CN112330657B - Image quality evaluation method and system based on gray scale characteristics - Google Patents

Image quality evaluation method and system based on gray scale characteristics Download PDF

Info

Publication number
CN112330657B
CN112330657B CN202011314950.3A CN202011314950A CN112330657B CN 112330657 B CN112330657 B CN 112330657B CN 202011314950 A CN202011314950 A CN 202011314950A CN 112330657 B CN112330657 B CN 112330657B
Authority
CN
China
Prior art keywords
sub
block image
feature
gray
image
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
Application number
CN202011314950.3A
Other languages
Chinese (zh)
Other versions
CN112330657A (en
Inventor
罗文峰
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hunan Upixels Technology Co ltd
Original Assignee
Hunan Upixels Technology Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Hunan Upixels Technology Co ltd filed Critical Hunan Upixels Technology Co ltd
Priority to CN202011314950.3A priority Critical patent/CN112330657B/en
Publication of CN112330657A publication Critical patent/CN112330657A/en
Application granted granted Critical
Publication of CN112330657B publication Critical patent/CN112330657B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/90—Dynamic range modification of images or parts thereof
    • 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
    • G06T2207/20021—Dividing image into blocks, subimages or windows
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30168—Image quality inspection

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Data Mining & Analysis (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Quality & Reliability (AREA)
  • Image Analysis (AREA)

Abstract

An image quality evaluation method and system based on gray characteristics, the method comprises the following steps: step S1: the method comprises the steps of performing blocking processing on a reference image and an image to be evaluated, and dividing the reference image and the image to be evaluated into a first sub-block image and a second sub-block image with preset sizes respectively; step S2: calculating gray characteristic indexes of each first sub-block image and each second sub-block image; step S3: dividing the first sub-block image and the second sub-block image into a first category and a second category according to the gray characteristic index; step S4: respectively extracting a first characteristic of each first sub-block image and a second characteristic of each second sub-block image in the first category; step S5: and respectively extracting a third characteristic of each first sub-block image and a fourth characteristic of each second sub-block image in the second category. The invention can accurately evaluate the quality of the image, has simple algorithm, fully considers the correlation among pixels and has high practical value.

Description

Image quality evaluation method and system based on gray scale characteristics
Technical Field
The present invention relates to the field of image processing technologies, and in particular, to an image quality evaluation method and system based on gray scale characteristics.
Background
With the rapid development of multimedia technology, digital images are widely favored by people due to the characteristics of intuitiveness, reality and richness. In the image processing process, factors such as an image imaging system, an image storage device, a transmission medium, a processing mechanism of an image at a terminal and the like of the digital image inevitably cause image distortion, and the image distortion degree can directly reflect the performance of a multimedia transmission system and the service quality of the multimedia transmission system. Therefore, the image quality evaluation algorithm is an important index for evaluating the performance of the multimedia transmission system as an objective evaluation criterion for the quality of the image.
The quality evaluation method can be classified into 3 kinds of full reference quality evaluation, no reference quality evaluation and half reference quality evaluation according to how much reference information is acquired. Wherein, the full reference image quality evaluation algorithm uses the original image as a reference image of the distorted image; only information of part of reference images is used in the semi-reference image quality evaluation algorithm; the no reference image quality assessment algorithm does not use any information in the reference image as a priori data.
The most common method at present is the full reference image quality assessment algorithm. The traditional full-reference objective image quality evaluation algorithms have mean square error and peak signal to noise ratio, and are widely used all the time due to the simple calculation method and clear physical meaning, but the algorithms only analyze the image in a statistical sense, and do not consider the correlation among pixels.
The foregoing description is provided for general background information and does not necessarily constitute prior art.
Disclosure of Invention
The invention aims to provide the image quality evaluation method and the system based on the gray scale characteristics, which can accurately evaluate the quality of the image, have simple algorithm, fully consider the correlation among pixels and have high practical value.
The invention provides an image quality evaluation method based on gray characteristics, which comprises the following steps: step S1: dividing the reference image and the image to be evaluated into a first sub-block image and a second sub-block image with preset sizes respectively, and marking the first sub-block image and the second sub-block image as respectivelyAnd/>Wherein N represents the number of all sub-blocks after partitioning; step S2: calculating gray characteristic indexes of each first sub-block image and each second sub-block image; step S3: dividing the first sub-block image and the second sub-block image into a first category and a second category according to the gray characteristic index; step S4: respectively extracting a first characteristic of each first sub-block image and a second characteristic of each second sub-block image in the first category; step S5: extracting a third feature of each first sub-block image and a fourth feature of each second sub-block image in the second category respectively; step S6: according to the first characteristic and the second characteristic, calculating to obtain a first similarity index of the first category; step S7: according to the third characteristic and the fourth characteristic, calculating to obtain a second similar index of the second category; step S8: and calculating the final similarity of the reference image and the image to be evaluated according to the first similarity index and the second similarity index.
