R of distortion _dis(x, y) } the characteristic vector that is used for reflection picture quality carry out linear weighted function, obtain S _disThe characteristic vector that is used for reflection picture quality, be designated as F _q

In this specific embodiment, step detailed process 4. is:

{fq}_{k} (x_{2}, y_{2}) = \frac{1}{N_{k}} \underset{(x_{3}, y_{3})}{Σ} \min {(\max (| L_{org} (x_{3}, y_{3}) - L_{dis} (x_{3}, y_{3}) | - J_{L} (x_{3}, y_{3}), 0), {ST}_{k})}^{2},

Wherein, k ∈ { p|1≤p≤4}, fq _k(x ₂, y ₂) expression distortion left visual point image { L _dis(x, y) } in block type be that in 8 * 8 sub-blocks of k, coordinate position is (x ₂, y ₂) the spatial noise intensity that is used for reflection picture quality of pixel, 1≤x ₂≤ 8,1≤y ₂≤ 8, N _kLeft visual point image { the L of expression distortion _dis(x, y) } in block type be the number of 8 * 8 sub-blocks of k, ST _kFor describing the saturation threshold value of error perception, in the present embodiment, ST _k=30, max () is for getting max function, and min () is for getting minimum value function, (x ₃, y ₃) expression distortion left visual point image { L _dis(x, y) } in block type be that in 8 * 8 sub-blocks of k, coordinate position is (x ₂, y ₂) pixel at undistorted left visual point image { L _org(x, y) } or undistorted left visual point image { L _org(x, y) } minimum discernable distorted image { J _L(x, y) } in coordinate position, 1≤x ₃≤ W, 1≤y ₃≤ H, L _org(x ₃, y ₃) expression { L _org(x, y) } in coordinate position be (x ₃, y ₃) the pixel value of pixel, L _dis(x ₃, y ₃) expression { L _dis(x, y) } in coordinate position be (x ₃, y ₃) the pixel value of pixel, J _L(x ₃, y ₃) expression { J _L(x, y) } in coordinate position be (x ₃, y ₃) the pixel value of pixel, " || " is for asking absolute value sign.

4.-2, with the left visual point image { L of distortion _dis(x, y) } in the spatial noise intensity set expression that is used for reflection picture quality of 8 * 8 sub-blocks of various block types be { fq _k(x ₂, y ₂) | 1≤k≤4}, then with { fq _k(x ₂, y ₂) | all elements in 1≤k≤4} is arranged in order and is obtained the First Characteristic vector, is designated as F ₁, wherein, F ₁Dimension be 256.

Left visual point image { L with distortion _dis(x, y) } in the singular value vector of l 8 * 8 sub-blocks be designated as Wherein, the dimension of singular value vector is 8,

1 \leq l \leq \frac{W \times H}{8 \times 8} .

Wherein, the left visual point image { L of l ' expression distortion _dis(x, y) } in block type be that 8 * 8 sub-blocks of k are at undistorted left visual point image { L _org(x, y) } or undistorted left visual point image { L _org(x, y) } minimum discernable distorted image { J _L(x, y) } in sequence number.

Then will

In all elements arrange in order and obtain the Second Characteristic vector, be designated as F ₂, wherein, F ₂Dimension be 32.

4.-6, with the First Characteristic vector F ₁With the Second Characteristic vector F ₂Form the New Characteristics vector, as the left visual point image { L of distortion _dis(x, y) } the characteristic vector that is used for reflection picture quality, be designated as F _L, F _L=[F ₁, F ₂], wherein, F _LDimension be 288, " [] " is the vector representation symbol, [F ₁, F ₂] represent the First Characteristic vector F ₁With the Second Characteristic vector F ₂Couple together and form a New Characteristics vector.

4.-7, to the right visual point image { R of distortion _dis(x, y) } adopt the operation identical with step 4.-1 to 4.-6, obtain the right visual point image { R of distortion _dis(x, y) } the characteristic vector that is used for reflection picture quality, be designated as F _R, wherein, F _RDimension be 288.

