CN103824292A - Three-dimensional image quality objective assessment method based on three-dimensional gradient amplitude - Google Patents

Abstract

The invention discloses a three-dimensional image quality objective assessment method based on a three-dimensional gradient amplitude. Firstly, a parallax error space picture of an undistorted three-dimensional image and a parallax error space picture of a distorted three-dimensional image to be assessed are calculated respectively, then a corresponding three-dimensional gradient amplitude is obtained by calculating the horizontal gradient, the vertical gradient and the view point gradient of each pixel point in the parallax error space picture of the undistorted three-dimensional image, a corresponding three-dimensional gradient amplitude is obtained by calculating the horizontal gradient, the vertical gradient and the view point gradient of each pixel point in the parallax error space picture of the distorted three-dimensional image to be assessed, and ultimately, according to the three-dimensional gradient amplitudes of each pixel point in the two parallax error space pictures, an image quality objective assessment prediction value of the distorted three-dimensional image to be assessed is obtained. The method has the advantages that the obtained three-dimensional gradient amplitudes are very stable and can well reflect quality changing conditions of the three-dimensional images and accordingly correlation of an objective assessment result and subjective perception can be effectively improved.

Description

A kind of objective evaluation method for quality of stereo images based on three-dimensional gradient amplitude

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 three-dimensional gradient amplitude.

Background technology

Along with developing rapidly of image coding technique and stereo display technique, stereo-picture technology has been subject to paying close attention to more and more widely and application, has become a current study hotspot.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, is merged and is formed binocular parallax, thereby enjoy the stereo-picture with depth perception and realism by brain.Owing to being subject to the impact of acquisition system, store compressed and transmission equipment, stereo-picture can inevitably be introduced a series of distortion, and compared with single channel image, stereo-picture need to guarantee the picture quality of two passages simultaneously, and therefore stereoscopic image is carried out quality assessment and had very important significance.But, lack at present effective method for objectively evaluating stereoscopic image quality and evaluate.Therefore, setting up effective stereo image quality objective evaluation model tool is of great significance.

Gradient amplitude is a kind of effectively picture structure information descriptor, evaluation method based on gradient amplitude has been applied to plane picture quality assessment, and for the stereo image quality evaluation based on gradient amplitude, need to solve following key issue: 1) three-dimensional perception evaluation reflects by parallax or depth information, how parallax or depth information are embedded in gradient amplitude and know characteristic to characterize truly stereoscopic sensation, remain one of difficulties in stereo image quality objective evaluation; 2) not all pixel all has strong structural information, and How to choose stable structure information is applied to quality assessment, and don't affects three-dimensional perceptual performance, is also the difficulties that needs solution in stereo image quality objective evaluation.

Summary of the invention

Technical matters to be solved by this invention is to provide a kind of objective evaluation method for quality of stereo images based on three-dimensional gradient amplitude, and it can improve the correlativity of objective evaluation result and subjective perception effectively.

The present invention solves the problems of the technologies described above adopted technical scheme: a kind of objective evaluation method for quality of stereo images based on three-dimensional gradient amplitude, is characterized in that comprising the following steps:

1. make S _orgrepresent original undistorted stereo-picture, make S _disrepresent the stereo-picture of distortion to be evaluated, by S _orgleft visual point image be designated as { L _org(x, y) }, by S _orgright visual point image be designated as { R _org(x, y) }, by S _disleft visual point image be designated as { L _dis(x, y) }, by S _disright visual point image be designated as { R _dis(x, y) }, wherein, (x, y) represent the coordinate position of the pixel in left visual point image and right visual point image, 1≤x≤W, 1≤y≤H, W represents the width of left visual point image and right visual point image, and H represents the height of left visual point image and right visual point image, L _org(x, y) represents { L _org(x, y) } in the pixel value of the coordinate position pixel that is (x, y), R _org(x, y) represents { R _org(x, y) } in the pixel value of the coordinate position pixel that is (x, y), L _dis(x, y) represents { L _dis(x, y) } in the pixel value of the coordinate position pixel that is (x, y), R _dis(x, y) represents { R _dis(x, y) } in the pixel value of the coordinate position pixel that is (x, y);

2. according to { L _org(x, y) } in each pixel and { R _org(x, y) } in the pixel of the respective coordinates position disparity space value under multiple parallax value, obtain S _orgdisparity space image, be designated as { DSI _org(x, y, d) }, and according to { L _dis(x, y) } in each pixel and { R _dis(x, y) } in the pixel of the respective coordinates position disparity space value under multiple parallax value, obtain S _disdisparity space image, be designated as { DSI _dis(x, y, d) }, wherein, DSI _org(x, y, d) represents { DSI _org(x, y, d) } in coordinate position be the disparity space value of the pixel of (x, y, d), DSI _dis(x, y, d) represents { DSI _dis(x, y, d) } in coordinate position be the disparity space value of the pixel of (x, y, d), 0≤d≤d _max, d _maxrepresent maximum disparity value;

3. calculate { DSI _org(x, y, d) } in horizontal direction gradient, vertical gradient and the viewpoint direction gradient of each pixel, by { DSI _org(x, y, d) } in coordinate position be that the horizontal direction gradient of the pixel of (x, y, d) is designated as gx _org(x, y, d), by { DSI _org(x, y, d) } in coordinate position be that the vertical gradient of the pixel of (x, y, d) is designated as gy _org(x, y, d), by { DSI _org(x, y, d) } in coordinate position be that the viewpoint direction gradient of the pixel of (x, y, d) is designated as gd _org(x, y, d);

Equally, calculate { DSI _dis(x, y, d) } in horizontal direction gradient, vertical gradient and the viewpoint direction gradient of each pixel, by { DSI _dis(x, y, d) } in coordinate position be that the horizontal direction gradient of the pixel of (x, y, d) is designated as gx _dis(x, y, d), by { DSI _dis(x, y, d) } in coordinate position be that the vertical gradient of the pixel of (x, y, d) is designated as gy _dis(x, y, d), by { DSI _dis(x, y, d) } in coordinate position be that the viewpoint direction gradient of the pixel of (x, y, d) is designated as gd _dis(x, y, d);

4. according to { DSI _org(x, y, d) } in horizontal direction gradient, vertical gradient and the viewpoint direction gradient of each pixel, calculate { DSI _org(x, y, d) } in the three-dimensional gradient amplitude of each pixel, by { DSI _org(x, y, d) } in coordinate position be that the three-dimensional gradient amplitude of the pixel of (x, y, d) is designated as m _org(x, y, d), $m_{\mathrm{org}}(x,y,d)=\sqrt{{\left({\mathrm{gx}}_{\mathrm{org}}(x,y,d)\right)}^2+{\left({\mathrm{gy}}_{\mathrm{org}}(x,y,d)\right)}^2+{\left({\mathrm{gd}}_{\mathrm{org}}(x,y,d)\right)}^2};$

Equally, according to { DSI _dis(x, y, d) } in horizontal direction gradient, vertical gradient and the viewpoint direction gradient of each pixel, calculate { DSI _dis(x, y, d) } in the three-dimensional gradient amplitude of each pixel, by { DSI _dis(x, y, d) } in coordinate position be that the three-dimensional gradient amplitude of the pixel of (x, y, d) is designated as m _dis(x, y, d), $m_{\mathrm{dis}}(x,y,d)=\sqrt{{\left({\mathrm{gx}}_{\mathrm{dis}}(x,y,d)\right)}^2+{\left({\mathrm{gy}}_{\mathrm{dis}}(x,y,d)\right)}^2+{\left({\mathrm{gd}}_{\mathrm{dis}}(x,y,d)\right)}^2};$

5. according to { DSI _org(x, y, d) } and { DSI _dis(x, y, d) } in the three-dimensional gradient amplitude of each pixel, calculate { DSI _dis(x, y, d) } in the objective evaluation metric of each pixel, by { DSI _dis(x, y, d) } in coordinate position be that the objective evaluation metric of the pixel of (x, y, d) is designated as Q _dSI(x, y, d), $Q_{\mathrm{DSI}}(x,y,d)=\frac{2\×m_{\mathrm{org}}(x,y,d)\×m_{\mathrm{dis}}(x,y,d)+C}{{\left(m_{\mathrm{org}}(x,y,d)\right)}^2+{\left(m_{\mathrm{dis}}(x,y,d)\right)}^2+C},$ Wherein, C is for controlling parameter;

6. according to { DSI _dis(x, y, d) } in the objective evaluation metric of each pixel, calculate S _dispicture quality objective evaluation predicted value, be designated as Q,

wherein, Ω represents { DSI _dis(x, y, d) } in the set of coordinate position of all pixels, N represents { DSI _dis(x, y, d) } in total number of the pixel that comprises.

