CN110717892B - Tone mapping image quality evaluation method - Google Patents
Tone mapping image quality evaluation method Download PDFInfo
- Publication number
- CN110717892B CN110717892B CN201910881340.2A CN201910881340A CN110717892B CN 110717892 B CN110717892 B CN 110717892B CN 201910881340 A CN201910881340 A CN 201910881340A CN 110717892 B CN110717892 B CN 110717892B
- Authority
- CN
- China
- Prior art keywords
- pixel
- pixel points
- tone mapping
- components
- value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000013507 mapping Methods 0.000 title claims abstract description 68
- 238000000034 method Methods 0.000 title claims abstract description 36
- 238000013441 quality evaluation Methods 0.000 title claims abstract description 12
- 239000013598 vector Substances 0.000 claims abstract description 93
- 238000012549 training Methods 0.000 claims abstract description 60
- 238000012360 testing method Methods 0.000 claims abstract description 44
- 239000011159 matrix material Substances 0.000 claims description 27
- 238000004364 calculation method Methods 0.000 claims description 21
- 230000006870 function Effects 0.000 claims description 13
- 239000000126 substance Substances 0.000 claims description 12
- 241000023320 Luma <angiosperm> Species 0.000 claims description 6
- OSWPMRLSEDHDFF-UHFFFAOYSA-N methyl salicylate Chemical compound COC(=O)C1=CC=CC=C1O OSWPMRLSEDHDFF-UHFFFAOYSA-N 0.000 claims description 6
- 238000012886 linear function Methods 0.000 claims description 5
- 238000010801 machine learning Methods 0.000 claims description 3
- 238000006243 chemical reaction Methods 0.000 claims 1
- 238000011156 evaluation Methods 0.000 abstract description 9
- 230000008447 perception Effects 0.000 abstract description 4
- 239000000284 extract Substances 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000012854 evaluation process Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012805 post-processing Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20172—Image enhancement details
- G06T2207/20208—High dynamic range [HDR] image processing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30168—Image quality inspection
Abstract
The invention discloses a tone mapping image quality evaluation method, which comprises the steps of taking the influence of bright area characteristics and dark area characteristics on tone mapping into consideration in a training stage, extracting bright and dark area characteristic vectors of a tone mapping image, extracting area contrast characteristic vectors of the tone mapping image, forming a global characteristic vector, training the global characteristic vectors of all tone mapping images in a training image set by using support vector regression, and constructing a quality prediction model; in the testing stage, the objective quality predicted value of the tone mapping image is obtained through calculating the global feature vector of the tone mapping image used for testing and predicting according to the quality prediction model constructed in the training stage, and because the obtained global feature vector information has stronger stability and can better reflect the quality change condition of the tone mapping image, the correlation between the objective evaluation result and the subjective perception is effectively improved.
Description
Technical Field
The invention relates to an image quality evaluation method, in particular to a tone mapping image quality evaluation method.
Background
With the rapid development of display technologies, High Dynamic Range (HDR) images have received increasing attention. The high dynamic range image has rich levels, and can achieve the light and shadow effect which is far closer to the reality than the common image. However, conventional display devices can only support low dynamic range display output. To solve the contradiction between the mismatch of the dynamic range of the real scene and the conventional display device, many Tone Mapping (Tone Mapping) algorithms for high dynamic range images are proposed. The aim of the tone mapping algorithm of the high dynamic range image is to compress the brightness of the high dynamic range image to a range acceptable by the traditional display equipment, meanwhile, the detail information of the original image is kept as much as possible, and image flaws are avoided. Therefore, how to accurately and objectively evaluate the performance of different tone mapping methods plays an important role in guiding content production and post-processing.
However, for the tone-mapped image quality evaluation, if the existing image quality evaluation method is directly applied to the tone-mapped image, the tone-mapped image has only a high dynamic range image as a reference, and thus the objective evaluation value cannot be accurately predicted. Therefore, how to effectively extract visual features in the evaluation process to enable objective evaluation results to be more perceptually accordant with the human visual system is a problem to be researched and solved in the process of carrying out objective quality evaluation on tone mapping images.
Disclosure of Invention
The invention aims to provide a tone mapping image quality evaluation method which can effectively improve the correlation between objective evaluation results and subjective perception.
The technical scheme adopted by the invention for solving the technical problems is as follows: a tone mapping image quality evaluation method is characterized by comprising a training stage and a testing stage;
the training stage process comprises the following specific steps:
firstly, 1, selecting N tone mapping images with the width of W and the height of H to form a training image set, and marking the kth tone mapping image in the training image set as a tone mapping image
Wherein, the first and the second end of the pipe are connected with each other,n is a positive integer, N is more than 1, k is a positive integer, the initial value of k is 1, and k is more than or equal to 1 and less than or equal to N;
firstly, 2, carrying out area division on each tone mapping image in the training image set, dividing each tone mapping image into a bright area, a dark area and a normal area, and dividing each tone mapping image into a bright area, a dark area and a normal area
The bright area, the dark area and the normal area are correspondingly recorded as
And
calculating a bright and dark region characteristic vector of each tone mapping image in the training image set according to a bright region and a dark region of each tone mapping image in the training image set, and calculating the characteristic vector of each tone mapping image in the training image set according to the bright region and the dark region
The light and dark region feature vector is recorded as
And calculating the regional contrast characteristic vector of each tone mapping image in the training image set according to the bright region, the dark region and the normal region of each tone mapping image in the training image set, and calculating the contrast characteristic vector of each tone mapping image in the training image set according to the bright region, the dark region and the normal region 
Is noted as
Wherein, the first and the second end of the pipe are connected with each other,
has a dimension of 3 x 1 (a),
dimension (d) is 8 × 1;
(I _ 4) collecting each training imageThe light and dark region feature vectors and the region contrast feature vectors of the tone-mapped image form a global feature vector
Is denoted as Fk，
Wherein, FkHas a dimension of 11X 1, symbol "[ 2 ]]"is a vector representing a symbol and,
show that
And
connected to form a vector;
firstly, 5, forming a training sample data set by global feature vectors and average subjective score difference values of all tone mapping images in the training image set, wherein the training sample data set comprises N global feature vectors and N average subjective score difference values; then, a support vector regression is adopted as a machine learning method to train all global feature vectors in the training sample data set, so that the error between a regression function value obtained through training and the average subjective score difference is minimum, and the optimal weight vector is obtained through fitting
And an optimal bias term
Then using the optimal weight vector
And an optimal bias term
Structural quality prediction model, as
Wherein the content of the first and second substances,
in functional representation, F is used to represent the global feature vector of the tone-mapped image, and as input vector to the quality prediction model, 
Is composed of
The method (2) is implemented by the following steps,
is a linear function of F;
the test stage process comprises the following specific steps:
for any tone mapped image I used as testtestObtaining I according to the same operation from the step I _2 to the step I _4testGlobal feature vector of (2), noted as Ftest(ii) a Then, according to the quality prediction model pair F constructed in the training stagetestTesting and predicting to obtain FtestCorresponding predicted value is taken as ItestThe predicted objective quality value of (2) is marked as Qtest，
Wherein, ItestHas a width W 'and a height H', FtestHas a dimension of 11 x 1,
is represented by FtestIs a linear function of (a).