5. existing studies show that, when the difference of the absolute difference image of the left and right visual point image of the absolute difference image of undistorted left and right visual point image and distortion surpasses some threshold values, the decline of depth perception is just easily discovered by human eye, therefore, can estimate with the similarity of the absolute difference image of the left and right visual point image of the absolute difference image of undistorted left and right visual point image and distortion the depth perception of stereo-picture, absolute difference image is more similar, and depth perception is stronger.Therefore, the present invention is according to undistorted left visual point image { L _org(x, y) } minimum discernable distorted image { J _L(x, y) } and undistorted right visual point image { R _org(x, y) } minimum discernable distorted image { J _R(x, y) }, by the left visual point image { L of calculated distortion _dis(x, y) } and the right visual point image { R of distortion _dis(x, y) } absolute difference image in various block types 8 * 8 sub-blocks the spatial noise intensity that is used for the reflection depth perception and be used for the space structure intensity of reflection depth perception, obtain S _disThe characteristic vector that is used for the reflection depth perception, be designated as F _s

In this specific embodiment, step detailed process 5. is:

5.-1, calculate respectively undistorted left visual point image { L _org(x, y) } and undistorted right visual point image { R _org(x, y) } absolute difference image, the left visual point image { L of distortion _dis(x, y) } and the right visual point image { R of distortion _dis(x, y) } absolute difference image and undistorted left visual point image { L _org(x, y) } minimum discernable distorted image { J _L(x, y) } and undistorted right visual point image { R _org(x, y) } minimum discernable distorted image { J _R(x, y) } absolute difference image, be designated as respectively { D _org(x, y) }, { D _dis(x, y) } and { Δ J (x, y) }, D _org(x, y)=| L _org(x, y)-R _org(x, y) |, D _dis(x, y)=| L _dis(x, y)-R _dis(x, y) |, Δ J (x, y)=| J _L(x, y)-J _R(x, y) |, wherein, D _org(x, y) represents { D _org(x, y) } in coordinate position be the pixel value of the pixel of (x, y), D _dis(x, y) represents { D _dis(x, y) } in coordinate position be the pixel value of the pixel of (x, y), the middle coordinate position of Δ J (x, y) expression { Δ J (x, y) } is the pixel value of the pixel of (x, y), " || " is for asking absolute value sign.

5. { D is obtained respectively in-2,3. identical operation of employing and step _org(x, y) } and { D _dis(x, y) } in the block type of each 8 * 8 sub-block.

{fd}_{k} (x_{2}, y_{2}) = \frac{1}{M_{k}} \underset{(x_{3}, y_{3})}{Σ} \min {(\max (| D_{org} (x_{4}, y_{4}) - D_{dis} (x_{4}, y_{4}) | - ΔJ (x_{4}, y_{4}), 0), {ST}_{k})}^{2},

Wherein, fd _k(x ₂, y ₂) expression { D _dis(x, y) } in block type be that in 8 * 8 sub-blocks of k, coordinate position is (x ₂, y ₂) the spatial noise intensity that is used for the reflection depth perception of pixel, 1≤x ₂≤ 8,1≤y ₂≤ 8, M _kExpression { D _dis(x, y) } in block type be the number of 8 * 8 sub-blocks of k, ST _kFor describing the saturation threshold value of error perception, (x ₄, y ₄) expression { D _dis(x, y) } in block type be that in 8 * 8 sub-blocks of k, coordinate position is (x ₂, y ₂) pixel at { D _org(x, y) } or { Δ J (x, y) } in coordinate position, 1≤x ₄≤ W, 1≤y ₄≤ H, D _org(x ₄, y ₄) expression { D _org(x, y) } in coordinate position be (x ₄, y ₄) the pixel value of pixel, D _dis(x ₄, y ₄) expression { D _dis(x, y) } in coordinate position be (x ₄, y ₄) the pixel value of pixel, Δ J (x ₄, y ₄) represent that in { Δ J (x, y) }, coordinate position is (x ₄, y ₄) the pixel value of pixel.