Described step is middle S 2. _orgthe acquisition process of disparity space image be:

2.-a1, by { L _org(x, y) } in current pending pixel be defined as current the first pixel, by { R _org(x, y) } in current pending pixel be defined as current the second pixel;

2.-a2, suppose that current the first pixel is { L _org(x, y) } in coordinate position be (x ₁, y ₁) pixel, suppose that current the second pixel is { R _org(x, y) } in coordinate position be (x ₁, y ₁) pixel, get parallax value d ₀=0, then calculate current the first pixel and current the second pixel at this parallax value d ₀under disparity space value, be designated as DSI _org(x ₁, y ₁, d ₀), DSI _org(x ₁, y ₁, d ₀)=| L _org(x ₁, y ₁)-R _org(x ₁-d ₀, y ₁) |, wherein, 1≤x ₁≤ W, 1≤y ₁≤ H, 0≤d ₀≤ d _max, d _maxrepresent maximum disparity value, L _org(x ₁, y ₁) expression { L _org(x, y) } in coordinate position be (x ₁, y ₁) the pixel value of pixel, R _org(x ₁-d ₀, y ₁) expression { R _org(x, y) } in coordinate position be (x ₁-d ₀, y ₁) the pixel value of pixel, " || " is the symbol that takes absolute value;

2.-a3, choose d _maxindividual and d ₀different parallax value, is designated as respectively

then calculate respectively current the first pixel and current the second pixel at this d _maxdisparity space value under individual different parallax value, corresponding is designated as respectively ${\mathrm{DSI}}_{\mathrm{org}}(x_1,{\mathrm{<figure-callout\; id="1"\; label="y"\; filenames="HDA0000469762880000012.png,HDA0000469762880000021.png"\; state="\{\{state\}\}">y</figure-callout>}}_1,d_1),{\mathrm{DSI}}_{\mathrm{org}}(x_1,{\mathrm{<figure-callout\; id="1"\; label="y"\; filenames="HDA0000469762880000012.png,HDA0000469762880000021.png"\; state="\{\{state\}\}">y</figure-callout>}}_1,d_2),...,{\mathrm{DSI}}_{\mathrm{org}}(x_1,y_1,d_i),...,{\mathrm{DSI}}_{\mathrm{org}}(x_1,y_1,d_{d_{\max }}),$ DSI _org(x ₁,y ₁,d ₁)=|L _org(x ₁,y ₁)-R _org(x ₁-d ₁,y ₁)|，DSI _org(x ₁,y ₁,d ₂)=|L _org(x ₁,y ₁)-R _org(x ₁-d ₂,y ₁)|，DSI _org(x ₁,y ₁,d _i)=|L _org(x ₁,y ₁)-R _org(x ₁-d _i,y ₁)|， ${\mathrm{DSI}}_{\mathrm{org}}(x_1,y_1,d_{d_{\max }})=|L_{\mathrm{org}}(x_1,y_1)-R_{\mathrm{org}}(x_1-d_{d_{\max }},y_1)|,$ Wherein, 1≤i≤d _max, d _i=d ₀+ i,

dSI _org(x ₁, y ₁, d ₁) represent that current the first pixel and current the second pixel are at parallax value d ₁under disparity space value, DSI _org(x ₁, y ₁, d ₂) represent that current the first pixel and current the second pixel are at parallax value d ₂under disparity space value, DSI _org(x ₁, y ₁, d _i) represent that current the first pixel and current the second pixel are at parallax value d _iunder disparity space value,

represent that current the first pixel and current the second pixel are in parallax value

under disparity space value, R _org(x ₁-d ₁, y ₁) expression { R _org(x, y) } in coordinate position be (x ₁-d ₁, y ₁) the pixel value of pixel, R _org(x ₁-d ₂, y ₁) expression { R _org(x, y) } in coordinate position be (x ₁-d ₂, y ₁) the pixel value of pixel, R _org(x ₁-d _i, y ₁) expression { R _org(x, y) } in coordinate position be (x ₁-d _i, y ₁) the pixel value of pixel,

represent { R _org(x, y) } in coordinate position be

the pixel value of pixel;

2.-a4, by { L _org(x, y) } in next pending pixel as current the first pixel, by { R _org(x, y) } in next pending pixel as current the second pixel, then return step 2.-a2 continues execution, until { L _org(x, y) } and { R _org(x, y) } in all pixels be disposed, obtain S _orgdisparity space image, be designated as { DSI _org(x, y, d) }, wherein, DSI _org(x, y, d) represents { DSI _org(x, y, d) } in coordinate position be the disparity space value of the pixel of (x, y, d), DSI _orgthe value of (x, y, d) is { L _org(x, y) } in coordinate position be (x, y) pixel and { R _org(x, y) } in the coordinate position disparity space value of pixel under parallax value d that be (x, y),

Described step is middle S 2. _disthe acquisition process of disparity space image be:

2.-b1, by { L _dis(x, y) } in current pending pixel be defined as current the first pixel, by { R _dis(x, y) } in current pending pixel be defined as current the second pixel;

2.-b2, suppose that current the first pixel is { L _dis(x, y) } in coordinate position be (x ₁, y ₁) pixel, suppose that current the second pixel is { R _dis(x, y) } in coordinate position be (x ₁, y ₁) pixel, get parallax value d ₀=0, then calculate current the first pixel and current the second pixel at this parallax value d ₀under disparity space value, be designated as DSI _dis(x ₁, y ₁, d ₀), DSI _dis(x ₁, y ₁, d ₀)=| L _dis(x ₁, y ₁)-R _dis(x ₁-d ₀, y ₁) |, wherein, 1≤x ₁≤ W, 1≤y ₁≤ H, 0≤d ₀≤ d _max, d _maxrepresent maximum disparity value, L _dis(x ₁, y ₁) expression { L _dis(x, y) } in coordinate position be (x ₁, y ₁) the pixel value of pixel, R _dis(x ₁-d ₀, y ₁) expression { R _dis(x, y) } in coordinate position be (x ₁-d ₀, y ₁) the pixel value of pixel, " || " is the symbol that takes absolute value;

2.-b3, choose d _maxindividual and d ₀different parallax value, is designated as respectively

then calculate respectively current the first pixel and current the second pixel at this d _maxdisparity space value under individual different parallax value, corresponding is designated as respectively ${\mathrm{DSI}}_{\mathrm{dis}}(x_1,{\mathrm{<figure-callout\; id="1"\; label="y"\; filenames="HDA0000469762880000012.png,HDA0000469762880000021.png"\; state="\{\{state\}\}">y</figure-callout>}}_1,d_1),{\mathrm{DSI}}_{\mathrm{dis}}(x_1,{\mathrm{<figure-callout\; id="1"\; label="y"\; filenames="HDA0000469762880000012.png,HDA0000469762880000021.png"\; state="\{\{state\}\}">y</figure-callout>}}_1,d_2),...,{\mathrm{DSI}}_{\mathrm{dis}}(x_1,y_1,d_i),...,{\mathrm{DSI}}_{\mathrm{dis}}(x_1,y_1,d_{d_{\max }}),$ DSI _dis(x ₁,y ₁,d ₁)=|L _dis(x ₁,y ₁)-R _dis(x ₁-d ₁,y ₁)|，DSI _dis(x ₁,y ₁,d ₂)=|L _dis(x ₁,y ₁)-R _dis(x ₁-d ₂,y ₁)|，DSI _dis(x ₁,y ₁,d _i)=|L _dis(x ₁,y ₁)-R _dis(x ₁-d _i,y ₁)|， ${\mathrm{DSI}}_{\mathrm{dis}}(x_1,y_1,d_{d_{\max }})=|L_{\mathrm{dis}}(x_1,y_1)-R_{\mathrm{dis}}(x_1-d_{d_{\max }},y_1)|,$ Wherein, 1≤i≤d _max, d _i=d ₀+ i,

dSI _dis(x ₁, y ₁, d ₁) represent that current the first pixel and current the second pixel are at parallax value d ₁under disparity space value, DSI _dis(x ₁, y ₁, d ₂) represent that current the first pixel and current the second pixel are at parallax value d ₂under disparity space value, DSI _dis(x ₁, y ₁, d _i) represent that current the first pixel and current the second pixel are at parallax value d _iunder disparity space value,

represent that current the first pixel and current the second pixel are in parallax value

under disparity space value, R _dis(x ₁-d ₁, y ₁) expression { R _dis(x, y) } in coordinate position be (x ₁-d ₁, y ₁) the pixel value of pixel, R _dis(x ₁-d ₂, y ₁) expression { R _dis(x, y) } in coordinate position be (x ₁-d ₂, y ₁) the pixel value of pixel, R _dis(x ₁-d _i, y ₁) expression { R _dis(x, y) } in coordinate position be (x ₁-d _i, y ₁) the pixel value of pixel, represent { R _dis(x, y) } in coordinate position be the pixel value of pixel;

2.-b4, by { L _dis(x, y) } in next pending pixel as current the first pixel, by { R _dis(x, y) } in next pending pixel as current the second pixel, then return step 2.-b2 continues execution, until { L _dis(x, y) } and { R _dis(x, y) } in all pixels be disposed, obtain S _disdisparity space image, be designated as { DSI _dis(x, y, d) }, wherein, DSI _dis(x, y, d) represents { DSI _dis(x, y, d) } in coordinate position be the disparity space value of the pixel of (x, y, d), DSI _disthe value of (x, y, d) is { L _dis(x, y) } in coordinate position be (x, y) pixel and { R _dis(x, y) } in the coordinate position disparity space value of pixel under parallax value d that be (x, y),

Described step 3. in { DSI _org(x, y, d) } in the acquisition process of horizontal direction gradient, vertical gradient and viewpoint direction gradient of each pixel be:

3.-a1, employing horizontal gradient operator are to { DSI _org(x, y, d) } carry out convolution, obtain { DSI _org(x, y, d) } in the horizontal direction gradient of each pixel, by { DSI _org(x, y, d) } in coordinate position be that the horizontal direction gradient of the pixel of (x, y, d) is designated as gx _org(x, y, d), ${\mathrm{gx}}_{\mathrm{org}}(x,y,d)=\underset{j=d-2}{\overset{j=d+2}{\mathrm{\&Sigma;}}}(-\underset{u=x-2}{\overset{u=x-1}{\mathrm{\&Sigma;}}}\underset{v=y-2}{\overset{v=y+2}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{org}}(u,v,j)+\underset{u=x+1}{\overset{u=x+2}{\mathrm{\&Sigma;}}}\underset{v=y-2}{\overset{v=y+2}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{org}}(u,v,j)),$ Wherein, DSI _org(u, v, j) represents { DSI _org(x, y, d) } in coordinate position be the disparity space value of the pixel of (u, v, j);

3.-a2, employing VG (vertical gradient) operator are to { DSI _org(x, y, d) } carry out convolution, obtain { DSI _org(x, y, d) } in the vertical gradient of each pixel, by { DSI _org(x, y, d) } in coordinate position be that the vertical gradient of the pixel of (x, y, d) is designated as gy _org(x, y, d), ${\mathrm{gy}}_{\mathrm{org}}(x,y,d)=\underset{j=d-2}{\overset{j=d+2}{\mathrm{\&Sigma;}}}(-\underset{u=x-2}{\overset{u=x+2}{\mathrm{\&Sigma;}}}\underset{v=y-2}{\overset{v=y-1}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{org}}(u,v,j)+\underset{u=x-2}{\overset{u=x+2}{\mathrm{\&Sigma;}}}\underset{v=y+1}{\overset{v=y+2}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{org}}(u,v,j));$