Said step (i _ 2)
And
the acquisition process comprises the following steps:
(ii) (-) 2 a)
R, G, and B components in the RGB color space are expressed as
Then calculate
Is recorded as a dark channel image
Will be provided with
The pixel value of the pixel point with the middle coordinate position (x, y) is recorded as
Wherein x is more than or equal to 1 and less than or equal to W, y is more than or equal to 1 and less than or equal to H, min () is a function for taking the minimum value, Cx,yRepresenting a set of coordinate positions of all pixel points within a 3 × 3 neighborhood range centered on a pixel point whose coordinate position is (x, y), (x1,y1) Is Cx,yAny one of the coordinate positions of (a) and (b),
to represent
The middle coordinate position is (x)1,y1) The pixel value of the pixel point of (a),
to represent
The middle coordinate position is (x) 1,y1) The pixel value of the pixel point of (a),
represent
The middle coordinate position is (x)1,y1) The pixel value of the pixel point of (1);
(r-2 b) calculation
Distribution of gray histogram of (1), noted as { hk(j) J is more than or equal to 1 and less than or equal to 256 }; then will { h }k(j) The coordinate of the node with the minimum coordinate in all nodes with the non-zero histogram value in the [ 1 is not less than j is not more than 256 ] is recorded as XminWill { h }k(j) The coordinate of the node with the maximum coordinate in all nodes with the non-zero histogram value in the [ 1 is not less than j is not more than 256 ] is recorded as XmaxWill be
The middle pixel value belongs to [ Xmin,Xmid]The set of pixel values of all the pixels in the range is recorded as omega1Will be
Middle pixel value belongs to (X)mid,Xmax]The set of pixel values of all the pixels in the range is recorded as omega2(ii) a Wherein j is a positive integer, j is more than or equal to 1 and less than or equal to 256, and hk(j) Represents { h }k(j) J is more than or equal to 1 and less than or equal to 256, the histogram value of the node with the coordinate of j,
(symbol)
is a rounding down operation symbol;
r 2c, by maximizing omega1Obtain a first threshold, denoted as X1 *，
And by maximizing Ω2Obtain a second threshold, denoted as X2 *，
Wherein the content of the first and second substances,
express the finding such that
X when the value of (A) is maximum1Value of (A), X1Is omega1Of any one pixel value, Pf(X1) Represents omega1In (A) is [ X ]min,X1) Probability density function, mu, of all pixel values in the rangef(X1) Represents omega1In (A) is [ X ] min,X1) Mean, σ, of all pixel values in the rangef(X1) Represents Ω1In (A) is [ X ]min,X1) Standard deviation, μ, of all pixel values within a rangeb(X1) Represents omega1In (A) is [ X ]1,Xmid]Mean, σ, of all pixel values in the rangeb(X1) Represents omega1In (A) is [ X ]1,Xmid]The standard deviation of all pixel values within the range,
express the finding such that
X when the value of (A) is maximum2Value of (A), X2Is omega2Of any one pixel value, Pf(X2) Represents omega2In (A) is [ X ]mid,X2) Probability density function, mu, of all pixel values within a rangef(X2) Represents omega2In (A) is [ X ]mid,X2) Mean, σ, of all pixel values in the rangef(X2) Represents omega2In (A) is [ X ]mid,X2) Standard deviation, μ, of all pixel values within a rangeb(X2) Represents omega2In (A) is [ X ]2,Xmax]Mean, σ, of all pixel values in the rangeb(X2) Represents omega2In (A) is [ X ]2,Xmax]Standard deviation of all pixel values within the range;
(ii) (-) 2 d)
Middle pixel value belongs to (X)2 *,Xmax]The area formed by all pixel points in the range is determined as a bright area
Will be provided with
The middle pixel value belongs to [ Xmin,X1 *) The area formed by all pixel points in the range is determined as a dark area
Will be provided with
The middle pixel value belongs to [ X1 *,X2 *]The area formed by all pixel points in the range is determined as a normal area
In the step (r _ 3)Is/are as follows
The acquisition process comprises the following steps:
r 3a1
From the RGB color space to the CIELAB color space, 
The three components in the CIELAB color space are a luma component, a first chroma component, and a second chroma component, respectively;
(r-3 b 1) providing
Dividing into M non-overlapping sub-blocks of size 8 × 8 if
The sub-blocks with the size of 8 multiplied by 8 can not be equally divided, and redundant pixel points are removed; then will be
The brightness components of all the pixel points in each sub-block form a matrix with dimension of 8 x 8, and the matrix is divided into two parts
The matrix with dimension of 8 multiplied by 8 formed by the brightness components of all the pixel points in the t-th sub-block is marked as zt(ii) a Wherein M is a positive integer, M is more than 1, t is a positive integer, the initial value of t is 1, and t is more than or equal to 1 and less than or equal to M;
r 3c1 pair
Performing two-dimensional discrete cosine transform on a matrix with dimension of 8 multiplied by 8 and formed by the brightness components of all pixel points in each sub-block to obtain a corresponding discrete cosine transform coefficient matrix, and converting z into ztThe corresponding matrix of discrete cosine transform coefficients is denoted as Zt(ii) a Then calculate
The sum of all high frequency coefficients and all intermediate frequency coefficients in the discrete cosine transform coefficient matrix corresponding to each sub-block in the Z-transform coefficient matrix is obtainedtThe sum of all high frequency coefficients and all medium frequency coefficients in (1) is denoted as St(ii) a Wherein Z istHas a dimension of 8 × 8;
(r-3 d 1) calculation
Is characterized by 
(r-3 e 1) calculation
The mean value and the standard deviation of the brightness components of all the pixel points in (1) are correspondingly recorded as
And
r 3f1
And
vectors constructed in a sequential arrangement as
Wherein the symbol "[ alpha ],")]"representing symbols for vectors，
Show that
And
connected to form a vector.