5.-4, with { D _dis(x, y) } in the spatial noise intensity that is used for the reflection depth perception of 8 * 8 sub-blocks of various block types with set expression { fd _k(x ₂, y ₂) | 1≤k≤4}, then with { fd _k(x ₂, y ₂) | all elements in 1≤k≤4} is arranged in order and is obtained the 3rd characteristic vector, is designated as F ₃, wherein, F ₃Dimension be 256.

Wherein, the dimension of singular value vector is 8,

5.-6, calculate { D _dis(x, y) } in all block types be the space structure intensity that is used for the reflection depth perception of 8 * 8 sub-blocks of k, be designated as

Wherein, l " expression { D _dis(x, y) } in block type be that 8 * 8 sub-blocks of k are at { D _org(x, y) } or { Δ J (x, y) } in sequence number.

Then will

In all elements arrange in order and obtain the 4th characteristic vector, be designated as F ₄, wherein, F ₄Dimension be 32.

6. with S _disThe characteristic vector F that is used for reflection picture quality _qWith the characteristic vector F that is used for the reflection depth perception _sForm the New Characteristics vector, as S _disCharacteristic vector, be designated as X, X=[F _q, F _s], " [] " is the vector representation symbol, [F _q, F _s] represent characteristic vector F _qWith characteristic vector F _sCouple together and form a New Characteristics vector.

7. adopt n undistorted stereo-picture, set up its distortion stereo-picture set under the different distortion levels of different type of distortion, this distortion stereo-picture set comprises the stereo-picture of several distortions, utilizes the subjective quality assessment method to obtain respectively the average subjective scoring difference of the stereo-picture of every width distortion in the set of distortion stereo-picture, is designated as DMOS, DMOS=100-MOS, wherein, MOS represents the subjective scoring average, DMOS ∈ [0,100], n 〉=1.

in the present embodiment, because the stereo-picture of testing obtains by H.264 encoding, therefore the type of distortion of training sample and test sample book should be consistent in support vector regression, utilize the stereo-picture as Fig. 2 a and Fig. 2 b formation, the stereo-picture that Fig. 3 a and Fig. 3 b consist of, the stereo-picture that Fig. 4 a and Fig. 4 b consist of, the stereo-picture that Fig. 5 a and Fig. 5 b consist of, the stereo-picture that Fig. 6 a and Fig. 6 b consist of, the stereo-picture that Fig. 7 a and Fig. 7 b consist of, the stereo-picture that Fig. 8 a and Fig. 8 b consist of, the stereo-picture that Fig. 9 a and Fig. 9 b consist of, the stereo-picture that Figure 10 a and Figure 10 b consist of, the stereo-picture that Figure 11 a and Figure 11 b consist of, the stereo-picture that Figure 12 a and Figure 12 b consist of, stereo-picture that Figure 13 a and Figure 13 b the consist of undistorted stereo-picture of totally 12 width (n=12) has been set up its distortion stereo-picture set under the different distortion levels of coding distortion type H.264, the stereo-picture of distortion has 72 width in this distortion stereo-picture set.

8. adopt and calculate S _disThe identical method of characteristic vector X, the characteristic vector of the stereo-picture of every width distortion in the set of calculated distortion stereo-picture respectively, the characteristic vector for the stereo-picture of i width distortion in the set of distortion stereo-picture is designated as X with it _i, wherein, 1≤i≤n ', the width number of the stereo-picture of the distortion that comprises in the set of n ' expression distortion stereo-picture.

In this specific embodiment, according to the stereoscopic vision masking effect inconsistent characteristic of human eye to different type of distortion, left visual point image and right visual point image to the stereo-picture of different type of distortion arrange different weights proportion, in the characteristic vector process that is used for reflection picture quality of the stereo-picture that calculates the JPEG compression artefacts, get w _L=0.50, w _R=0.50; In the characteristic vector process that is used for reflection picture quality of the stereo-picture that calculates the JPEG2000 compression artefacts, get w _L=0.15, w _R=0.85; In the characteristic vector process that is used for reflection picture quality of the stereo-picture that calculates the Gaussian Blur distortion, get w _L=0.10, w _R=0.90; In the characteristic vector process that is used for reflection picture quality of the stereo-picture that calculates the white noise distortion, get w _L=0.20, w _R=0.80; Calculating the H.264 characteristic vector process that is used for reflection picture quality of the stereo-picture of coding distortion, get w _L=0.10, w _R=0.90.