Further, any one of the first sub-block images is selectedThe step S2 includes: step S21: random at/>Selecting 10 points, and taking the 10 points as the center and the diameter as/>In the field of mm, calculating a gray average value to obtain 10 gray average values/>Statistics of the variability/>, of 10 of the gray-scale averagesThe specific formula is as follows: /(I)Wherein/>; Step 22: when the degree of difference/>When the first sub-block is made to imageAnd the corresponding second sub-block image/>Gray characteristic index/>; Otherwise/>。
Further, in the step S3, the first category is specifically: the gray characteristic indexFirst sub-block image set/>And with said/>Corresponding second sub-block image set/>Wherein/>Representing the gray characteristic index/>The number of sub-blocks of (a); the second category is specifically: the gray characteristic index/>Is a first sub-block image set of (1)And with said/>Corresponding second sub-block image setWherein/>Representing the gray characteristic index/>Is a number of sub-blocks of (c).
Further, the method comprises the steps ofAny first sub-block image/>The step S4 includes: step S41: will said/>Dividing into sixteen equal parts, and counting the gray level histogram of the first sub-block image to obtain a vector with ten six dimensions, which is recorded as/>; Step S42: calculating the mean and variance of the first sub-block images, and storing the mean and variance as features, so that each first sub-block image obtains a two-dimensional vector, which is recorded as/>; Step S43: will said/>And said/>Together, obtain the first sub-block image/>The first feature of (2) is an eighteen-dimensional vector, and the first feature is finally obtained and is marked as/>; Selecting the/>Any second sub-block image/>The calculation method of the second feature is identical to the calculation methods of the step S41, the step S42 and the step S43, and finally the second feature is obtained and is marked as/>。
Further, the method comprises the steps ofAny first sub-block image/>The step S5 includes: step S51: calculating the/>, respectively, by using sobel operatorsHorizontal gradient information of/>And vertical gradient information/>：/>Wherein/>，/>Representing a convolution operation; step S52: by means of said/>And said/>Calculate the first gradient magnitude/>And a first gradient direction/>The specific formula is as follows: /(I); Step S53: according to the first gradient direction/>Finally, the third feature is calculated as/>The method specifically comprises the following steps: /(I); Selecting the saidAny second sub-block image/>The calculation mode of the fourth feature is identical to the calculation modes of the step S51, the step S52 and the step S53, and finally the second gradient amplitude/> isobtainedAnd the fourth feature is:/>。
Further, step S6 includes: step S61: optionally one of theAnd said/>Corresponding said/>Obtaining the similarity degree/>, of each pair of sub-block images：Wherein/>Representing a summation operation; step S62: for all said/>Summing to obtain the first similarity index/>：/>。
Further, the step S7 includes: step S71: optionally one of theAnd with the saidCorresponding said/>Obtaining the similarity degree/>, of each pair of sub-block images：Wherein/>Representing a summation operation,/>Representing a dot product operation; step S72: for all said/>Summing to obtain the second similarity index/>：/>。
Further, the step S8 specifically includes: the similarity E is the sum of the first similarity index and the second similarity index, and the specific formula is as follows:。
the invention also provides an image quality evaluation system based on gray characteristics, which comprises a separation module, a classification module, an extraction module and a calculation module, wherein the separation module is used for carrying out blocking processing on the reference image and the image to be evaluated, and dividing the reference image and the image to be evaluated into a first sub-block image and a second sub-block image with preset sizes respectively, which are respectively recorded as AndWherein N represents the number of all sub-blocks after partitioning; the calculation module is used for calculating gray characteristic indexes of each first sub-block image and each second sub-block image; the classification module is used for classifying the first sub-block image and the second sub-block image into a first category and a second category according to the gray characteristic index; the extraction module is used for respectively extracting a first feature of each first sub-block image and a second feature of each second sub-block image in the first category, and respectively extracting a third feature of each first sub-block image and a fourth feature of each second sub-block image in the second category; the calculation module is further configured to calculate a first similarity index of the first class according to the first feature and the second feature, calculate a second similarity index of the second class according to the third feature and the fourth feature, and calculate a final similarity of the reference image and the image to be evaluated according to the first similarity index and the second similarity index.