Said step (r-3)
The acquisition process comprises the following steps:
r 3a2
From the RGB color space to the CIELAB color space,
the three components in the CIELAB color space are a luma component, a first chroma component, and a second chroma component, respectively;
(r-3 b 2) calculation
Sum of luminance components of all pixel points in
The first regional contrast of the brightness components of all the pixels in (1), which is recorded as
And calculate
Sum of luminance components of all pixel points in
The second regional contrast of the brightness components of all the pixel points in (1), is recorded as
Wherein the symbol "|" is an absolute value symbol,
to represent
The average of the luminance components of all the pixel points in (a),
to represent
The standard deviation of the luminance components of all the pixel points in (a),
to represent
The average of the luminance components of all the pixel points in (a),
to represent
The standard deviation of the brightness components of all the pixel points in the image is xi, which is a control parameter;
(r-3 c 2) calculation
Sum of luminance components of all pixel points in 
The first regional contrast of the brightness components of all the pixel points in (1) is recorded as
And calculate
Luminance component sum of all pixel points in
The second regional contrast of the brightness components of all the pixel points in (1) is recorded as
Wherein the content of the first and second substances,
to represent
The average of the luminance components of all the pixel points in (a),
to represent
The standard deviation of the luminance components of all the pixel points in (1);
(r-3 d 2) calculation
First chrominance component sum of all pixel points in
The first regional contrast of the first chrominance components of all the pixel points in (1) is recorded as
And calculate
First chrominance component sum of all pixel points in
The second regional contrast of the first chrominance components of all the pixel points in (1) is recorded as
Wherein the content of the first and second substances,
to represent
The average of the first chrominance components of all the pixels in (a),
to represent
The standard deviation of the first chrominance components of all the pixel points in (a),
to represent
The average of the first chrominance components of all the pixels in (a),
to represent
The standard deviation of the first chrominance components of all the pixel points in (1);
(r-3 e 2) calculation
First chrominance component sum of all pixel points in
The first regional contrast of the first chrominance components of all the pixel points in (1) is recorded as
And calculate
First chrominance component sum of all pixel points in 
The second regional contrast of the first chrominance components of all the pixel points in (1) is recorded as
Wherein the content of the first and second substances,
to represent
The average of the first chrominance components of all the pixels in (a),
to represent
The standard deviation of the first chrominance components of all the pixel points in (1);
r 3f2
Vectors constructed in a sequential arrangement as
Wherein the symbol "[ alpha ],")]"is a vector representing a symbol and,
show that
And
connected to form a vector.
Compared with the prior art, the invention has the advantages that:
the method takes the influence of bright area characteristics and dark area characteristics on tone mapping into consideration, extracts bright and dark area characteristic vectors of a tone mapping image, extracts area contrast characteristic vectors of the tone mapping image at the same time, then constructs a global characteristic vector, and trains the global characteristic vectors of all tone mapping images in a training image set by using support vector regression to construct a quality prediction model; in the testing stage, the objective quality predicted value of the tone mapping image is obtained through calculating the global feature vector of the tone mapping image used for testing and predicting according to the quality prediction model constructed in the training stage, and because the obtained global feature vector information has stronger stability and can better reflect the quality change condition of the tone mapping image, the correlation between the objective evaluation result and the subjective perception is effectively improved.
Drawings
Fig. 1 is a block diagram of a general implementation of the method of the present invention.
Detailed Description
The invention is described in further detail below with reference to the following examples of the drawings.
The general implementation block diagram of the tone mapping image quality evaluation method provided by the invention is shown in fig. 1, and the method comprises a training stage and a testing stage;
the specific steps of the training phase process are as follows:
firstly, 1, selecting N tone mapping images with width W and height H to form a training image set, and recording the kth tone mapping image in the training image set as
Wherein, N is a positive integer, N is more than 1, if N is 1000, k is a positive integer, the initial value of k is 1, and k is more than or equal to 1 and less than or equal to N.
Firstly, 2, dividing each tone mapping image in the training image set into a bright area, a dark area and a normal area, and dividing the bright area, the dark area and the normal area
The bright area, the dark area and the normal area are correspondingly recorded as
And
in this embodiment, in step (r _ 2)
And
the acquisition process comprises the following steps:
(ii) (-) 2 a)
R, G, and B components in the RGB color space are expressed as
Then calculate
Is recorded as a dark channel image
Will be provided with
The pixel value of the pixel point with the middle coordinate position (x, y) is recorded as
Wherein x is more than or equal to 1 and less than or equal to W, y is more than or equal to 1 and less than or equal to H, min () is a function for taking the minimum value, C x,yRepresenting a set of coordinate positions of all pixel points within a 3 × 3 neighborhood range centered on a pixel point whose coordinate position is (x, y), (x1,y1) Is Cx,yAny one of the coordinate positions of (a) and (b),
to represent
The middle coordinate position is (x)1,y1) The pixel value of the pixel point of (a),
to represent
The middle coordinate position is (x)1,y1) The pixel value of the pixel point of (a),
to represent
The middle coordinate position is (x)1,y1) The pixel value of the pixel point of (1).