3.-a3, employing viewpoint gradient operator are to { DSI _org(x, y, d) } carry out convolution, obtain { DSI _org(x, y, d) } in the viewpoint direction gradient of each pixel, by { DSI _org(x, y, d) } in coordinate position be that the viewpoint direction gradient of the pixel of (x, y, d) is designated as gd _org(x, y, d), ${\mathrm{gd}}_{\mathrm{org}}(x,y,d)=\underset{j=d-2}{\overset{j=d+2}{\mathrm{\&Sigma;}}}(\mathrm{sign}(j-d)\&times;\underset{u=x-2}{\overset{u=x+2}{\mathrm{\&Sigma;}}}\underset{v=y-2}{\overset{v=y+2}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{org}}(u,v,j)),$ Wherein, sign () is step function,

Above-mentioned steps 3.-a1 to step 3.-a3 in, if u<1, DSI _orgthe value of (u, v, j) is by DSI _orgthe value of (1, v, j) substitutes, if u>W, DSI _orgthe value of (u, v, j) is by DSI _orgthe value of (W, v, j) substitutes, if v<1, DSI _orgthe value of (u, v, j) is by DSI _org(u, 1, value j) substitutes, if v>H, DSI _orgthe value of (u, v, j) is by DSI _orgthe value of (u, H, j) substitutes, if j<0, DSI _orgthe value of (u, v, j) is by DSI _orgthe value of (u, v, 0) substitutes, if j>d _max, DSI _orgthe value of (u, v, j) is by DSI _org(u, v, d _max) value substitute, DSI _org(1, v, j) represents { DSI _org(x, y, d) } in coordinate position be the disparity space value of the pixel of (1, v, j), DSI _org(W, v, j) represents { DSI _org(x, y, d) } in coordinate position be the disparity space value of the pixel of (W, v, j), DSI _org(u, 1, j) represent { DSI _org(x, y, d) } in coordinate position be (u, 1, the disparity space value of pixel j), DSI _org(u, H, j) represents { DSI _org(x, y, d) } in coordinate position be the disparity space value of the pixel of (u, H, j), DSI _org(u, v, 0) represents { DSI _org(x, y, d) } in coordinate position be the disparity space value of the pixel of (u, v, 0), DSI _org(u, v, d _max) expression { DSI _org(x, y, d) } in coordinate position be (u, v, d _max) the disparity space value of pixel;

Described step 3. in { DSI _dis(x, y, d) } in the acquisition process of horizontal direction gradient, vertical gradient and viewpoint direction gradient of each pixel be:

3.-b1, employing horizontal gradient operator are to { DSI _dis(x, y, d) } carry out convolution, obtain { DSI _dis(x, y, d) } in the horizontal direction gradient of each pixel, by { DSI _dis(x, y, d) } in coordinate position be that the horizontal direction gradient of the pixel of (x, y, d) is designated as gx _dis(x, y, d), ${\mathrm{gx}}_{\mathrm{dis}}(x,y,d)=\underset{j=d-2}{\overset{j=d+2}{\mathrm{\&Sigma;}}}(-\underset{u=x-2}{\overset{u=x-1}{\mathrm{\&Sigma;}}}\underset{v=y-2}{\overset{v=y+2}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{dis}}(u,v,j)+\underset{u=x+1}{\overset{u=x+2}{\mathrm{\&Sigma;}}}\underset{v=y-2}{\overset{v=y+2}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{dis}}(u,v,j)),$ Wherein, DSI _dis(u, v, j) represents { DSI _dis(x, y, d) } in coordinate position be the disparity space value of the pixel of (u, v, j);

3.-b2, employing VG (vertical gradient) operator are to { DSI _dis(x, y, d) } carry out convolution, obtain { DSI _dis(x, y, d) } in the vertical gradient of each pixel, by { DSI _dis(x, y, d) } in coordinate position be that the vertical gradient of the pixel of (x, y, d) is designated as gy _dis(x, y, d), ${\mathrm{gy}}_{\mathrm{dis}}(x,y,d)=\underset{j=d-2}{\overset{j=d+2}{\mathrm{\&Sigma;}}}(-\underset{u=x-2}{\overset{u=x+2}{\mathrm{\&Sigma;}}}\underset{v=y-2}{\overset{v=y-1}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{dis}}(u,v,j)+\underset{u=x-2}{\overset{u=x+2}{\mathrm{\&Sigma;}}}\underset{v=y+1}{\overset{v=y+2}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{dis}}(u,v,j));$

3.-b3, employing viewpoint gradient operator are to { DSI _dis(x, y, d) } carry out convolution, obtain { DSI _dis(x, y, d) } in the viewpoint direction gradient of each pixel, by { DSI _dis(x, y, d) } in coordinate position be that the parallax directions gradient of the pixel of (x, y, d) is designated as gd _dis(x, y, d), ${\mathrm{gd}}_{\mathrm{dis}}(x,y,d)=\underset{j=d-2}{\overset{j=d+2}{\mathrm{\&Sigma;}}}(\mathrm{sign}(j-d)\&times;\underset{u=x-2}{\overset{u=x+2}{\mathrm{\&Sigma;}}}\underset{v=y-2}{\overset{v=y+2}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{dis}}(u,v,j)),$ Wherein, sign () is step function,

Above-mentioned steps 3.-b1 to step 3.-b3 in, if u<1, DSI _disthe value of (u, v, j) is by DSI _disthe value of (1, v, j) substitutes, if u>W, DSI _disthe value of (u, v, j) is by DSI _disthe value of (W, v, j) substitutes, if v<1, DSI _disthe value of (u, v, j) is by DSI _dis(u, 1, value j) substitutes, if v>H, DSI _disthe value of (u, v, j) is by DSI _disthe value of (u, H, j) substitutes, if j<0, DSI _disthe value of (u, v, j) is by DSI _disthe value of (u, v, 0) substitutes, if j>d _max, DSI _disthe value of (u, v, j) is by DSI _dis(u, v, d _max) value substitute, DSI _dis(1, v, j) represents { DSI _dis(x, y, d) } in coordinate position be the disparity space value of the pixel of (1, v, j), DSI _dis(W, v, j) represents { DSI _dis(x, y, d) } in coordinate position be the disparity space value of the pixel of (W, v, j), DSI _dis(u, 1, j) represent { DSI _dis(x, y, d) } in coordinate position be (u, 1, the disparity space value of pixel j), DSI _dis(u, H, j) represents { DSI _dis(x, y, d) } in coordinate position be the disparity space value of the pixel of (u, H, j), DSI _dis(u, v, 0) represents { DSI _dis(x, y, d) } in coordinate position be the disparity space value of the pixel of (u, v, 0), DSI _dis(u, v, d _max) expression { DSI _dis(x, y, d) } in coordinate position be (u, v, d _max) the disparity space value of pixel.

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

1) the inventive method is considered the impact of parallax on three-dimensional perception, therefore construct respectively the disparity space image of the disparity space image of original undistorted stereo-picture and the stereo-picture of distortion to be evaluated, avoided so complicated disparity estimation operation, and the disparity space image of constructing can reflect the impact of different parallax stereoscopic image quality well.

2) the inventive method is by calculating horizontal direction gradient, vertical gradient and the viewpoint direction gradient of the each pixel in disparity space image, obtain the three-dimensional gradient amplitude of the each pixel in disparity space image, the three-dimensional gradient amplitude obtaining has stronger stability and can reflect preferably the mass change situation of stereo-picture, therefore can effectively improve the correlativity of objective evaluation result and subjective perception.

Accompanying drawing explanation

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

Fig. 2 is horizontal gradient operator template;

Fig. 3 is VG (vertical gradient) operator template;

Fig. 4 is viewpoint gradient operator template;

Fig. 5 is the picture quality objective evaluation predicted value of stereo-picture and the scatter diagram of average subjective scoring difference that utilizes the every width distortion in University Of Ningbo's stereo-picture storehouse that the inventive method obtains;

Fig. 6 is the picture quality objective evaluation predicted value of stereo-picture and the scatter diagram of average subjective scoring difference that utilizes the every width distortion in the LIVE stereo-picture storehouse that the inventive method obtains.

Embodiment

Below in conjunction with accompanying drawing, embodiment is described in further detail the present invention.

A kind of objective evaluation method for quality of stereo images based on three-dimensional gradient amplitude that the present invention proposes, it totally realizes block diagram as shown in Figure 1, and it specifically comprises the following steps:

1. make S _orgrepresent original undistorted stereo-picture, make S _disrepresent the stereo-picture of distortion to be evaluated, by S _orgleft visual point image be designated as { L _org(x, y) }, by S _orgright visual point image be designated as { R _org(x, y) }, by S _disleft visual point image be designated as { L _dis(x, y) }, by S _disright visual point image be designated as { R _dis(x, y) }, wherein, (x, y) represent the coordinate position of the pixel in left visual point image and right visual point image, 1≤x≤W, 1≤y≤H, W represents the width of left visual point image and right visual point image, and H represents the height of left visual point image and right visual point image, L _org(x, y) represents { L _org(x, y) } in the pixel value of the coordinate position pixel that is (x, y), R _org(x, y) represents { R _org(x, y) } in the pixel value of the coordinate position pixel that is (x, y), L _dis(x, y) represents { L _dis(x, y) } in the pixel value of the coordinate position pixel that is (x, y), R _dis(x, y) represents { R _dis(x, y) } in the pixel value of the coordinate position pixel that is (x, y).

2. according to { L _org(x, y) } in each pixel and { R _org(x, y) } in the pixel of the respective coordinates position disparity space value under multiple parallax value, obtain S _orgdisparity space image, be designated as { DSI _org(x, y, d) }, and according to { L _dis(x, y) } in each pixel and { R _dis(x, y) } in the pixel of the respective coordinates position disparity space value under multiple parallax value, obtain S _disdisparity space image, be designated as { DSI _dis(x, y, d) }, wherein, DSI _org(x, y, d) represents { DSI _org(x, y, d) } in coordinate position be the disparity space value of the pixel of (x, y, d), DSI _dis(x, y, d) represents { DSI _dis(x, y, d) } in coordinate position be the disparity space value of the pixel of (x, y, d), 0≤d≤d _max, d _maxrepresent maximum disparity value, get in the present embodiment d _max=31.