Further, any one of the first sub-block images is selectedThe gray characteristic index is obtained by the following steps: random at/>Selecting 10 points, and taking the 10 points as the center and the diameter as/>In the field of mm, calculating a gray average value to obtain 10 gray average values/>Statistics of the variability/>, of 10 of the gray-scale averagesThe specific formula is as follows: /(I)Wherein/>; When the degree of difference/>When the first sub-block is made to imageAnd the corresponding second sub-block image/>Gray characteristic index/>; Otherwise/>; The first category is specifically: the gray characteristic index/>Is a first sub-block image set of (1)And with said/>Corresponding second sub-block image setWherein/>Representing the gray characteristic index/>The number of sub-blocks of (a); the second category is specifically: the gray characteristic index/>First sub-block image set/>And with said/>Corresponding second sub-block image set/>WhereinRepresenting the gray characteristic index/>The number of sub-blocks of (a); selecting the/>Any first sub-block image/>The first characteristic is obtained by the following steps: will said/>Dividing into sixteen equal parts, and counting the gray level histogram of the first sub-block image to obtain a vector with ten six dimensions, which is recorded as/>; Calculating the mean and variance of the first sub-block images, and storing the mean and variance as features, so that each first sub-block image obtains a two-dimensional vector, which is recorded as/>; Will said/>And said/>Together, obtain the first sub-block image/>The first feature of (2) is an eighteen-dimensional vector, and the first feature is finally obtained and is marked as/>; Selecting the saidAny second sub-block image/>The second characteristic is obtained by the following steps: the second characteristic is finally obtained in accordance with the acquisition mode of the first characteristic and is marked as/>; Selecting the saidAny first sub-block image/>The third characteristic is obtained by the following steps: calculating the/>, respectively, by using sobel operatorsHorizontal gradient information of/>And vertical gradient information/>：Wherein/>，/>Representing a convolution operation; by means of said/>And said/>Calculate the first gradient magnitude/>And a first gradient direction/>The specific formula is as follows: /(I); According to the first gradient direction/>Finally, the third feature is calculated as/>The method specifically comprises the following steps: ; selecting the/> Any second sub-block image/>The fourth feature is obtained by the following steps: consistent with the acquisition mode of the third characteristic, finally obtaining the second gradient amplitude/>And the fourth feature is:/>; The first similarity index is obtained in the following manner: optionally one of said/>And said/>Corresponding said/>Obtaining the similarity degree/>, of each pair of sub-block images：/>Wherein/>Representing a summation operation; for all said/>Summing to obtain the first similarity index/>：/>; The second similar index is obtained by the following steps: optionally one of said/>And said/>Corresponding saidObtaining the similarity degree/>, of each pair of sub-block images：/>Wherein/>Representing a summation operation,/>Representing a dot product operation; for all said/>Summing to obtain the second similar index：/>; The final similarity between the reference image and the image to be evaluated is obtained by the following steps: the similarity E is the sum of the first similarity index and the second similarity index, and the specific formula is as follows: /(I)。
According to the image quality evaluation method and system based on the gray characteristic, the final similarity between the reference image and the image to be evaluated is calculated, so that the quality of the image can be accurately evaluated, the algorithm is simple, the correlation between pixels is fully considered, and the method and system have high practical value.
Drawings
Fig. 1 is a schematic flow chart of an image quality evaluation method based on gray scale characteristics according to an embodiment of the invention.
Fig. 2 is a schematic diagram of a second flow of the image quality evaluation method based on gray characteristics shown in fig. 1.
Fig. 3 is a flowchart showing a specific procedure for calculating a gray characteristic index in the gray characteristic-based image quality evaluation method shown in fig. 1.
Fig. 4 is a specific flowchart for extracting the first feature in the image quality evaluation method based on gray characteristics shown in fig. 1.
Fig. 5 is a flowchart showing a specific procedure for extracting a third feature in the image quality evaluation method based on gray characteristics shown in fig. 1.
Fig. 6 is a flowchart showing a specific procedure for calculating the first similarity index in the image quality evaluation method based on gray characteristics shown in fig. 1.
Fig. 7 is a specific flowchart of calculating the second similarity index in the image quality evaluation method based on gray characteristics shown in fig. 1.