(r-2 b) calculation
Distribution of gray histogram of (1), noted as { hk(j) J is more than or equal to 1 and less than or equal to 256 }; then will { h }k(j) The coordinate of the node with the minimum coordinate in all nodes with the non-zero histogram value in the [ 1 is not less than j is not more than 256 ] is recorded as XminWill { h }k(j) The coordinate of the node with the maximum coordinate in all nodes with the non-zero histogram value in the [ 1 is not less than j is not more than 256 ] is recorded as XmaxWill be
The middle pixel value belongs to [ Xmin,Xmid]The set of pixel values of all the pixels in the range is recorded as omega1Will be
Middle pixel value belongs to (X)mid,Xmax]The set of pixel values of all the pixels in the range is recorded as omega2(ii) a Wherein j is a positive integer, j is more than or equal to 1 and less than or equal to 256, and hk(j) Represents { h }k(j) J is more than or equal to 1 and less than or equal to 256, the histogram value of the node with the coordinate of j,
(symbol)
to round the operator down.
R 2c, by maximizing omega1Obtain a first threshold, denoted as X1 *，
And by maximizing omega 2Obtain a second threshold, denoted as X2*，
Wherein, the first and the second end of the pipe are connected with each other,
express the solution such that
X when the value of (A) is maximum1Value of (A), X1Is omega1Of any one pixel value, Pf(X1) Represents omega1In (A) is [ X ]min,X1) Probability density function, mu, of all pixel values within a rangef(X1) Represents omega1In (A) is [ X ]min,X1) Mean, σ, of all pixel values in the rangef(X1) Represents omega1In (A) is [ X ]min,X1) Standard deviation, μ, of all pixel values within a rangeb(X1) Represents omega1In (A) is [ X ]1,Xmid]Mean, σ, of all pixel values in the rangeb(X1) Represents omega1In (A) is [ X ]1,Xmid]The standard deviation of all pixel values within the range,
express the finding such that
X when the value of (A) is maximum2Value of (A), X2Is omega2Of any one pixel value, Pf(X2) Watch (A)Show omega2In (A) is [ X ]mid,X2) Probability density function, mu, of all pixel values within a rangef(X2) Represents omega2In (A) is [ X ]mid,X2) Mean, σ, of all pixel values in the rangef(X2) Represents omega2In (A) is [ X ]mid,X2) Standard deviation, μ, of all pixel values within a rangeb(X2) Represents omega2In (A) is [ X ]2,Xmax]Mean, σ, of all pixel values in the rangeb(X2) Represents omega2In (A) is [ X ]2,Xmax]Standard deviation of all pixel values within the range.
(ii) (-) 2 d)
Middle pixel value belongs to (X)2 *,Xmax]The area formed by all pixel points in the range is determined as a bright area
Will be provided with
The middle pixel value belongs to [ X min,X1 *) Determining the area formed by all pixel points in the range as a dark area
Will be provided with
The middle pixel value belongs to [ X1 *,X2 *]The area formed by all pixel points in the range is determined as a normal area
Calculating a bright and dark region characteristic vector of each tone mapping image in the training image set according to a bright region and a dark region of each tone mapping image in the training image set, and calculating the characteristic vector of each tone mapping image in the training image set according to the bright region and the dark region
The light and dark region feature vector is recorded as
And calculating the regional contrast characteristic vector of each tone mapping image in the training image set according to the bright region, the dark region and the normal region of each tone mapping image in the training image set, and calculating the contrast characteristic vector of each tone mapping image in the training image set according to the bright region, the dark region and the normal region
Is noted as the area contrast feature vector
Wherein the content of the first and second substances,
has a dimension of 3 x 1,
dimension (d) is 8 × 1.
In this embodiment, in step (r _ 3)
The acquisition process comprises the following steps:
r 3a1
From the RGB color space to the CIELAB color space,
the three components in the CIELAB color space are the luma component, the first chroma component (referred to as component a) and the second chroma component (referred to as component b), respectively.
R 3b1
Dividing into M non-overlapping sub-blocks of size 8 x 8If, if
The subblocks with the size of 8 multiplied by 8 can not be equally divided, redundant pixel points are removed, namely the redundant pixel points are not considered; then will be 
The brightness components of all pixel points in each sub-block form a matrix with dimension of 8 x 8, and the matrix is divided into two parts
The matrix with dimension of 8 multiplied by 8 formed by the brightness components of all the pixel points in the t-th sub-block is marked as zt(ii) a Wherein M is a positive integer, M is more than 1, t is a positive integer, the initial value of t is 1, and t is more than or equal to 1 and less than or equal to M.
R 3c1 pair
Performing two-dimensional discrete cosine transform on a matrix with dimension of 8 multiplied by 8 and formed by the brightness components of all pixel points in each sub-block to obtain a corresponding discrete cosine transform coefficient matrix, and converting z into ztThe corresponding matrix of discrete cosine transform coefficients is denoted as Zt(ii) a Then calculate
The sum of all high frequency coefficients and all intermediate frequency coefficients in the discrete cosine transform coefficient matrix corresponding to each sub-block in the Z-transform coefficient matrix is obtainedtThe sum of all high frequency coefficients and all medium frequency coefficients in (1) is denoted as St(ii) a Wherein Z istThe dimension of (2) is 8 multiplied by 8, the upper left corner part in the discrete cosine transform coefficient matrix is direct current and low frequency coefficients, the lower right corner part is high frequency coefficients, and the middle part is intermediate frequency coefficients.
(r-3 d 1) calculation
Is characterized by
(r-3 e 1) calculation
The mean and standard deviation of the brightness components of all the pixels in (1) are correspondingly recorded as
And
r 3f1
And
vectors constructed in a sequential arrangement as 
Wherein the symbol "[ 2 ]]"is a vector representing a symbol and,
show that
And
connected to form a vector.