9. the characteristic vector due to the stereo-picture of distortion is the higher dimensional space vector, need to construct linear decision function and realize non-linear decision function in former space in higher dimensional space, support vector regression (Support Vector Regression, SVR) is a kind of reasonable method that realizes non-linear higher dimensional space conversion.Therefore the inventive method adopts support vector regression that the characteristic vector of the stereo-picture of all distortions of identical type of distortion in the set of distortion stereo-picture is trained, and the support vector regression training pattern of utilizing training to obtain is tested the stereo-picture of every width distortion of same type of distortion, calculate the evaluating objective quality predicted value of the stereo-picture of every width distortion of identical type of distortion in the set of distortion stereo-picture, evaluating objective quality predicted value for the stereo-picture of i width distortion in the set of distortion stereo-picture is designated as Q with it _i, Q _i=f (X _i), f () is the function representation form, Q _i=f (X _i) expression Q _iBe X _iFunction, wherein, 1≤i≤n ', the width number of the stereo-picture of the distortion that comprises in the set of n ' expression distortion stereo-picture.

In this specific embodiment, step detailed process 9. is:

9.-1, the stereo-picture with all distortions of same type of distortion in the set of distortion stereo-picture is divided into mutually disjoint 5 groups of subsets, selects arbitrarily 4 groups of subset composing training sample datas set wherein, is designated as Ω _q, { X _k, DMOS _k∈ Ω _q, wherein, q represents training sample data set omega _qIn the width number of stereo-picture of the distortion that comprises, X _kExpression training sample data set omega _qIn the characteristic vector of stereo-picture of k width distortion, DMOS _kExpression training sample data set omega _qIn the average subjective scoring difference of stereo-picture of k width distortion, 1≤k≤q.

9.-2, structure X _kRegression function f (X _k), Wherein, f () is the function representation form, and w is weight vector, w ^TBe the transposed matrix of w, b is bias term,

D(X _k, X _l) be the kernel function in support vector regression,

X _lBe training sample data set omega _qIn the characteristic vector of stereo-picture of l width distortion, γ is nuclear parameter, is used for the scope of reflection input sample value, the scope of sample value is larger, and the γ value is also just larger, the exponential function of exp () expression take e the end of as, e=2.71828183, " || || for asking the Euclidean distance symbol.

In the present embodiment, JPEG compression artefacts, JPEG 2000 compression artefacts, Gaussian Blur distortion, white noise distortion and H.264 the γ value of coding distortion get respectively 42,52,54,130 and 116.

(w^{opt}, b^{opt}) = \underset{(w, b) &Element; Ψ}{\arg \min} Σ_{k = 1}^{q} {(f (X_{k}) - {DMOS}_{k})}^{2},

Expression minimizes probability density function, X _inpExpress support for the input vector of vector regression training pattern, (w ^opt) ^TBe w ^optTransposed matrix,

Express support for the input vector X of vector regression training pattern _inpLinear function.

9.-4, according to the support vector regression training pattern, the stereo-picture that remains the every width distortion in 1 group of subset is tested, prediction obtains the evaluating objective quality predicted value of the stereo-picture of every width distortion in this group subset, evaluating objective quality predicted value for the stereo-picture of j width distortion in this group subset is designated as Q with it _j, Q _j=f (X _j), Wherein, X _jThe characteristic vector that represents the stereo-picture of j width distortion in this group subset,

The linear function that represents the stereo-picture of j width distortion in this group subset.