Fig. 8 is a schematic structural diagram of an image quality evaluation system based on gray scale characteristics according to an embodiment of the present invention.
Detailed Description
The following describes in further detail the embodiments of the present invention with reference to the drawings and examples. The following examples are illustrative of the invention and are not intended to limit the scope of the invention.
As shown in fig. 1 to 7, in the present embodiment, there is provided an image quality evaluation method based on gradation characteristics, including the steps of:
step S1: dividing the reference image and the image to be evaluated into a first sub-block image and a second sub-block image with preset sizes respectively, and marking the first sub-block image and the second sub-block image as respectively And/>Where N represents the number of all sub-blocks after partitioning.
In this embodiment, the reference image and the image to be evaluated are both RGB (optical three primary colors, R represents red, G represents green, and B represents blue) images, which are the best color modes. In this embodiment, the reference image and the image to be evaluated are divided into predetermined sizesA first sub-block image and a second sub-block image of mm. In other embodiments, the preset size may be other values such as 50×50mm, 80×80mm, etc.
Step S2: a gradation characteristic index of each of the first sub-block image and the second sub-block image is calculated.
Specifically, a gradation characteristic index is calculated based on the human gradation characteristics. In a specific application example, any first sub-block image is selectedA specific flowchart for calculating the gradation characteristic index in the image quality evaluation method based on the gradation characteristic shown in fig. 3 will be described. In other embodiments, any of the second sub-block images/>, may also be selectedAn explanation is given. The detailed flow of the step S2 of the present invention includes:
step S21: random at Selecting 10 points, and taking 10 points as the center and the diameter as/>In the field of mm, a gray average value is calculated to obtain 10 gray average values/>Statistics of the variance/>, of the 10 gray-scale averagesThe specific formula is as follows: /(I)Wherein/>。
Step 22: when the degree of difference isAt this time, let the first sub-block image/>Corresponding second sub-block imageGray characteristic index/>; Otherwise/>。
Step S3: the first sub-block image and the second sub-block image are divided into a first category and a second category according to the gray characteristic index.
The first category is specifically: gray scale characteristic indexIs the first sub-block image setAnd/>Corresponding second sub-block image setWherein/>Representing gray characteristic index/>Is a number of sub-blocks of (c).
The second category is specifically: gray scale characteristic indexIs the first sub-block image setAnd/>Corresponding second sub-block image setWherein/>Representing gray characteristic index/>Is a number of sub-blocks of (c).
Step S4: the first feature of each first sub-block image and the second feature of each second sub-block image in the first category are extracted respectively.
First, a first feature of each first sub-block image in the first category is extracted, as shown in a specific flowchart of extracting the first feature in the image quality evaluation method based on gray-scale characteristics shown in fig. 4. In a specific application example, selectAny first sub-block image/>An explanation is given. The detailed flow of the step S4 of the present invention includes:
step S41: will be Dividing into sixteen equal parts, and counting the gray level histogram of the first sub-block image to obtain a vector with ten six dimensions, which is recorded as/>。
Step S42: calculating the mean and variance of the first sub-block images, and storing the mean and variance as features so that each first sub-block image obtains a two-dimensional vector, which is recorded as。
Step S43: will beAnd/>Together, a first sub-block image/>The first feature of (a) is an eighteen-dimensional vector, and finally the first feature is obtained and is recorded as/>。
Next, a second feature of each second sub-block image in the first category is extracted. In a specific application example, selectAny second sub-block image/>An explanation is given. The calculation method of the second feature is identical to the calculation methods of the above-described step S41, step S42, and step S43, and will not be described in detail here. Finally, a second characteristic is obtained, which is marked as/>。
Step S5: the third feature of each first sub-block image and the fourth feature of each second sub-block image in the second category are extracted separately.
First, a third feature of each first sub-block image in the second category is extracted, as in the specific flowchart of extracting the third feature in the image quality evaluation method based on gray-scale characteristics shown in fig. 5. In a specific application example, selectAny first sub-block image/>An explanation is given. The detailed flow of the step S5 of the present invention includes:
step S51: calculation by sobel operator Horizontal gradient information of/>And vertical gradient information/>. The specific formula is as follows: /(I)Wherein/>，/>Representing a convolution operation.