In this embodiment, in step (r _ 3)
Is obtained by：
R 3a2
From the RGB color space to the CIELAB color space,
the three components in the CIELAB color space are the luma component, the first chroma component (referred to as component a) and the second chroma component (referred to as component b), respectively.
(r-3 b 2) calculation
Sum of luminance components of all pixel points in
The first regional contrast of the brightness components of all the pixels in (1), which is recorded as
And calculate
Sum of luminance components of all pixel points in
The second regional contrast of the brightness components of all the pixel points in (1), is recorded as
Wherein the symbol "|" is an absolute value symbol,
to represent
The average of the luminance components of all the pixel points in (a),
to represent
The standard deviation of the luminance components of all the pixel points in (a),
to represent
The average of the luminance components of all the pixel points in (a),
to represent
The standard deviation of the luminance components of all the pixels in (1), ξ is a control parameter, which is 10 in this embodiment-6。
(r-3 c 2) calculation
Sum of luminance components of all pixel points in
The first regional contrast of the brightness components of all the pixels in (1), which is recorded as 
And calculate
Luminance component sum of all pixel points in
The second regional contrast of the brightness components of all the pixel points in (1) is recorded as
Wherein, the first and the second end of the pipe are connected with each other,
represent
The average of the luminance components of all the pixel points in (a),
to represent
The standard deviation of the luminance components of all the pixel points in (1).
(r-3 d 2) calculation
First chrominance component sum of all pixel points in
The first regional contrast of the first chrominance components of all the pixel points in (1) is recorded as
And calculate
First chrominance component sum of all pixel points in
The second regional contrast of the first chrominance components of all the pixel points in (1) is recorded as
Wherein the content of the first and second substances,
to represent
The average of the first chrominance components of all the pixels in (a),
to represent
The standard deviation of the first chrominance components of all the pixel points in (a),
to represent
The average of the first chrominance components of all the pixels in (a),
to represent
The standard deviation of the first chrominance components of all the pixel points in (1).
(r-3 e 2) calculation
First chrominance component sum of all pixel points in
The first regional contrast of the first chrominance components of all the pixel points in (1) is recorded as
And calculate
First chrominance component sum of all pixel points in
The second regional contrast of the first chrominance components of all the pixel points in (1) is recorded as 
Wherein, the first and the second end of the pipe are connected with each other,
to represent
The average of the first chrominance components of all the pixels in (a),
to represent
The standard deviation of the first chrominance components of all the pixel points in (1).
R 3f2
Vectors constructed in a sequential arrangement as
Wherein the symbol "[ alpha ],")]"is a vector representing a symbol and,
show that
And
connected to form a vector.
Firstly, 4, forming a global feature vector by a bright and dark region feature vector and a region contrast feature vector of each tone mapping image in the training image set, and then carrying out color matching on the global feature vector
Is noted as Fk，
Wherein, FkHas a dimension of 11X 1, symbol "[ 2 ]]"is a vector representing a symbol and,
show that
And
connected to form a vector.
Firstly, 5, forming a training sample data set by global feature vectors and average subjective score difference values of all tone mapping images in the training image set, wherein the training sample data set comprises N global feature vectors and N average subjective score difference values;then, a support vector regression is adopted as a machine learning method to train all global feature vectors in the training sample data set, so that the error between a regression function value obtained through training and the average subjective score difference is minimum, and the optimal weight vector is obtained through fitting 
And an optimum bias term
Then using the optimal weight vector
And an optimum bias term
Structural quality prediction model, as
Wherein, the first and the second end of the pipe are connected with each other,
in functional representation, F is used to represent the global feature vector of the tone-mapped image, and as input vector to the quality prediction model,
is composed of
The transpose of (a) is performed,
as a linear function of F.
The test stage process comprises the following specific steps:
② for any tone mapping image I used as testtestPush-buttonObtaining I according to the same operation from the step (r _ 2) to the step (r _ 4)testGlobal feature vector of (2), noted as Ftest(ii) a Then, according to the quality prediction model pair F constructed in the training stagetestTesting and predicting to obtain FtestCorresponding predicted value is taken as ItestThe objective quality prediction value of (1) is recorded as
Wherein, ItestHas a width W 'and a height H', FtestHas a dimension of 11 x 1,
is represented by FtestIs a linear function of (a).
In the present embodiment, as the tone mapping image database, the TMID database established by the university of luugu, canada and the ESPL-LIVE database established by the austin division, university of texas, usa were used, and the TMID database included 120 tone mapping images and the ESPL-LIVE database included 1811 tone mapping images. 2 common objective parameters of the evaluation method for evaluating the image quality are used as evaluation indexes, namely Pearson Linear Correlation Coefficient (PLCC) and Spearman Rank Order Correlation Coefficient (SROCC) under the condition of nonlinear regression. The higher PLCC and SROCC indicate the better correlation between the evaluation results of the method of the invention and the difference of the average subjective scores. Table 1 shows the correlation between the objective quality prediction value obtained by the method of the present invention and the mean subjective score difference.
TABLE 1 correlation between objective quality prediction values and mean subjective score differences obtained by the method of the invention
| Database with a plurality of databases | PLCC | SROCC |
| TMID | 0.827 | 0.758 |
| ESPL-LIVE | 0.658 | 0.660 |
As can be seen from Table 1, the correlation between the objective quality prediction value of the tone mapping image obtained by the method of the present invention and the average subjective score difference is very high, which indicates that the objective evaluation result is more consistent with the result of human eye subjective perception, and is sufficient to illustrate the effectiveness of the method of the present invention.