In this specific embodiment, step is middle S 2. _orgthe acquisition process of disparity space image be:

2.-a1, by { L _org(x, y) } in current pending pixel be defined as current the first pixel, by { R _org(x, y) } in current pending pixel be defined as current the second pixel.

2.-a2, suppose that current the first pixel is { L _org(x, y) } in coordinate position be (x ₁, y ₁) pixel, suppose that current the second pixel is { R _org(x, y) } in coordinate position be (x ₁, y ₁) pixel, current the first pixel is identical with the coordinate position of current the second pixel, gets parallax value d ₀=0, then calculate current the first pixel and current the second pixel at this parallax value d ₀under disparity space value, be designated as DSI _org(x ₁, y ₁, d ₀), DSI _org(x ₁, y ₁, d ₀)=| L _org(x ₁, y ₁)-R _org(x ₁-d ₀, y ₁) |, wherein, 1≤x ₁≤ W, 1≤y ₁≤ H, 0≤d ₀≤ d _max, d _maxrepresent maximum disparity value, L _org(x ₁, y ₁) expression { L _org(x, y) } in coordinate position be (x ₁, y ₁) the pixel value of pixel, R _org(x ₁-d ₀, y ₁) expression { R _org(x, y) } in coordinate position be (x ₁-d ₀, y ₁) the pixel value of pixel, " || " is the symbol that takes absolute value.

2.-a3, choose d _maxindividual and d ₀different parallax value, is designated as respectively

then calculate respectively current the first pixel and current the second pixel at this d _maxdisparity space value under individual different parallax value, corresponding is designated as respectively ${\mathrm{DSI}}_{\mathrm{org}}(x_1,{\mathrm{<figure-callout\; id="1"\; label="y"\; filenames="HDA0000469762880000012.png,HDA0000469762880000021.png"\; state="\{\{state\}\}">y</figure-callout>}}_1,d_1),{\mathrm{DSI}}_{\mathrm{org}}(x_1,{\mathrm{<figure-callout\; id="1"\; label="y"\; filenames="HDA0000469762880000012.png,HDA0000469762880000021.png"\; state="\{\{state\}\}">y</figure-callout>}}_1,d_2),...,{\mathrm{DSI}}_{\mathrm{org}}(x_1,{\mathrm{<figure-callout\; id="1"\; label="y"\; filenames="HDA0000469762880000012.png,HDA0000469762880000021.png"\; state="\{\{state\}\}">y</figure-callout>}}_1,d_i),...,{\mathrm{DSI}}_{\mathrm{org}}(x_1,y_1,d_{d_{\max }}),$ DSI _org(x ₁,y ₁,d ₁)=|L _org(x ₁,y ₁)-R _org(x ₁-d ₁,y ₁)|，DSI _org(x ₁,y ₁,d ₂)=|L _org(x ₁,y ₁)-R _org(x ₁-d ₂,y ₁)|，DSI _org(x ₁,y ₁,d _i)=|L _org(x ₁,y ₁)-R _org(x ₁-d _i,y ₁)|， ${\mathrm{DSI}}_{\mathrm{org}}(x_1,y_1,d_{d_{\max }})=|L_{\mathrm{org}}(x_1,y_1)-R_{\mathrm{org}}(x_1-d_{d_{\max }},y_1)|,$ Wherein, 1≤i≤d _max, d _i=d ₀+ i,

dSI _org(x ₁, y ₁, d ₁) represent that current the first pixel and current the second pixel are at parallax value d ₁under disparity space value, DSI _org(x ₁, y ₁, d ₂) represent that current the first pixel and current the second pixel are at parallax value d ₂under disparity space value, DSI _org(x ₁, y ₁, d _i) represent that current the first pixel and current the second pixel are at parallax value d _iunder disparity space value,

represent that current the first pixel and current the second pixel are in parallax value

under disparity space value, R _org(x ₁-d ₁, y ₁) expression { R _org(x, y) } in coordinate position be (x ₁-d ₁, y ₁) the pixel value of pixel, R _org(x ₁-d ₂, y ₁) expression { R _org(x, y) } in coordinate position be (x ₁-d ₂, y ₁) the pixel value of pixel, R _org(x ₁-d _i, y ₁) expression { R _org(x, y) } in coordinate position be (x ₁-d _i, y ₁) the pixel value of pixel,

represent { R _org(x, y) } in coordinate position be

the pixel value of pixel.

2.-a4, by { L _org(x, y) } in next pending pixel as current the first pixel, by { R _org(x, y) } in next pending pixel as current the second pixel, then return step 2.-a2 continues execution, until { L _org(x, y) } and { R _org(x, y) } in all pixels be disposed, obtain S _orgdisparity space image, be designated as { DSI _org(x, y, d) }, wherein, DSI _org(x, y, d) represents { DSI _org(x, y, d) } in coordinate position be the disparity space value of the pixel of (x, y, d), DSI _orgthe value of (x, y, d) is { L _org(x, y) } in coordinate position be (x, y) pixel and { R _org(x, y) } in the coordinate position disparity space value of pixel under parallax value d that be (x, y),

In this specific embodiment, step is middle S 2. _disthe acquisition process of disparity space image be:

2.-b1, by { L _dis(x, y) } in current pending pixel be defined as current the first pixel, by { R _dis(x, y) } in current pending pixel be defined as current the second pixel.

2.-b2, suppose that current the first pixel is { L _dis(x, y) } in coordinate position be (x ₁, y ₁) pixel, suppose that current the second pixel is { R _dis(x, y) } in coordinate position be (x ₁, y ₁) pixel, current the first pixel is identical with the coordinate position of current the second pixel, gets parallax value d ₀=0, then calculate current the first pixel and current the second pixel at this parallax value d ₀under disparity space value, be designated as DSI _dis(x ₁, y ₁, d ₀), DSI _dis(x ₁, y ₁, d ₀)=| L _dis(x ₁, y ₁)-R _dis(x ₁-d ₀, y ₁) |, wherein, 1≤x ₁≤ W, 1≤y ₁≤ H, 0≤d ₀≤ d _max, d _maxrepresent maximum disparity value, L _dis(x ₁, y ₁) expression { L _dis(x, y) } in coordinate position be (x ₁, y ₁) the pixel value of pixel, R _dis(x ₁-d ₀, y ₁) expression { R _dis(x, y) } in coordinate position be (x ₁-d ₀, y ₁) the pixel value of pixel, " || " is the symbol that takes absolute value.

2.-b3, choose d _maxindividual and d ₀different parallax value, is designated as respectively

then calculate respectively current the first pixel and current the second pixel at this d _maxdisparity space value under individual different parallax value, corresponding is designated as respectively ${\mathrm{DSI}}_{\mathrm{dis}}(x_1,{\mathrm{<figure-callout\; id="1"\; label="y"\; filenames="HDA0000469762880000012.png,HDA0000469762880000021.png"\; state="\{\{state\}\}">y</figure-callout>}}_1,d_1),{\mathrm{DSI}}_{\mathrm{dis}}(x_1,{\mathrm{<figure-callout\; id="1"\; label="y"\; filenames="HDA0000469762880000012.png,HDA0000469762880000021.png"\; state="\{\{state\}\}">y</figure-callout>}}_1,d_2),...,{\mathrm{DSI}}_{\mathrm{dis}}(x_1,{\mathrm{<figure-callout\; id="1"\; label="y"\; filenames="HDA0000469762880000012.png,HDA0000469762880000021.png"\; state="\{\{state\}\}">y</figure-callout>}}_1,d_i),...,{\mathrm{DSI}}_{\mathrm{dis}}(x_1,y_1,d_{d_{\max }}),$ DSI _dis(x ₁,y ₁,d ₁)=|L _dis(x ₁,y ₁)-R _dis(x ₁-d ₁,y ₁)|，DSI _dis(x ₁,y ₁,d ₂)=|L _dis(x ₁,y ₁)-R _dis(x ₁-d ₂,y ₁)|，DSI _dis(x ₁,y ₁,d _i)=|L _dis(x ₁,y ₁)-R _dis(x ₁-d _i,y ₁)|， ${\mathrm{DSI}}_{\mathrm{dis}}(x_1,y_1,d_{d_{\max }})=|L_{\mathrm{dis}}(x_1,y_1)-R_{\mathrm{dis}}(x_1-d_{d_{\max }},y_1)|,$ Wherein, 1≤i≤d _max, d _i=d ₀+ i,

dSI _dis(x ₁, y ₁, d ₁) represent that current the first pixel and current the second pixel are at parallax value d ₁under disparity space value, DSI _dis(x ₁, y ₁, d ₂) represent that current the first pixel and current the second pixel are at parallax value d ₂under disparity space value, DSI _dis(x ₁, y ₁, d _i) represent that current the first pixel and current the second pixel are at parallax value d _iunder disparity space value,

represent that current the first pixel and current the second pixel are in parallax value

under disparity space value, R _dis(x ₁-d ₁, y ₁) expression { R _dis(x, y) } in coordinate position be (x ₁-d ₁, y ₁) the pixel value of pixel, R _dis(x ₁-d ₂, y ₁) expression { R _dis(x, y) } in coordinate position be (x ₁-d ₂, y ₁) the pixel value of pixel, R _dis(x ₁-d _i, y ₁) expression { R _dis(x, y) } in coordinate position be (x ₁-d _i, y ₁) the pixel value of pixel, represent { R _dis(x, y) } in coordinate position be

the pixel value of pixel.

2.-b4, by { L _dis(x, y) } in next pending pixel as current the first pixel, by { R _dis(x, y) } in next pending pixel as current the second pixel, then return step 2.-b2 continues execution, until { L _dis(x, y) } and { R _dis(x, y) } in all pixels be disposed, obtain S _disdisparity space image, be designated as { DSI _dis(x, y, d) }, wherein, DSI _dis(x, y, d) represents { DSI _dis(x, y, d) } in coordinate position be the disparity space value of the pixel of (x, y, d), DSI _disthe value of (x, y, d) is { L _dis(x, y) } in coordinate position be (x, y) pixel and { R _dis(x, y) } in the coordinate position disparity space value of pixel under parallax value d that be (x, y),

3. calculate { DSI _org(x, y, d) } in horizontal direction gradient, vertical gradient and the viewpoint direction gradient of each pixel, by { DSI _org(x, y, d) } in coordinate position be that the horizontal direction gradient of the pixel of (x, y, d) is designated as gx _org(x, y, d), by { DSI _org(x, y, d) } in coordinate position be that the vertical gradient of the pixel of (x, y, d) is designated as gy _org(x, y, d), by { DSI _org(x, y, d) } in coordinate position be that the viewpoint direction gradient of the pixel of (x, y, d) is designated as gd _org(x, y, d).