Step S52: by means ofAnd/>Calculate the first gradient magnitude/>And a first gradient direction/>The specific formula is as follows: /(I)。
Step S53: according to the first gradient directionFinally, the third feature is calculated as/>The method specifically comprises the following steps: /(I)。
Next, a fourth feature of each second sub-block image in the second category is extracted. In a specific application example, selectAny second sub-block image/>An explanation is given. The calculation method of the fourth feature is identical to the calculation methods of the above-described step S51, step S52, and step S53, and will not be described in detail here. Finally, the second gradient amplitude/>And the fourth characteristic is:/>。
Step S6: and calculating a first similarity index of the first category according to the first characteristic and the second characteristic.
In the present embodiment, specifically, the gradation characteristic index is calculatedSuch as a specific flowchart for calculating the first similarity index in the image quality evaluation method based on gray characteristics shown in fig. 6. The detailed flow of the step S6 of the present invention includes:
Step S61: optionally one And/>Corresponding/>Obtaining the similarity degree/>, of each pair of sub-block images：/>Wherein/>Representing a summation operation.
Step S62: for all ofSumming to obtain a first similarity index/>：/>。
In step S62, for allSumming, i.e. for gray-scale characteristic index/>Is summed up for all sub-block pairs.
Step S7: and calculating a second similarity index of the second category according to the third characteristic and the fourth characteristic.
In the present embodiment, specifically, the gradation characteristic index is calculatedA specific flowchart for calculating the second similarity index in the image quality evaluation method based on gray characteristics shown in fig. 7. The detailed flow of the step S7 of the present invention includes:
step S71: optionally one And/>Corresponding/>Obtaining the similarity degree/>, of each pair of sub-block images：/>Wherein/>Representing a summation operation,/>Representing a dot product operation.
Step S72: for all ofSumming to obtain a second similarity index/>：/>。
In step S72, for allSumming, i.e. for gray-scale characteristic index/>Is summed up for all sub-block pairs.
Step S8: and calculating the final similarity between the reference image and the image to be evaluated according to the first similarity index and the second similarity index.
The step S8 specifically comprises the following steps: the similarity E is the sum of the first similarity index and the second similarity index, and the specific formula is as follows:。
according to the method, the final similarity between the reference image and the image to be evaluated is calculated, so that the quality of the image can be accurately evaluated, the algorithm is simple, the correlation among pixels is fully considered, and the method has high practical value.
As shown in fig. 8, the present invention further provides an image quality evaluation system based on gray characteristics, which includes a separation module 80, a classification module 81, an extraction module 82, and a calculation module 83.
The separation module 80 is used for performing block processing on the reference image and the image to be evaluated, and dividing the reference image and the image to be evaluated into a first sub-block image and a second sub-block image with preset sizes respectively, which are respectively recorded asAnd/>Where N represents the number of all sub-blocks after partitioning.
In this embodiment, the preset size isMm. In other embodiments, the preset size may be other values such as 50×50mm, 80×80mm, etc.
The calculation module 83 is configured to calculate a gray characteristic index of each of the first sub-block image and the second sub-block image.
Specifically, any first sub-block image is selectedThe gray characteristic index is calculated, and the specific acquisition mode of the gray characteristic index is as follows:
random at Selecting 10 points, and taking 10 points as the center and the diameter as/>In the field of mm, a gray average value is calculated to obtain 10 gray average values/>Statistics of the variance/>, of the 10 gray-scale averagesThe specific formula is as follows: /(I)Wherein/>; When the degree of difference/>At this time, let the first sub-block image/>Corresponding second sub-block image/>Gray characteristic index/>; Otherwise/>。
The classification module 81 is configured to classify the first sub-block image and the second sub-block image into a first class and a second class according to the gray characteristic index.
Specifically, the first category is specifically: gray scale characteristic indexIs a first sub-block image set of (1)And/>Corresponding second sub-block image setWherein/>Representing gray characteristic index/>Is a number of sub-blocks of (c).
Specifically, the second category is specifically: gray scale characteristic indexIs a first sub-block image set of (1)And/>Corresponding second sub-block image setWherein/>Representing gray characteristic index/>The number of sub-blocks of (a);
The extraction module 82 is configured to extract a first feature of each first sub-block image and a second feature of each second sub-block image in the first category, and extract a third feature of each first sub-block image and a fourth feature of each second sub-block image in the second category, respectively.