Claims (2)
1. A tone mapping image quality evaluation method is characterized by comprising a training stage and a testing stage;
the specific steps of the training phase process are as follows:
firstly, 1, selecting N tone mapping images with width W and height H to form a training image set, and recording the kth tone mapping image in the training image set as
Wherein N is a positive integer, N is more than 1, k is a positive integer, the initial value of k is 1, and k is more than or equal to 1 and less than or equal to N;
firstly, 2, dividing each tone mapping image in the training image set into a bright area, a dark area and a normal area, and dividing the bright area, the dark area and the normal area
The bright area, the dark area and the normal area are correspondingly recorded as
And
calculating a bright and dark region characteristic vector of each tone mapping image in the training image set according to a bright region and a dark region of each tone mapping image in the training image set, and calculating the characteristic vector of each tone mapping image in the training image set according to the bright region and the dark region 
The light and dark region feature vector of
And calculating the regional contrast characteristic vector of each tone mapping image in the training image set according to the bright region, the dark region and the normal region of each tone mapping image in the training image set, and calculating the contrast characteristic vector of each tone mapping image in the training image set according to the bright region, the dark region and the normal region
Is noted as the area contrast feature vector
Wherein, the first and the second end of the pipe are connected with each other,
has a dimension of 3 x 1,
has a dimension of 8 × 1;
said step (r-3)
The acquisition process comprises the following steps:
r 3a1
Conversion from RGB color spaceThe color space of the CIELAB is such that,
the three components in the CIELAB color space are a luma component, a first chroma component, and a second chroma component, respectively;
r 3b1
Dividing into M non-overlapping sub-blocks of size 8 × 8 if
The sub-blocks with the size of 8 multiplied by 8 can not be equally divided, and redundant pixel points are removed; then will be
The brightness components of all pixel points in each sub-block form a matrix with dimension of 8 x 8, and the matrix is divided into two parts
The matrix with dimension of 8 multiplied by 8 formed by the brightness components of all the pixel points in the t-th sub-block is marked as zt(ii) a Wherein M is a positive integer, M is more than 1, t is a positive integer, the initial value of t is 1, and t is more than or equal to 1 and less than or equal to M;
r 3c1 pair
Performing two-dimensional discrete cosine transform on a matrix with dimension of 8 multiplied by 8 and formed by the brightness components of all pixel points in each sub-block to obtain a corresponding discrete cosine transform coefficient matrix, and performing z-component transform on the matrix tThe corresponding matrix of discrete cosine transform coefficients is denoted as Zt(ii) a Then calculate
The sum of all high frequency coefficients and all intermediate frequency coefficients in the discrete cosine transform coefficient matrix corresponding to each sub-block in the Z-transform coefficient matrix is obtained by dividing the Z-transform coefficient matrix into Z-transform coefficientstIs given as S, the sum of all high frequency coefficients and all medium frequency coefficientst(ii) a Wherein, ZtDimension of (d) is 8 x 8;
(r-3 d 1) calculation
Is characterized by being
(r-3 e 1) calculation
The mean and standard deviation of the brightness components of all the pixels in (1) are correspondingly recorded as
And
r 3f1
And
vectors constructed in a sequential arrangement as
Wherein the symbol "[ alpha ],")]"is a vector representing a symbol and,
show that
And
connected to form a vector;
said step (r-3)
The acquisition process comprises the following steps:
r 3a2
From the RGB color space to the CIELAB color space,
the three components in the CIELAB color space are a luma component, a first chroma component, and a second chroma component, respectively;
(r-3 b 2) calculation
Sum of luminance components of all pixel points in
The first regional contrast of the brightness components of all the pixels in (1), which is recorded as
And calculate
Sum of luminance components of all pixel points in
The second regional contrast of the brightness components of all the pixel points in (1), is recorded as
Wherein the symbol "|" is an absolute value symbol, 
Represent
The average of the luminance components of all the pixel points in (a),
represent
The standard deviation of the luminance components of all the pixel points in (a),
to represent
The average of the luminance components of all the pixel points in (a),
to represent
The standard deviation of the brightness components of all the pixel points in the image is xi, which is a control parameter;
(r-3 c 2) calculation
Sum of luminance components of all pixel points in
The first regional contrast of the brightness components of all the pixels in (1), which is recorded as
And calculate
Sum of luminance components of all pixel points in
The second regional contrast of the brightness components of all the pixel points in (1), is recorded as
Wherein the content of the first and second substances,
to represent
The average of the luminance components of all the pixel points in (a),
to represent
The standard deviation of the luminance components of all the pixel points in (1);
(r-3 d 2) calculation
First chrominance component sum of all pixel points in
The first regional contrast of the first chrominance components of all the pixel points in (1) is recorded as
And calculate
First chrominance component sum of all pixel points in
The second regional contrast of the first chrominance components of all the pixel points in (1) is recorded as
Wherein the content of the first and second substances,
to represent
The average of the first chrominance components of all the pixels in (a),
to represent
The standard deviation of the first chrominance components of all the pixel points in (a), 
To represent
The average of the first chrominance components of all the pixels in (a),
to represent
All ofA standard deviation of a first chrominance component of the pixel;
(r-3 e 2) calculation
First chrominance component sum of all pixel points in
The first regional contrast of the first chrominance components of all the pixel points in (1) is recorded as
And calculate
First chrominance component sum of all pixel points in
The second regional contrast of the first chrominance components of all the pixel points in (1) is recorded as
Wherein the content of the first and second substances,
to represent
The average of the first chrominance components of all the pixels in (a),
to represent
The standard deviation of the first chrominance components of all the pixel points in (1);
r 3f2
Vectors constructed in a sequential arrangement as
Wherein the symbol "[ alpha ],")]"is a vector representing a symbol and,
show that
And
connected to form a vector;
firstly, 4, forming a global feature vector by a bright and dark region feature vector and a region contrast feature vector of each tone mapping image in the training image set, and then carrying out color matching on the global feature vector
Is noted as Fk，
Wherein, FkHas a dimension of 11X 1, symbol "[ 2 ]]"is a vector representing a symbol and,
show that
And
connected to form a vector;
(r-5) to trainThe global feature vectors and the average subjective score difference values of all tone mapping images in the training image set form a training sample data set, and the training sample data set comprises N global feature vectors and N average subjective score difference values; then, a support vector regression is adopted as a machine learning method to train all global feature vectors in the training sample data set, so that the error between a regression function value obtained through training and the average subjective score difference is minimum, and the optimal weight vector is obtained through fitting 
And an optimum bias term
Then using the optimal weight vector
And an optimal bias term
Structural quality prediction model, as
Wherein the content of the first and second substances,
in functional representation, F is used to represent the global feature vector of the tone-mapped image, and as input vector to the quality prediction model,
is composed of
The transpose of (a) is performed,
is a linearity of FA function;
the test stage process comprises the following specific steps:
② for any tone mapping image I used as testtestObtaining I according to the same operation from the step I _2 to the step I _4testGlobal feature vector of (2), noted as Ftest(ii) a Then, according to the quality prediction model pair F constructed in the training stagetestTesting and predicting to obtain FtestCorresponding predicted value is taken as ItestThe objective quality prediction value of (1) is recorded as
Wherein, ItestHas a width W 'and a height H', FtestHas a dimension of 11 x 1,
is represented by FtestIs a linear function of (a).