Adopt 12 undistorted stereo-pictures shown in Fig. 2 a to Figure 13 b to analyze objective image quality evaluation predicted value and the average correlation between the subjective scoring difference of the stereo-picture of the distortion that the present embodiment obtains at the stereo-picture of in various degree JPEG compression, JPEG2000 compression, Gaussian Blur, white noise and H.264 312 width distortions in the coding distortion situation.Here, utilize 2 objective parameters commonly used of evaluate image quality evaluating method as evaluation index, be Pearson correlation coefficient (the Correlation Coefficient under the nonlinear regression condition, CC), Spearman coefficient correlation (Rank-Order Correlation Coefficient, ROCC), the stereo-picture of CC reflection distortion is estimated the accuracy of objective models, and ROCC reflects its monotonicity.The objective image evaluation quality predicted value of the stereo-picture of the distortion that will calculate by the present embodiment is done four parameter L ogistic function nonlinear fittings, and the higher explanation method for objectively evaluating of CC and ROCC value is better with average subjective scoring difference correlation.Table 1 has been listed the image quality evaluation predicted value of stereo-picture of the distortion that the present embodiment obtains and correlation between subjective scoring, from the listed data of table 1 as can be known, correlation between the final objective image quality evaluation predicted value of the stereo-picture of the distortion that obtains by the present embodiment and average subjective scoring difference is very high, the result that shows objective evaluation result and human eye subjective perception is more consistent, is enough to illustrate the validity of the inventive method.

Table 2 has provided the image quality evaluation predicted value of the stereo-picture that adopts the distortion that the different characteristic vector obtains and correlation between subjective scoring, as can be seen from Table 2, only adopt evaluation predicted value single or that two characteristic vectors obtain all and between subjective scoring all to have larger correlation, the feature extracting method that the inventive method is described is effective, and the characteristic vector in conjunction with reflection picture quality and depth perception, the evaluation predicted value and the correlation between subjective scoring that obtain are stronger, are enough to illustrate that this method is effective.

The image quality evaluation predicted value of the stereo-picture of the distortion that table 1 the present embodiment obtains and the correlation between subjective scoring

The image quality evaluation predicted value of the stereo-picture of the distortion that table 2 employing different characteristic vector obtains and the correlation between subjective scoring

Claims

1. objective evaluation method for quality of stereo images based on visually-perceptible is characterized in that comprising the following steps:

Described step detailed process 4. is:

{fq}_{k} (x_{2}, y_{2}) = \frac{1}{N_{k}} \underset{(x_{3}, y_{3})}{Σ} \min {(\max (| L_{org} (x_{3}, y_{3}) - L_{dis} (x_{3}, y_{3}) | - J_{L} (x_{3}, y_{3}), 0), {ST}_{k})}^{2},

Wherein, the dimension of singular value vector is 8,

1 \leq l \leq \frac{W \times H}{8 \times 8};

Wherein, l' represents the left visual point image { L of distortion _dis(x, y) } in block type be that 8 * 8 sub-blocks of k are at undistorted left visual point image { L _org(x, y) } or undistorted left visual point image { L _org(x, y) } minimum discernable distorted image { J _L(x, y) } in sequence number;

4.-5, with the left visual point image { L of distortion _dis(x, y) } in the space structure intensity set expression that is used for reflection picture quality of 8 * 8 sub-blocks of various block types be Then will

4.-8, to the left visual point image { L of distortion _dis(x, y) } the characteristic vector F that is used for reflection picture quality _LRight visual point image { R with distortion _dis(x, y) } the characteristic vector F that is used for reflection picture quality _RCarry out linear weighted function, obtain S _disThe characteristic vector that is used for reflection picture quality, be designated as F _q, F _q=w _L* F _L+ w _R* F _R, wherein, w _LLeft visual point image { the L of expression distortion _dis(x, y) } weights proportion, w _RRight visual point image { the R of expression distortion _dis(x, y) } weights proportion, w _L+ w _R=1;

Described step detailed process 5. is:

{fd}_{k} (x_{2}, y_{2}) = \frac{1}{M_{k}} \underset{(x 3, y 3)}{Σ} \min {(\max (| D_{org} (x_{4}, y_{4}) - D_{dis} (x_{4}, y_{4}) | - ΔJ (x_{4}, y_{4}), 0), {ST}_{k})}^{2},

5.-5, to { D _org(x, y) } and { D _dis(x, y) } in each 8 * 8 sub-block implement respectively singular value decomposition, obtain respectively { D _org(x, y) } and { D _dis(x, y) } in each self-corresponding singular value vector of each 8 * 8 sub-block, with { D _org(x, y) } in the singular value vector of l 8 * 8 sub-blocks be designated as With { D _dis(x, y) } in the singular value vector of l 8 * 8 sub-blocks be designated as

Wherein, the dimension of singular value vector is 8,

Wherein, l'' represents { D _dis(x, y) } in block type be that 8 * 8 sub-blocks of k are at { D _org(x, y) } or { Δ J (x, y) } in sequence number;

Then will

5.-8, with the 3rd characteristic vector F ₃With the 4th characteristic vector F ₄Form the New Characteristics vector, as S _disThe characteristic vector that is used for the reflection depth perception, be designated as F _s, F _s=[F ₃, F ₄], wherein, F _sDimension be 288, " [] " is the vector representation symbol, [F ₃, F ₄] represent the 3rd characteristic vector F ₃With the 4th characteristic vector F ₄Couple together and form a New Characteristics vector;

8. adopt and calculate S _disThe identical method of characteristic vector X, the characteristic vector of the stereo-picture of every width distortion in the set of calculated distortion stereo-picture respectively, the characteristic vector for the stereo-picture of i width distortion in the set of distortion stereo-picture is designated as X with it _i, wherein, 1≤i≤n', n' represent the width number of the stereo-picture of the distortion that comprises in the set of distortion stereo-picture;

9. adopt support vector regression that the characteristic vector of the stereo-picture of all distortions of identical type of distortion in the set of distortion stereo-picture is trained, and the support vector regression training pattern of utilizing training to obtain is tested the stereo-picture of every width distortion of same type of distortion, calculate the evaluating objective quality predicted value of the stereo-picture of every width distortion of identical type of distortion in the set of distortion stereo-picture, evaluating objective quality predicted value for the stereo-picture of i width distortion in the set of distortion stereo-picture is designated as Q with it _i, Q _i=f (X _i), f () is the function representation form, Q _i=f (X _i) expression Q _iBe X _iFunction, wherein, 1≤i≤n', n' represent the width number of the stereo-picture of the distortion that comprises in the set of distortion stereo-picture;

Described step detailed process 9. is:

9.-2, structure X _kRegression function f (X _k),

D(X _k, X _l) be the kernel function in support vector regression,

X _lBe training sample data set omega _qIn the characteristic vector of stereo-picture of l width distortion, γ is nuclear parameter, is used for the scope of reflection input sample value, the scope of sample value is larger, and the γ value is also just larger, the exponential function of exp () expression take e the end of as, e=2.71828183, " || || " for asking the Euclidean distance symbol;

(w^{opt}, b^{opt}) = \underset{(w, b) &Element; Ψ}{\arg \min} Σ_{k = 1}^{q} {(f (X_{k}) - {DMOS}_{k})}^{2},

Express support for the input vector X of vector regression training pattern _inpLinear function;

2. a kind of objective evaluation method for quality of stereo images based on visually-perceptible according to claim 1 is characterized in that described step detailed process 2. is:

2.-3, to undistorted left visual point image { L _org(x, y) } the visual threshold value set { T of visual masking effect of background illumination _l(x, y) } and the visual threshold value set { T of the visual masking effect of texture _t(x, y) } merge, obtain undistorted left visual point image { L _org(x, y) } minimum discernable distorted image, be designated as { J _L(x, y) }, J _L(x, y)=T _l(x, y)+T _t(x, y)-C _l,t* min{T _l(x, y), T _t(x, y) }, wherein, C _l,tThe parameter of the visual masking effect eclipse effect of background illumination and texture, 0＜C are controlled in expression _l,t＜1, min{} is for getting minimum value function;