In this specific embodiment, step 3. in { DSI _org(x, y, d) } in the acquisition process of horizontal direction gradient, vertical gradient and viewpoint direction gradient of each pixel be:

3.-a1, employing horizontal gradient operator are as shown in Figure 2 to { DSI _org(x, y, d) } carry out convolution, obtain { DSI _org(x, y, d) } in the horizontal direction gradient of each pixel, by { DSI _org(x, y, d) } in coordinate position be that the horizontal direction gradient of the pixel of (x, y, d) is designated as gx _org(x, y, d), ${\mathrm{gx}}_{\mathrm{org}}(x,y,d)=\underset{j=d-2}{\overset{j=d+2}{\mathrm{\&Sigma;}}}(-\underset{u=x-2}{\overset{u=x-1}{\mathrm{\&Sigma;}}}\underset{v=y-2}{\overset{v=y+2}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{org}}(u,v,j)+\underset{u=x+1}{\overset{u=x+2}{\mathrm{\&Sigma;}}}\underset{v=y-2}{\overset{v=y+2}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{org}}(u,v,j)),$ Wherein, DSI _org(u, v, j) represents { DSI _org(x, y, d) } in coordinate position be the disparity space value of the pixel of (u, v, j).

3.-a2, employing VG (vertical gradient) operator are as shown in Figure 3 to { DSI _org(x, y, d) } carry out convolution, obtain { DSI _org(x, y, d) } in the vertical gradient of each pixel, by { DSI _org(x, y, d) } in coordinate position be that the vertical gradient of the pixel of (x, y, d) is designated as gy _org(x, y, d), ${\mathrm{gy}}_{\mathrm{org}}(x,y,d)=\underset{j=d-2}{\overset{j=d+2}{\mathrm{\&Sigma;}}}(-\underset{u=x-2}{\overset{u=x+2}{\mathrm{\&Sigma;}}}\underset{v=y-2}{\overset{v=y-1}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{org}}(u,v,j)+\underset{u=x-2}{\overset{u=x+2}{\mathrm{\&Sigma;}}}\underset{v=y+1}{\overset{v=y+2}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{org}}(u,v,j)).$

3.-a3, employing viewpoint gradient operator are as shown in Figure 4 to { DSI _org(x, y, d) } carry out convolution, obtain { DSI _org(x, y, d) } in the viewpoint direction gradient of each pixel, by { DSI _org(x, y, d) } in coordinate position be that the viewpoint direction gradient of the pixel of (x, y, d) is designated as gd _org(x, y, d), ${\mathrm{gd}}_{\mathrm{org}}(x,y,d)=\underset{j=d-2}{\overset{j=d+2}{\mathrm{\&Sigma;}}}(\mathrm{sign}(j-d)\&times;\underset{u=x-2}{\overset{u=x+2}{\mathrm{\&Sigma;}}}\underset{v=y-2}{\overset{v=y+2}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{org}}(u,v,j)),$ Wherein, sign () is step function,

Above-mentioned steps 3.-a1 to step 3.-a3 in, if u<1, DSI _orgthe value of (u, v, j) is by DSI _orgthe value of (1, v, j) substitutes, if u>W, DSI _orgthe value of (u, v, j) is by DSI _orgthe value of (W, v, j) substitutes, if v<1, DSI _orgthe value of (u, v, j) is by DSI _org(u, 1, value j) substitutes, if v>H, DSI _orgthe value of (u, v, j) is by DSI _orgthe value of (u, H, j) substitutes, if j<0, DSI _orgthe value of (u, v, j) is by DSI _orgthe value of (u, v, 0) substitutes, if j>d _max, DSI _orgthe value of (u, v, j) is by DSI _org(u, v, d _max) value substitute, DSI _org(1, v, j) represents { DSI _org(x, y, d) } in coordinate position be the disparity space value of the pixel of (1, v, j), DSI _org(W, v, j) represents { DSI _org(x, y, d) } in coordinate position be the disparity space value of the pixel of (W, v, j), DSI _org(u, 1, j) represent { DSI _org(x, y, d) } in coordinate position be (u, 1, the disparity space value of pixel j), DSI _org(u, H, j) represents { DSI _org(x, y, d) } in coordinate position be the disparity space value of the pixel of (u, H, j), DSI _org(u, v, 0) represents { DSI _org(x, y, d) } in coordinate position be the disparity space value of the pixel of (u, v, 0), DSI _org(u, v, d _max) expression { DSI _org(x, y, d) } in coordinate position be (u, v, d _max) the disparity space value of pixel.

Equally, calculate { DSI _dis(x, y, d) } in horizontal direction gradient, vertical gradient and the viewpoint direction gradient of each pixel, by { DSI _dis(x, y, d) } in coordinate position be that the horizontal direction gradient of the pixel of (x, y, d) is designated as gx _dis(x, y, d), by { DSI _dis(x, y, d) } in coordinate position be that the vertical gradient of the pixel of (x, y, d) is designated as gy _dis(x, y, d), by { DSI _dis(x, y, d) } in coordinate position be that the viewpoint direction gradient of the pixel of (x, y, d) is designated as gd _dis(x, y, d).

In this specific embodiment, step 3. in { DSI _dis(x, y, d) } in the acquisition process of horizontal direction gradient, vertical gradient and viewpoint direction gradient of each pixel be:

3.-b1, employing horizontal gradient operator are as shown in Figure 2 to { DSI _dis(x, y, d) } carry out convolution, obtain { DSI _dis(x, y, d) } in the horizontal direction gradient of each pixel, by { DSI _dis(x, y, d) } in coordinate position be that the horizontal direction gradient of the pixel of (x, y, d) is designated as gx _dis(x, y, d), ${\mathrm{gx}}_{\mathrm{dis}}(x,y,d)=\underset{j=d-2}{\overset{j=d+2}{\mathrm{\&Sigma;}}}(-\underset{u=x-2}{\overset{u=x-1}{\mathrm{\&Sigma;}}}\underset{v=y-2}{\overset{v=y+2}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{dis}}(u,v,j)+\underset{u=x+1}{\overset{u=x+2}{\mathrm{\&Sigma;}}}\underset{v=y-2}{\overset{v=y+2}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{dis}}(u,v,j)),$ Wherein, DSI _dis(u, v, j) represents { DSI _dis(x, y, d) } in coordinate position be the disparity space value of the pixel of (u, v, j).

3.-b2, employing VG (vertical gradient) operator are as shown in Figure 3 to { DSI _dis(x, y, d) } carry out convolution, obtain { DSI _dis(x, y, d) } in the vertical gradient of each pixel, by { DSI _dis(x, y, d) } in coordinate position be that the vertical gradient of the pixel of (x, y, d) is designated as gy _dis(x, y, d), ${\mathrm{gy}}_{\mathrm{dis}}(x,y,d)=\underset{j=d-2}{\overset{j=d+2}{\mathrm{\&Sigma;}}}(-\underset{u=x-2}{\overset{u=x+2}{\mathrm{\&Sigma;}}}\underset{v=y-2}{\overset{v=y-1}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{dis}}(u,v,j)+\underset{u=x-2}{\overset{u=x+2}{\mathrm{\&Sigma;}}}\underset{v=y+1}{\overset{v=y+2}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{dis}}(u,v,j)).$

3.-b3, employing viewpoint gradient operator are as shown in Figure 4 to { DSI _dis(x, y, d) } carry out convolution, obtain { DSI _dis(x, y, d) } in the viewpoint direction gradient of each pixel, by { DSI _dis(x, y, d) } in coordinate position be that the parallax directions gradient of the pixel of (x, y, d) is designated as gd _dis(x, y, d), ${\mathrm{gd}}_{\mathrm{dis}}(x,y,d)=\underset{j=d-2}{\overset{j=d+2}{\mathrm{\&Sigma;}}}(\mathrm{sign}(j-d)\&times;\underset{u=x-2}{\overset{u=x+2}{\mathrm{\&Sigma;}}}\underset{v=y-2}{\overset{v=y+2}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{dis}}(u,v,j)),$ Wherein, sign () is step function,

Above-mentioned steps 3.-b1 to step 3.-b3 in, if u<1, DSI _disthe value of (u, v, j) is by DSI _disthe value of (1, v, j) substitutes, if u>W, DSI _disthe value of (u, v, j) is by DSI _disthe value of (W, v, j) substitutes, if v<1, DSI _disthe value of (u, v, j) is by DSI _dis(u, 1, value j) substitutes, if v>H, DSI _disthe value of (u, v, j) is by DSI _disthe value of (u, H, j) substitutes, if j<0, DSI _disthe value of (u, v, j) is by DSI _disthe value of (u, v, 0) substitutes, if j>d _max, DSI _disthe value of (u, v, j) is by DSI _dis(u, v, d _max) value substitute, DSI _dis(1, v, j) represents { DSI _dis(x, y, d) } in coordinate position be the disparity space value of the pixel of (1, v, j), DSI _dis(W, v, j) represents { DSI _dis(x, y, d) } in coordinate position be the disparity space value of the pixel of (W, v, j), DSI _dis(u, 1, j) represent { DSI _dis(x, y, d) } in coordinate position be (u, 1, the disparity space value of pixel j), DSI _dis(u, H, j) represents { DSI _dis(x, y, d) } in coordinate position be the disparity space value of the pixel of (u, H, j), DSI _dis(u, v, 0) represents { DSI _dis(x, y, d) } in coordinate position be the disparity space value of the pixel of (u, v, 0), DSI _dis(u, v, d _max) expression { DSI _dis(x, y, d) } in coordinate position be (u, v, d _max) the disparity space value of pixel.