2. The method for evaluating the quality of a tone-mapped image according to claim 1, wherein said step (r _ 2) is
And
the acquisition process comprises the following steps:
(ii) (-) 2 a)
R, G, B components in the RGB color space are correspondingly denoted as
Then calculate
Is recorded as a dark channel image
Will be provided with
The pixel value of the pixel point with the middle coordinate position (x, y) is recorded as 
Wherein x is more than or equal to 1 and less than or equal to W, y is more than or equal to 1 and less than or equal to H, min () is a minimum value taking function, Cx,yRepresents a set of coordinate positions of all pixel points in a 3 × 3 neighborhood range centered on a pixel point whose coordinate position is (x, y), (x)1,y1) Is Cx,yIs determined at a position in any one of the coordinates,
represent
The middle coordinate position is (x)1,y1) The pixel value of the pixel point of (a),
to represent
The middle coordinate position is (x)1,y1) The pixel value of the pixel point of (a),
to represent
The middle coordinate position is (x)1,y1) The pixel value of the pixel point of (1);
(r-2 b) calculation
Distribution of gray histogram of (1), noted as { hk(j) J is more than or equal to 1 and less than or equal to 256 }; then { h } will bek(j) The coordinate of the node with the minimum coordinate in all nodes with the non-zero histogram value in the [ 1 is not less than j is not more than 256 ] is recorded as XminWill { h }k(j) The coordinate of the node with the maximum coordinate in all nodes with the non-zero histogram value in the [ 1 is not less than j is not more than 256 ] is recorded as XmaxWill be
The middle pixel value belongs to [ Xmin,Xmid]The set of pixel values of all the pixels in the range is recorded as omega1Will be
Middle pixel value belongs to (X)mid,Xmax]The set of pixel values of all the pixels in the range is recorded as omega2(ii) a Wherein j is a positive integer, j is more than or equal to 1 and less than or equal to 256, and hk(j) Represents { h }k(j) J is more than or equal to 1 and less than or equal to 256, the histogram value of the node with the coordinate of j,
(symbol)
is a rounding down operation symbol;
(r 2 c) by maximizing Ω 1Obtain a first threshold, denoted as X1 *，
And by maximizing omega2Obtain a second threshold, denoted as X2 *，
Wherein the content of the first and second substances,
express the finding such that
X when the value of (A) is maximum1Value of (A), X1Is omega1Of any one pixel value, Pf(X1) Represents omega1In (A) is [ X ]min,X1) Probability density function, mu, of all pixel values within a rangef(X1) Represents omega1In (A) is [ X ]min,X1) Mean, σ, of all pixel values in the rangef(X1) Represents omega1In (A) is [ X ]min,X1) Standard deviation, μ, of all pixel values within a rangeb(X1) Represents omega1In (A) is [ X ]1,Xmid]Mean, σ, of all pixel values in the rangeb(X1) Represents omega1In (A) is [ X ]1,Xmid]The standard deviation of all pixel values within the range,
express the finding such that
X when the value of (A) is maximum2Value of (A), X2Is omega2Of any one pixel value, Pf(X2) Represents omega2In (A) is [ X ]mid,X2) Probability density function, mu, of all pixel values within a rangef(X2) Represents omega2In (A) is [ X ]mid,X2) Mean, σ, of all pixel values in the rangef(X2) Represents omega2In (A) is [ X ]mid,X2) Standard deviation, μ, of all pixel values within a rangeb(X2) Represents omega2In (A) is [ X ]2,Xmax]Mean, σ, of all pixel values in the rangeb(X2) Represents omega2In (A) is [ X ]2,Xmax]Standard deviation of all pixel values within the range;
(ii) (-) 2 d)
Middle pixel value belongs to (X)2 *,Xmax]The area formed by all pixel points in the range is determined as a bright area 
Will be provided with
The middle pixel value belongs to [ X ]min,X1 *) Determining the area formed by all pixel points in the range as a dark area
Will be provided with
The middle pixel value belongs to [ X1 *,X2 *]The area formed by all pixel points in the range is determined as a normal area
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910881340.2A CN110717892B (en) | 2019-09-18 | 2019-09-18 | Tone mapping image quality evaluation method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910881340.2A CN110717892B (en) | 2019-09-18 | 2019-09-18 | Tone mapping image quality evaluation method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110717892A CN110717892A (en) | 2020-01-21 |
| CN110717892B true CN110717892B (en) | 2022-06-28 |
Family
ID=69209939
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910881340.2A Active CN110717892B (en) | 2019-09-18 | 2019-09-18 | Tone mapping image quality evaluation method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110717892B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112950596B (en) * | 2021-03-09 | 2023-06-02 | 宁波大学 | Tone mapping omnidirectional image quality evaluation method based on multiple areas and multiple levels |
| CN116630447B (en) * | 2023-07-24 | 2023-10-20 | 成都海风锐智科技有限责任公司 | Weather prediction method based on image processing |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101540048A (en) * | 2009-04-21 | 2009-09-23 | 北京航空航天大学 | Image quality evaluating method based on support vector machine |
| CN105741328A (en) * | 2016-01-22 | 2016-07-06 | 西安电子科技大学 | Shot image quality evaluation method based on visual perception |
| CN105761227A (en) * | 2016-03-04 | 2016-07-13 | 天津大学 | Underwater image enhancement method based on dark channel prior algorithm and white balance |
| CN107105223A (en) * | 2017-03-20 | 2017-08-29 | 宁波大学 | A kind of tone mapping method for objectively evaluating image quality based on global characteristics |
| CN107172418A (en) * | 2017-06-08 | 2017-09-15 | 宁波大学 | A kind of tone scale map image quality evaluating method analyzed based on exposure status |