3. a kind of objective evaluation method for quality of stereo images based on visually-perceptible according to claim 1 and 2 is characterized in that the detailed process of the regional detection algorithm during described step is 3. is:

8 * 8 sub-blocks of individual non-overlapping copies define undistorted left visual point image { L _org(x, y) } in l 8 * 8 sub-blocks be current the first sub-block, be designated as { f _l ^o(x ₂, y ₂), the left visual point image { L of definition distortion _dis(x, y) } in l 8 * 8 sub-blocks be current the second sub-block, be designated as { f _l ^d(x ₂, y ₂), wherein,

(x ₂, y ₂) current the first sub-block { f of expression _l ^o(x ₂, y ₂) and current the second sub-block { f _l ^d(x ₂, y ₂) in the coordinate position of pixel, 1≤x ₂≤ 8,1≤y ₂≤ 8, f _l ^o(x ₂, y ₂) current the first sub-block { f of expression _l ^o(x ₂, y ₂) in coordinate position be (x ₂, y ₂) the pixel value of pixel, f _l ^d(x ₂, y ₂) current the second sub-block { f of expression _l ^d(x ₂, y ₂) in coordinate position be (x ₂, y ₂) the pixel value of pixel;

3.-2, calculate respectively current the first sub-block { f _l ^o(x ₂, y ₂) and current the second sub-block { f _l ^d(x ₂, y ₂) in the Grad of all pixels, for current the first sub-block { f _l ^o(x ₂, y ₂) in coordinate position be (x ₂', y ₂') pixel, its Grad is designated as P _o(x ₂', y ₂'), P _o(x ₂', y ₂')=| G _ox(x ₂', y ₂') |+| G _oy(x ₂', y ₂') |, for current the second sub-block { f _l ^d(x ₂, y ₂) in coordinate position be (x ₂', y ₂') pixel, its Grad is designated as P _d(x ₂', y ₂'), P _d(x ₂', y ₂')=| G _dx(x ₂', y ₂') |+| G _dy(x ₂', y ₂') |, wherein, 1≤x ₂'≤8,1≤y ₂'≤8, G _ox(x ₂', y ₂') current the first sub-block { f of expression _l ^o(x ₂, y ₂) in coordinate position be (x ₂', y ₂') the horizontal gradient value of pixel, G _oy(x ₂', y ₂') current the first sub-block { f of expression _l ^o(x ₂, y ₂) in coordinate position be (x ₂', y ₂') the vertical gradient value of pixel, G _dx(x ₂', y ₂') current the second sub-block { f of expression _l ^d(x ₂, y ₂) in coordinate position be (x ₂', y ₂') the horizontal gradient value of pixel, G _dy(x ₂', y ₂') current the second sub-block { f of expression _l ^d(x ₂, y ₂) in coordinate position be (x ₂', y ₂') the vertical gradient value of pixel, " || " is for asking absolute value sign;

3.-3, find out current the first sub-block { f _l ^o(x ₂, y ₂) in the maximum of Grad of all pixels, be designated as G _max, then according to G _maxCalculate the first Grads threshold and the second Grads threshold, be designated as respectively T ₁And T ₂, T ₁=0.12 * G _max, T ₂=0.06 * G _max

3.-4, for current the first sub-block { f _l ^o(x ₂, y ₂) in coordinate position be (x ₂', y ₂') pixel and current the second sub-block { f _l ^d(x ₂, y ₂) in coordinate position be (x ₂', y ₂') pixel, the judgement P _o(x ₂', y ₂') T ₁And P _d(x ₂', y ₂') T ₁Whether set up, if so, judge current the first sub-block { f _l ^o(x ₂, y ₂) in coordinate position be (x ₂', y ₂') pixel and current the second sub-block { f _l ^d(x ₂, y ₂) in coordinate position be (x ₂', y ₂') pixel be strong fringe region, Num ₁=Num ₁+ 1, then execution in step 3.-8, otherwise, execution in step 3.-5, wherein, Num ₁Initial value be 0;