4. according to { DSI _org(x, y, d) } in horizontal direction gradient, vertical gradient and the viewpoint direction gradient of each pixel, calculate { DSI _org(x, y, d) } in the three-dimensional gradient amplitude of each pixel, by { DSI _org(x, y, d) } in coordinate position be that the three-dimensional gradient amplitude of the pixel of (x, y, d) is designated as m _org(x, y, d), $m_{\mathrm{org}}(x,y,d)=\sqrt{{\left({\mathrm{gx}}_{\mathrm{org}}(x,y,d)\right)}^2+{\left({\mathrm{gy}}_{\mathrm{org}}(x,y,d)\right)}^2+{\left({\mathrm{gd}}_{\mathrm{org}}(x,y,d)\right)}^2}.$

Equally, according to { DSI _dis(x, y, d) } in horizontal direction gradient, vertical gradient and the viewpoint direction gradient of each pixel, calculate { DSI _dis(x, y, d) } in the three-dimensional gradient amplitude of each pixel, by { DSI _dis(x, y, d) } in coordinate position be that the three-dimensional gradient amplitude of the pixel of (x, y, d) is designated as m _dis(x, y, d), $m_{\mathrm{dis}}(x,y,d)=\sqrt{{\left({\mathrm{gx}}_{\mathrm{dis}}(x,y,d)\right)}^2+{\left({\mathrm{gy}}_{\mathrm{dis}}(x,y,d)\right)}^2+{\left({\mathrm{gd}}_{\mathrm{dis}}(x,y,d)\right)}^2}.$

5. according to { DSI _org(x, y, d) } and { DSI _dis(x, y, d) } in the three-dimensional gradient amplitude of each pixel, calculate { DSI _dis(x, y, d) } in the objective evaluation metric of each pixel, by { DSI _dis(x, y, d) } in coordinate position be that the objective evaluation metric of the pixel of (x, y, d) is designated as Q _dSI(x, y, d), $Q_{\mathrm{DSI}}(x,y,d)=\frac{2\&times;m_{\mathrm{org}}(x,y,d)\&times;m_{\mathrm{dis}}(x,y,d)+C}{{\left(m_{\mathrm{org}}(x,y,d)\right)}^2+{\left(m_{\mathrm{dis}}(x,y,d)\right)}^2+C},$ Wherein, C, for controlling parameter, gets C=0.85 in the present embodiment.

6. according to { DSI _dis(x, y, d) } in the objective evaluation metric of each pixel, calculate S _dispicture quality objective evaluation predicted value, be designated as Q,

wherein, Ω represents { DSI _dis(x, y, d) } in the set of coordinate position of all pixels, N represents { DSI _dis(x, y, d) } in total number of the pixel that comprises.

At this, adopt University Of Ningbo's stereo-picture storehouse and LIVE stereo-picture storehouse to analyze the picture quality objective evaluation predicted value and the average correlativity between subjective scoring difference of the stereo-picture of the distortion that the present embodiment obtains.The stereo-picture of 72 width distortions in the stereo-picture of 60 width distortions in stereo-picture, the white Gaussian noise situation of 60 width distortions in stereo-picture, the Gaussian Blur situation of 60 width distortions in stereo-picture, the JPEG2000 compression situation of the 60 width distortions in the JPEG of different distortion levels compression situation by 12 undistorted stereo-pictures of University Of Ningbo's stereo-picture storehouse and H.264 coding distortion situation forms.The stereo-picture of 80 width distortions in stereo-picture and the Fast Fading distortion situation of 80 width distortions in stereo-picture, the white Gaussian noise situation of 45 width distortions in stereo-picture, the Gaussian Blur situation of 80 width distortions in stereo-picture, the JPEG2000 compression situation of the 80 width distortions in the JPEG of different distortion levels compression situation by 20 undistorted stereo-pictures of LIVE stereo-picture storehouse forms.

Here, utilize 4 conventional objective parameters of evaluate image quality evaluating method as evaluation index, be Pearson correlation coefficient (the Pearson linear correlation coefficient under non-linear regression condition, PLCC), Spearman related coefficient (Spearman rank order correlation coefficient, SROCC), Kendall related coefficient (Kendall rank-order correlation coefficient, KROCC), square error (root mean squared error, RMSE), the accuracy of the three-dimensional image objective evaluation result of PLCC and RMSE reflection distortion, SROCC and KROCC reflect its monotonicity.

Utilize the inventive method to calculate the picture quality objective evaluation predicted value of the stereo-picture of the every width distortion in picture quality objective evaluation predicted value and the LIVE stereo-picture storehouse of stereo-picture of the every width distortion in University Of Ningbo's stereo-picture storehouse, recycle the average subjective scoring difference of the stereo-picture of the every width distortion in average subjective scoring difference and the LIVE stereo-picture storehouse of stereo-picture that existing subjective evaluation method obtains the every width distortion in University Of Ningbo's stereo-picture storehouse.The picture quality objective evaluation predicted value of the stereo-picture of the distortion calculating by the inventive method is done to five parameter L ogistic function nonlinear fittings, PLCC, SROCC and KROCC value are higher, and the lower explanation method for objectively evaluating of RMSE value is better with average subjective scoring difference correlativity.Table 1, table 2, table 3 and table 4 have provided the picture quality objective evaluation predicted value of the stereo-picture that adopts the distortion that the inventive method obtains and average Pearson correlation coefficient, Spearman related coefficient, Kendall related coefficient and the square error between subjective scoring difference.From table 1, table 2, table 3 and table 4, can find out, correlativity between the final picture quality objective evaluation predicted value of the stereo-picture of the distortion that employing the inventive method obtains 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.

Fig. 5 has provided the scatter diagram of picture quality objective evaluation predicted value with the average subjective scoring difference of the stereo-picture that utilizes the every width distortion in University Of Ningbo's stereo-picture storehouse that the inventive method obtains, Fig. 6 has provided the scatter diagram of picture quality objective evaluation predicted value with the average subjective scoring difference of the stereo-picture that utilizes the every width distortion in the LIVE stereo-picture storehouse that the inventive method obtains, loose point is more concentrated, illustrates that the consistance of objective evaluation result and subjective perception is better.As can be known from Fig. 5 and Fig. 6, adopt the scatter diagram that obtains of the inventive method more concentrated, and the goodness of fit between subjective assessment data is higher.

Table 1 utilizes the Pearson correlation coefficient comparison between picture quality objective evaluation predicted value and the average subjective scoring difference of stereo-picture of the distortion that the inventive method obtains

Table 2 utilizes the Spearman related coefficient comparison between picture quality objective evaluation predicted value and the average subjective scoring difference of stereo-picture of the distortion that the inventive method obtains

Table 3 utilizes the Kendall related coefficient comparison between picture quality objective evaluation predicted value and the average subjective scoring difference of stereo-picture of the distortion that the inventive method obtains

Table 4 utilizes the square error comparison between picture quality objective evaluation predicted value and the average subjective scoring difference of stereo-picture of the distortion that the inventive method obtains

Claims (3)

1. the objective evaluation method for quality of stereo images based on three-dimensional gradient amplitude, is characterized in that comprising the following steps:

1. make S _orgrepresent original undistorted stereo-picture, make S _disrepresent the stereo-picture of distortion to be evaluated, by S _orgleft visual point image be designated as { L _org(x, y) }, by S _orgright visual point image be designated as { R _org(x, y) }, by S _disleft visual point image be designated as { L _dis(x, y) }, by S _disright visual point image be designated as { R _dis(x, y) }, wherein, (x, y) represent the coordinate position of the pixel in left visual point image and right visual point image, 1≤x≤W, 1≤y≤H, W represents the width of left visual point image and right visual point image, and H represents the height of left visual point image and right visual point image, L _org(x, y) represents { L _org(x, y) } in the pixel value of the coordinate position pixel that is (x, y), R _org(x, y) represents { R _org(x, y) } in the pixel value of the coordinate position pixel that is (x, y), L _dis(x, y) represents { L _dis(x, y) } in the pixel value of the coordinate position pixel that is (x, y), R _dis(x, y) represents { R _dis(x, y) } in the pixel value of the coordinate position pixel that is (x, y);

2. according to { L _org(x, y) } in each pixel and { R _org(x, y) } in the pixel of the respective coordinates position disparity space value under multiple parallax value, obtain S _orgdisparity space image, be designated as { DSI _org(x, y, d) }, and according to { L _dis(x, y) } in each pixel and { R _dis(x, y) } in the pixel of the respective coordinates position disparity space value under multiple parallax value, obtain S _disdisparity space image, be designated as { DSI _dis(x, y, d) }, wherein, DSI _org(x, y, d) represents { DSI _org(x, y, d) } in coordinate position be the disparity space value of the pixel of (x, y, d), DSI _dis(x, y, d) represents { DSI _dis(x, y, d) } in coordinate position be the disparity space value of the pixel of (x, y, d), 0≤d≤d _max, d _maxrepresent maximum disparity value;

3. calculate { DSI _org(x, y, d) } in horizontal direction gradient, vertical gradient and the viewpoint direction gradient of each pixel, by { DSI _org(x, y, d) } in coordinate position be that the horizontal direction gradient of the pixel of (x, y, d) is designated as gx _org(x, y, d), by { DSI _org(x, y, d) } in coordinate position be that the vertical gradient of the pixel of (x, y, d) is designated as gy _org(x, y, d), by { DSI _org(x, y, d) } in coordinate position be that the viewpoint direction gradient of the pixel of (x, y, d) is designated as gd _org(x, y, d);

Equally, calculate { DSI _dis(x, y, d) } in horizontal direction gradient, vertical gradient and the viewpoint direction gradient of each pixel, by { DSI _dis(x, y, d) } in coordinate position be that the horizontal direction gradient of the pixel of (x, y, d) is designated as gx _dis(x, y, d), by { DSI _dis(x, y, d) } in coordinate position be that the vertical gradient of the pixel of (x, y, d) is designated as gy _dis(x, y, d), by { DSI _dis(x, y, d) } in coordinate position be that the viewpoint direction gradient of the pixel of (x, y, d) is designated as gd _dis(x, y, d);