| KR101846743B1 (en) * | 2016-11-28 | 2018-04-09 | 연세대학교 산학협력단 | Objective quality assessment method and apparatus for tone mapped images |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10467735B2 (en) * | 2015-08-25 | 2019-11-05 | Interdigital Vc Holdings, Inc. | Inverse tone mapping based on luminance zones |
| EP3319013A1 (en) * | 2016-11-03 | 2018-05-09 | Thomson Licensing | Method and device for estimating cast shadow regions and/or highlight regions in images |
-
2019
- 2019-09-18 CN CN201910881340.2A patent/CN110717892B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101540048A (en) * | 2009-04-21 | 2009-09-23 | 北京航空航天大学 | Image quality evaluating method based on support vector machine |
| CN105741328A (en) * | 2016-01-22 | 2016-07-06 | 西安电子科技大学 | Shot image quality evaluation method based on visual perception |
| CN105761227A (en) * | 2016-03-04 | 2016-07-13 | 天津大学 | Underwater image enhancement method based on dark channel prior algorithm and white balance |
| KR101846743B1 (en) * | 2016-11-28 | 2018-04-09 | 연세대학교 산학협력단 | Objective quality assessment method and apparatus for tone mapped images |
| CN107105223A (en) * | 2017-03-20 | 2017-08-29 | 宁波大学 | A kind of tone mapping method for objectively evaluating image quality based on global characteristics |
| CN107172418A (en) * | 2017-06-08 | 2017-09-15 | 宁波大学 | A kind of tone scale map image quality evaluating method analyzed based on exposure status |
Non-Patent Citations (4)
| Title |
|---|
| An Entropy-based Inverse Tone Mapping Operator for High Dynamic Range Applications;P. Mohammadi, M. T. Pourazad and P. Nasiopoulos;《2018 9th IFIP International Conference on New Technologies, Mobility and Security (NTMS)》;20180402;全文 * |
| Blind quality index for tone-mapped images based on luminance partition;Chen P, Li L, Zhang X, et al;《Pattern Recognition》;20190108;第110-113页 * |
| Transform domain measure of enhancement — TDME — For security imaging applications;Samani, A. , K. Panetta , and S. Agaian;《IEEE International Conference on Technologies for Homeland Security IEEE, 2014》;20140106;全文 * |
| 基于改进的最大类间方差法的水果图像识别研究;陈雪鑫，卜庆凯;《青海大学学报（工程技术版）》;20190531;第33-35页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110717892A (en) | 2020-01-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102333233B (en) | Stereo image quality objective evaluation method based on visual perception | |
| CN103996192B (en) | Non-reference image quality evaluation method based on high-quality natural image statistical magnitude model | |
| CN104902267B (en) | No-reference image quality evaluation method based on gradient information | |
| CN108074239B (en) | No-reference image quality objective evaluation method based on prior perception quality characteristic diagram | |
| CN107105223B (en) | A kind of tone mapping method for objectively evaluating image quality based on global characteristics | |
| CN106600597B (en) | It is a kind of based on local binary patterns without reference color image quality evaluation method | |
| CN105574901B (en) | A kind of general non-reference picture quality appraisement method based on local contrast pattern | |
| CN109978854B (en) | Screen content image quality evaluation method based on edge and structural features | |
| CN102547368B (en) | Objective evaluation method for quality of stereo images | |
| CN109218716B (en) | No-reference tone mapping image quality evaluation method based on color statistics and information entropy | |
| CN107146220B (en) | A kind of universal non-reference picture quality appraisement method | |
| CN106651829B (en) | A kind of non-reference picture method for evaluating objective quality based on energy and texture analysis | |
| CN110717892B (en) | Tone mapping image quality evaluation method | |
| CN112950596B (en) | Tone mapping omnidirectional image quality evaluation method based on multiple areas and multiple levels | |
| CN110120034B (en) | Image quality evaluation method related to visual perception | |
| CN110443800A (en) | The evaluation method of video image quality | |
| CN114598864A (en) | Full-reference ultrahigh-definition video quality objective evaluation method based on deep learning | |
| CN106023152B (en) | It is a kind of without with reference to objective evaluation method for quality of stereo images | |
| CN110910347A (en) | Image segmentation-based tone mapping image no-reference quality evaluation method | |
| CN106683079A (en) | No-reference image objective quality evaluation method based on structural distortion | |
| CN107292331B (en) | Based on unsupervised feature learning without reference screen image quality evaluating method | |
| CN114067006B (en) | Screen content image quality evaluation method based on discrete cosine transform | |
| Fu et al. | Image quality assessment using edge and contrast similarity | |
| Yan et al. | Blind image quality assessment based on natural redundancy statistics | |
| CN112950479B (en) | Image gray level region stretching algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20231226 Address after: 315000 North of First Floor, Fifth Avenue, No. 719 Zhongxing Road, Yinzhou District, Ningbo City, Zhejiang Province Patentee after: Ningbo Frontier Digital Technology Co.,Ltd. Address before: 200120 building C, No. 888, Huanhu West 2nd Road, Lingang New Area, China (Shanghai) pilot Free Trade Zone, Pudong New Area, Shanghai Patentee before: Shanghai ruishenglian Information Technology Co.,Ltd. Effective date of registration: 20231226 Address after: 200120 building C, No. 888, Huanhu West 2nd Road, Lingang New Area, China (Shanghai) pilot Free Trade Zone, Pudong New Area, Shanghai Patentee after: Shanghai ruishenglian Information Technology Co.,Ltd. Address before: 315211, Fenghua Road, Jiangbei District, Zhejiang, Ningbo 818 Patentee before: Ningbo University |