3.-5, judgement P _o(x ₂', y ₂') T ₁And P _d(x ₂', y ₂')＜=T ₁, perhaps P _d(x ₂', y ₂') T ₁And P _o(x ₂', y ₂')＜=T ₁Whether set up, if so, judge current the first sub-block { f _l ^o(x ₂, y ₂) in coordinate position be (x ₂', y ₂') pixel and current the second sub-block { f _l ^d(x ₂, y ₂) in coordinate position be (x ₂', y ₂') pixel be weak fringe region, Num ₂=Num ₂+ 1, then execution in step 3.-8, otherwise, execution in step 3.-6, wherein, Num ₂Initial value be 0;

3.-6, judgement P _o(x ₂', y ₂')＜T ₂And P _d(x ₂', y ₂')＜T ₁Whether set up, if so, judge current the first sub-block { f _l ^o(x ₂, y ₂) in coordinate position be (x ₂', y ₂') pixel and current the second sub-block { f _l ^d(x ₂, y ₂) in coordinate position be (x ₂', y ₂') pixel be smooth region, Num ₃=Num ₃+ 1, then execution in step 3.-8, otherwise, execution in step 3.-7, wherein, Num ₃Initial value be 0;

3.-7, judge current the first sub-block { f _l ^o(x ₂, y ₂) in coordinate position be (x ₂', y ₂') pixel and current the second sub-block { f _l ^d(x ₂, y ₂) in coordinate position be (x ₂', y ₂') pixel be texture region, Num ₄=Num ₄+ 1, wherein, Num ₄Initial value be 0;

3.-8, returning to step 3.-4 continues current the first sub-block { f _l ^o(x ₂, y ₂) and current the second sub-block { f _l ^d(x ₂, y ₂) in remaining pixel process, until current the first sub-block { f _l ^o(x ₂, y ₂) and current the second sub-block { f _l ^d(x ₂, y ₂) in 8 * 8 pixels all be disposed;

3.-9, with Num ₁, Num ₂, Num ₃And Num ₄In the corresponding area type of maximum as current the first sub-block { f _l ^o(x ₂, y ₂) and current the second sub-block { f _l ^d(x ₂, y ₂) block type, be designated as p, wherein, p ∈ 1,2,3,4}, p=1 represents Strong edge blocks, p=2 represents weak edge block, p=3 represents smooth block, p=4 represents texture block;

3.-10, make l''=l+1, l=l'' is with undistorted left visual point image { L _org(x, y) } in the next one 8 * 8 sub-blocks as current the first sub-block, with the left visual point image { L of distortion _dis(x, y) } in the next one 8 * 8 sub-blocks as current the second sub-block, return to step 3.-2 and continue to carry out, until undistorted left visual point image { L _org(x, y) } and the left visual point image { L of distortion _dis(x, y) } in

8 * 8 sub-blocks of individual non-overlapping copies all are disposed, and obtain undistorted left visual point image { L _org(x, y) } and the left visual point image { L of distortion _dis(x, y) } in the block type of all 8 * 8 sub-blocks, wherein, the initial value of l'' is 0;

4. a kind of objective evaluation method for quality of stereo images based on visually-perceptible according to claim 1, it is characterized in that described step 4. with step 8. in the characteristic vector process that is used for reflection picture quality of the stereo-picture that calculates the JPEG compression artefacts, get w _L=0.50, w _R=0.50; In the characteristic vector process that is used for reflection picture quality of the stereo-picture that calculates the JPEG2000 compression artefacts, get w _L=0.15, w _R=0.85; In the characteristic vector process that is used for reflection picture quality of the stereo-picture that calculates the Gaussian Blur distortion, get w _L=0.10, w _R=0.90; In the characteristic vector process that is used for reflection picture quality of the stereo-picture that calculates the white noise distortion, get w _L=0.20, w _R=0.80; Calculating the H.264 characteristic vector process that is used for reflection picture quality of the stereo-picture of coding distortion, get w _L=0.10, w _R=0.90.