4. according to { DSI _org(x, y, d) } in horizontal direction gradient, vertical gradient and the viewpoint direction gradient of each pixel, calculate { DSI _org(x, y, d) } in the three-dimensional gradient amplitude of each pixel, by { DSI _org(x, y, d) } in coordinate position be that the three-dimensional gradient amplitude of the pixel of (x, y, d) is designated as m _org(x, y, d), $m_{\mathrm{org}}(x,y,d)=\sqrt{{\left({\mathrm{gx}}_{\mathrm{org}}(x,y,d)\right)}^2+{\left({\mathrm{gy}}_{\mathrm{org}}(x,y,d)\right)}^2+{\left({\mathrm{gd}}_{\mathrm{org}}(x,y,d)\right)}^2};$

Equally, according to { DSI _dis(x, y, d) } in horizontal direction gradient, vertical gradient and the viewpoint direction gradient of each pixel, calculate { DSI _dis(x, y, d) } in the three-dimensional gradient amplitude of each pixel, by { DSI _dis(x, y, d) } in coordinate position be that the three-dimensional gradient amplitude of the pixel of (x, y, d) is designated as m _dis(x, y, d), $m_{\mathrm{dis}}(x,y,d)=\sqrt{{\left({\mathrm{gx}}_{\mathrm{dis}}(x,y,d)\right)}^2+{\left({\mathrm{gy}}_{\mathrm{dis}}(x,y,d)\right)}^2+{\left({\mathrm{gd}}_{\mathrm{dis}}(x,y,d)\right)}^2};$

5. according to { DSI _org(x, y, d) } and { DSI _dis(x, y, d) } in the three-dimensional gradient amplitude of each pixel, calculate { DSI _dis(x, y, d) } in the objective evaluation metric of each pixel, by { DSI _dis(x, y, d) } in coordinate position be that the objective evaluation metric of the pixel of (x, y, d) is designated as Q _dSI(x, y, d), $Q_{\mathrm{DSI}}(x,y,d)=\frac{2\&times;m_{\mathrm{org}}(x,y,d)\&times;m_{\mathrm{dis}}(x,y,d)+C}{{\left(m_{\mathrm{org}}(x,y,d)\right)}^2+{\left(m_{\mathrm{dis}}(x,y,d)\right)}^2+C},$ Wherein, C is for controlling parameter;

6. according to { DSI _dis(x, y, d) } in the objective evaluation metric of each pixel, calculate S _dispicture quality objective evaluation predicted value, be designated as Q,

wherein, Ω represents { DSI _dis(x, y, d) } in the set of coordinate position of all pixels, N represents { DSI _dis(x, y, d) } in total number of the pixel that comprises.

2. a kind of objective evaluation method for quality of stereo images based on three-dimensional gradient amplitude according to claim 1, is characterized in that 2. middle S of described step _orgthe acquisition process of disparity space image be:

2.-a1, by { L _org(x, y) } in current pending pixel be defined as current the first pixel, by { R _org(x, y) } in current pending pixel be defined as current the second pixel;

2.-a2, suppose that current the first pixel is { L _org(x, y) } in coordinate position be (x ₁, y ₁) pixel, suppose that current the second pixel is { R _org(x, y) } in coordinate position be (x ₁, y ₁) pixel, get parallax value d ₀=0, then calculate current the first pixel and current the second pixel at this parallax value d ₀under disparity space value, be designated as DSI _org(x ₁, y ₁, d ₀), DSI _org(x ₁, y ₁, d ₀)=| L _org(x ₁, y ₁)-R _org(x ₁-d ₀, y ₁) |, wherein, 1≤x ₁≤ W, 1≤y ₁≤ H, 0≤d ₀≤ d _max, d _maxrepresent maximum disparity value, L _org(x ₁, y ₁) expression { L _org(x, y) } in coordinate position be (x ₁, y ₁) the pixel value of pixel, R _org(x ₁-d ₀, y ₁) expression { R _org(x, y) } in coordinate position be (x ₁-d ₀, y ₁) the pixel value of pixel, " || " is the symbol that takes absolute value;

2.-a3, choose d _maxindividual and d ₀different parallax value, is designated as respectively

then calculate respectively current the first pixel and current the second pixel at this d _maxdisparity space value under individual different parallax value, corresponding is designated as respectively ${\mathrm{DSI}}_{\mathrm{org}}(x_1,y_1,d_1),{\mathrm{DSI}}_{\mathrm{org}}(x_1,y_1,d_2),...,{\mathrm{DSI}}_{\mathrm{org}}(x_1,y_1,d_i),...,{\mathrm{DSI}}_{\mathrm{org}}(x_1,y_1,d_{d_{\max }}),$ DSI _org(x ₁,y ₁,d ₁)=|L _org(x ₁,y ₁)-R _org(x ₁-d ₁,y ₁)|，DSI _org(x ₁,y ₁,d ₂)=|L _org(x ₁,y ₁)-R _org(x ₁-d ₂,y ₁)|，DSI _org(x ₁,y ₁,d _i)=|L _org(x ₁,y ₁)-R _org(x ₁-d _i,y ₁)|， ${\mathrm{DSI}}_{\mathrm{org}}(x_1,y_1,d_{d_{\max }})=|L_{\mathrm{org}}(x_1,y_1)-R_{\mathrm{org}}(x_1-d_{d_{\max }},y_1)|,$ Wherein, 1≤i≤d _max, d _i=d ₀+ i,

dSI _org(x ₁, y ₁, d ₁) represent that current the first pixel and current the second pixel are at parallax value d ₁under disparity space value, DSI _org(x ₁, y ₁, d ₂) represent that current the first pixel and current the second pixel are at parallax value d ₂under disparity space value, DSI _org(x ₁, y ₁, d _i) represent that current the first pixel and current the second pixel are at parallax value d _iunder disparity space value,

represent that current the first pixel and current the second pixel are in parallax value

under disparity space value, R _org(x ₁-d ₁, y ₁) expression { R _org(x, y) } in coordinate position be (x ₁-d ₁, y ₁) the pixel value of pixel, R _org(x ₁-d ₂, y ₁) expression { R _org(x, y) } in coordinate position be (x ₁-d ₂, y ₁) the pixel value of pixel, R _org(x ₁-d _i, y ₁) expression { R _org(x, y) } in coordinate position be (x ₁-d _i, y ₁) the pixel value of pixel,

represent { R _org(x, y) } in coordinate position be

the pixel value of pixel;

2.-a4, by { L _org(x, y) } in next pending pixel as current the first pixel, by { R _org(x, y) } in next pending pixel as current the second pixel, then return step 2.-a2 continues execution, until { L _org(x, y) } and { R _org(x, y) } in all pixels be disposed, obtain S _orgdisparity space image, be designated as { DSI _org(x, y, d) }, wherein, DSI _org(x, y, d) represents { DSI _org(x, y, d) } in coordinate position be the disparity space value of the pixel of (x, y, d), DSI _orgthe value of (x, y, d) is { L _org(x, y) } in coordinate position be (x, y) pixel and { R _org(x, y) } in the coordinate position disparity space value of pixel under parallax value d that be (x, y),

Described step is middle S 2. _disthe acquisition process of disparity space image be:

2.-b1, by { L _dis(x, y) } in current pending pixel be defined as current the first pixel, by { R _dis(x, y) } in current pending pixel be defined as current the second pixel;

2.-b2, suppose that current the first pixel is { L _dis(x, y) } in coordinate position be (x ₁, y ₁) pixel, suppose that current the second pixel is { R _dis(x, y) } in coordinate position be (x ₁, y ₁) pixel, get parallax value d ₀=0, then calculate current the first pixel and current the second pixel at this parallax value d ₀under disparity space value, be designated as DSI _dis(x ₁, y ₁, d ₀), DSI _dis(x ₁, y ₁, d ₀)=| L _dis(x ₁, y ₁)-R _dis(x ₁-d ₀, y ₁) |, wherein, 1≤x ₁≤ W, 1≤y ₁≤ H, 0≤d ₀≤ d _max, d _maxrepresent maximum disparity value, L _dis(x ₁, y ₁) expression { L _dis(x, y) } in coordinate position be (x ₁, y ₁) the pixel value of pixel, R _dis(x ₁-d ₀, y ₁) expression { R _dis(x, y) } in coordinate position be (x ₁-d ₀, y ₁) the pixel value of pixel, " || " is the symbol that takes absolute value;

2.-b3, choose d _maxindividual and d ₀different parallax value, is designated as respectively

then calculate respectively current the first pixel and current the second pixel at this d _maxdisparity space value under individual different parallax value, corresponding is designated as respectively ${\mathrm{DSI}}_{\mathrm{dis}}(x_1,y_1,d_1),{\mathrm{DSI}}_{\mathrm{dis}}(x_1,y_1,d_2),...,{\mathrm{DSI}}_{\mathrm{dis}}(x_1,y_1,d_i),...,{\mathrm{DSI}}_{\mathrm{dis}}(x_1,y_1,d_{d_{\max }}),$ DSI _dis(x ₁,y ₁,d ₁)=|L _dis(x ₁,y ₁)-R _dis(x ₁-d ₁,y ₁)|，DSI _dis(x ₁,y ₁,d ₂)=|L _dis(x ₁,y ₁)-R _dis(x ₁-d ₂,y ₁)|，DSI _dis(x ₁,y ₁,d _i)=|L _dis(x ₁,y ₁)-R _dis(x ₁-d _i,y ₁)|， ${\mathrm{DSI}}_{\mathrm{dis}}(x_1,y_1,d_{d_{\max }})=|L_{\mathrm{dis}}(x_1,y_1)-R_{\mathrm{dis}}(x_1-d_{d_{\max }},y_1)|,$ Wherein, 1≤i≤d _max, d _i=d ₀+ i,

dSI _dis(x ₁, y ₁, d ₁) represent that current the first pixel and current the second pixel are at parallax value d ₁under disparity space value, DSI _dis(x ₁, y ₁, d ₂) represent that current the first pixel and current the second pixel are at parallax value d ₂under disparity space value, DSI _dis(x ₁, y ₁, d _i) represent that current the first pixel and current the second pixel are at parallax value d _iunder disparity space value, represent that current the first pixel and current the second pixel are in parallax value

under disparity space value, R _dis(x ₁-d ₁, y ₁) expression { R _dis(x, y) } in coordinate position be (x ₁-d ₁, y ₁) the pixel value of pixel, R _dis(x ₁-d ₂, y ₁) expression { R _dis(x, y) } in coordinate position be (x ₁-d ₂, y ₁) the pixel value of pixel, R _dis(x ₁-d _i, y ₁) expression { R _dis(x, y) } in coordinate position be (x ₁-d _i, y ₁) the pixel value of pixel,

represent { R _dis(x, y) } in coordinate position be the pixel value of pixel;

2.-b4, by { L _dis(x, y) } in next pending pixel as current the first pixel, by { R _dis(x, y) } in next pending pixel as current the second pixel, then return step 2.-b2 continues execution, until { L _dis(x, y) } and { R _dis(x, y) } in all pixels be disposed, obtain S _disdisparity space image, be designated as { DSI _dis(x, y, d) }, wherein, DSI _dis(x, y, d) represents { DSI _dis(x, y, d) } in coordinate position be the disparity space value of the pixel of (x, y, d), DSI _disthe value of (x, y, d) is { L _dis(x, y) } in coordinate position be (x, y) pixel and { R _dis(x, y) } in the coordinate position disparity space value of pixel under parallax value d that be (x, y),

3. a kind of objective evaluation method for quality of stereo images based on three-dimensional gradient amplitude according to claim 1 and 2, is characterized in that { DSI during described step 3. _org(x, y, d) } in the acquisition process of horizontal direction gradient, vertical gradient and viewpoint direction gradient of each pixel be:

3.-a1, employing horizontal gradient operator are to { DSI _org(x, y, d) } carry out convolution, obtain { DSI _org(x, y, d) } in the horizontal direction gradient of each pixel, by { DSI _org(x, y, d) } in coordinate position be that the horizontal direction gradient of the pixel of (x, y, d) is designated as gx _org(x, y, d), ${\mathrm{gx}}_{\mathrm{org}}(x,y,d)=\underset{j=d-2}{\overset{j=d+2}{\mathrm{\&Sigma;}}}(-\underset{u=x-2}{\overset{u=x-1}{\mathrm{\&Sigma;}}}\underset{v=y-2}{\overset{v=y+2}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{org}}(u,v,j)+\underset{u=x+1}{\overset{u=x+2}{\mathrm{\&Sigma;}}}\underset{v=y-2}{\overset{v=y+2}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{org}}(u,v,j)),$ Wherein, DSI _org(u, v, j) represents { DSI _org(x, y, d) } in coordinate position be the disparity space value of the pixel of (u, v, j);

3.-a2, employing VG (vertical gradient) operator are to { DSI _org(x, y, d) } carry out convolution, obtain { DSI _org(x, y, d) } in the vertical gradient of each pixel, by { DSI _org(x, y, d) } in coordinate position be that the vertical gradient of the pixel of (x, y, d) is designated as gy _org(x, y, d), ${\mathrm{gy}}_{\mathrm{org}}(x,y,d)=\underset{j=d-2}{\overset{j=d+2}{\mathrm{\&Sigma;}}}(-\underset{u=x-2}{\overset{u=x+2}{\mathrm{\&Sigma;}}}\underset{v=y-2}{\overset{v=y-1}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{org}}(u,v,j)+\underset{u=x-2}{\overset{u=x+2}{\mathrm{\&Sigma;}}}\underset{v=y+1}{\overset{v=y+2}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{org}}(u,v,j));$

3.-a3, employing viewpoint gradient operator are to { DSI _org(x, y, d) } carry out convolution, obtain { DSI _org(x, y, d) } in the viewpoint direction gradient of each pixel, by { DSI _org(x, y, d) } in coordinate position be that the viewpoint direction gradient of the pixel of (x, y, d) is designated as gd _org(x, y, d), ${\mathrm{gd}}_{\mathrm{org}}(x,y,d)=\underset{j=d-2}{\overset{j=d+2}{\mathrm{\&Sigma;}}}(\mathrm{sign}(j-d)\&times;\underset{u=x-2}{\overset{u=x+2}{\mathrm{\&Sigma;}}}\underset{v=y-2}{\overset{v=y+2}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{org}}(u,v,j)),$ Wherein, sign () is step function,

Above-mentioned steps 3.-a1 to step 3.-a3 in, if u<1, DSI _orgthe value of (u, v, j) is by DSI _orgthe value of (1, v, j) substitutes, if u>W, DSI _orgthe value of (u, v, j) is by DSI _orgthe value of (W, v, j) substitutes, if v<1, DSI _orgthe value of (u, v, j) is by DSI _org(u, 1, value j) substitutes, if v>H, DSI _orgthe value of (u, v, j) is by DSI _orgthe value of (u, H, j) substitutes, if j<0, DSI _orgthe value of (u, v, j) is by DSI _orgthe value of (u, v, 0) substitutes, if j>d _max, DSI _orgthe value of (u, v, j) is by DSI _org(u, v, d _max) value substitute, DSI _org(1, v, j) represents { DSI _org(x, y, d) } in coordinate position be the disparity space value of the pixel of (1, v, j), DSI _org(W, v, j) represents { DSI _org(x, y, d) } in coordinate position be the disparity space value of the pixel of (W, v, j), DSI _org(u, 1, j) represent { DSI _org(x, y, d) } in coordinate position be (u, 1, the disparity space value of pixel j), DSI _org(u, H, j) represents { DSI _org(x, y, d) } in coordinate position be the disparity space value of the pixel of (u, H, j), DSI _org(u, v, 0) represents { DSI _org(x, y, d) } in coordinate position be the disparity space value of the pixel of (u, v, 0), DSI _org(u, v, d _max) expression { DSI _org(x, y, d) } in coordinate position be (u, v, d _max) the disparity space value of pixel;

Described step 3. in { DSI _dis(x, y, d) } in the acquisition process of horizontal direction gradient, vertical gradient and viewpoint direction gradient of each pixel be:

3.-b1, employing horizontal gradient operator are to { DSI _dis(x, y, d) } carry out convolution, obtain { DSI _dis(x, y, d) } in the horizontal direction gradient of each pixel, by { DSI _dis(x, y, d) } in coordinate position be that the horizontal direction gradient of the pixel of (x, y, d) is designated as gx _dis(x, y, d), ${\mathrm{gx}}_{\mathrm{dis}}(x,y,d)=\underset{j=d-2}{\overset{j=d+2}{\mathrm{\&Sigma;}}}(-\underset{u=x-2}{\overset{u=x-1}{\mathrm{\&Sigma;}}}\underset{v=y-2}{\overset{v=y+2}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{dis}}(u,v,j)+\underset{u=x+1}{\overset{u=x+2}{\mathrm{\&Sigma;}}}\underset{v=y-2}{\overset{v=y+2}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{dis}}(u,v,j)),$ Wherein, DSI _dis(u, v, j) represents { DSI _dis(x, y, d) } in coordinate position be the disparity space value of the pixel of (u, v, j);

3.-b2, employing VG (vertical gradient) operator are to { DSI _dis(x, y, d) } carry out convolution, obtain { DSI _dis(x, y, d) } in the vertical gradient of each pixel, by { DSI _dis(x, y, d) } in coordinate position be that the vertical gradient of the pixel of (x, y, d) is designated as gy _dis(x, y, d), ${\mathrm{gy}}_{\mathrm{dis}}(x,y,d)=\underset{j=d-2}{\overset{j=d+2}{\mathrm{\&Sigma;}}}(-\underset{u=x-2}{\overset{u=x+2}{\mathrm{\&Sigma;}}}\underset{v=y-2}{\overset{v=y-1}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{dis}}(u,v,j)+\underset{u=x-2}{\overset{u=x+2}{\mathrm{\&Sigma;}}}\underset{v=y+1}{\overset{v=y+2}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{dis}}(u,v,j));$

3.-b3, employing viewpoint gradient operator are to { DSI _dis(x, y, d) } carry out convolution, obtain { DSI _dis(x, y, d) } in the viewpoint direction gradient of each pixel, by { DSI _dis(x, y, d) } in coordinate position be that the parallax directions gradient of the pixel of (x, y, d) is designated as gd _dis(x, y, d), ${\mathrm{gd}}_{\mathrm{dis}}(x,y,d)=\underset{j=d-2}{\overset{j=d+2}{\mathrm{\&Sigma;}}}(\mathrm{sign}(j-d)\&times;\underset{u=x-2}{\overset{u=x+2}{\mathrm{\&Sigma;}}}\underset{v=y-2}{\overset{v=y+2}{\mathrm{\&Sigma;}}}{\mathrm{DSI}}_{\mathrm{dis}}(u,v,j)),$ Wherein, sign () is step function,

Above-mentioned steps 3.-b1 to step 3.-b3 in, if u<1, DSI _disthe value of (u, v, j) is by DSI _disthe value of (1, v, j) substitutes, if u>W, DSI _disthe value of (u, v, j) is by DSI _disthe value of (W, v, j) substitutes, if v<1, DSI _disthe value of (u, v, j) is by DSI _dis(u, 1, value j) substitutes, if v>H, DSI _disthe value of (u, v, j) is by DSI _disthe value of (u, H, j) substitutes, if j<0, DSI _disthe value of (u, v, j) is by DSI _disthe value of (u, v, 0) substitutes, if j>d _max, DSI _disthe value of (u, v, j) is by DSI _dis(u, v, d _max) value substitute, DSI _dis(1, v, j) represents { DSI _dis(x, y, d) } in coordinate position be the disparity space value of the pixel of (1, v, j), DSI _dis(W, v, j) represents { DSI _dis(x, y, d) } in coordinate position be the disparity space value of the pixel of (W, v, j), DSI _dis(u, 1, j) represent { DSI _dis(x, y, d) } in coordinate position be (u, 1, the disparity space value of pixel j), DSI _dis(u, H, j) represents { DSI _dis(x, y, d) } in coordinate position be the disparity space value of the pixel of (u, H, j), DSI _dis(u, v, 0) represents { DSI _dis(x, y, d) } in coordinate position be the disparity space value of the pixel of (u, v, 0), DSI _dis(u, v, d _max) expression { DSI _dis(x, y, d) } in coordinate position be (u, v, d _max) the disparity space value of pixel.

Images