The specific acquisition modes of the first feature, the second feature, the third feature and the fourth feature are as follows:
Selecting Any first sub-block image/>To illustrate, the first feature is obtained by:
Will be Dividing into sixteen equal parts, and counting the gray level histogram of the first sub-block image to obtain a vector with ten six dimensions, which is recorded as/>; Calculating the mean and variance of the first sub-block images, and storing the mean and variance as features so that each first sub-block image obtains a two-dimensional vector, which is recorded as/>; Will/>And/>Together, obtain a first sub-block imageThe first feature of (a) is an eighteen-dimensional vector, and finally the first feature is obtained and is recorded as/>。
SelectingAny second sub-block image/>To illustrate, the second feature is obtained by:
the second characteristic is finally obtained in accordance with the acquisition mode of the first characteristic and is recorded as 。
SelectingAny first sub-block image/>To illustrate, the third feature is obtained by:
calculation by sobel operator Horizontal gradient information of/>And vertical gradient information：/>Wherein/>，/>Representing a convolution operation; utilization/>And/>Calculate the first gradient magnitude/>And a first gradient direction/>The specific formula is as follows: /(I); According to the first gradient direction/>Finally, the third feature is calculated as/>The method specifically comprises the following steps:。
Selecting Any second sub-block image/>To illustrate, the fourth feature is obtained by:
Consistent with the acquisition mode of the third characteristic, finally obtaining the second gradient amplitude And the fourth feature is:。
The calculating module 83 is further configured to calculate a first similarity index of the first class according to the first feature and the second feature, calculate a second similarity index of the second class according to the third feature and the fourth feature, and calculate a final similarity between the reference image and the image to be evaluated according to the first similarity index and the second similarity index.
The first similarity index is obtained by the following steps:
Optionally one And/>Corresponding/>Obtaining the similarity degree/>, of each pair of sub-block images：/>Wherein/>Representing a summation operation; for all/>Summing to obtain a first similarity index/>：/>。
The second similar index is obtained by the following steps:
Optionally one And/>Corresponding/>Obtaining the similarity degree/>, of each pair of sub-block images：/>Wherein/>Representing a summation operation,/>Representing a dot product operation; for all/>Summing to obtain a second similarity index/>：/>。
The final similarity between the reference image and the image to be evaluated is obtained by the following steps:
the similarity E is the sum of the first similarity index and the second similarity index, and the specific formula is as follows: 。
in this document, the terms "upper", "lower", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", "vertical", "horizontal", etc. refer to the directions or positional relationships based on those shown in the drawings, and are merely for clarity and convenience of description of the expression technical solution, and thus should not be construed as limiting the present invention.
In this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a list of elements is included, and may include other elements not expressly listed.
The foregoing is merely illustrative of the present invention, and the present invention is not limited thereto, and any person skilled in the art will readily recognize that variations or substitutions are within the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims (7)

1. An image quality evaluation method based on gray characteristics, comprising the steps of:
Step S1: dividing the reference image and the image to be evaluated into a first sub-block image and a second sub-block image with preset sizes respectively, wherein the first sub-block image and the second sub-block image are respectively marked as { A n (x, y) |n=1, …, N } and { B n (x, y) |n=1, …, N }, wherein N represents the number of all sub-blocks after the segmentation;
Step S2: calculating gray characteristic indexes of each first sub-block image and each second sub-block image;
Step S3: dividing the first sub-block image and the second sub-block image into a first category and a second category according to the gray characteristic index;
Step S4: respectively extracting a first characteristic of each first sub-block image and a second characteristic of each second sub-block image in the first category;
step S5: extracting a third feature of each first sub-block image and a fourth feature of each second sub-block image in the second category respectively;
step S6: according to the first characteristic and the second characteristic, calculating to obtain a first similarity index of the first category;
Step S7: according to the third characteristic and the fourth characteristic, calculating to obtain a second similar index of the second category;
step S8: calculating the final similarity of the reference image and the image to be evaluated according to the first similarity index and the second similarity index;
The step S6 comprises the following steps:
Step S61: optionally, one of the A1 n1 (x, y) and the B1 n1 (x, y) corresponding to the A1 n1 (x, y) result in a similarity degree E1 n1 for each pair of sub-block images:
Where sum represents the sum operation;
Step S62: summing all the E1 n1 to obtain the first similarity index E1: ；
The A1 n1 (x, y) is any one of a first sub-block image set { A1 n (x, y) |n=1, …, N1} of the first sub-block image set having the gray characteristic index lable =1;
Any one of a second sub-block image set { B1 n (x, y) |n=1, …, N1} of the second sub-block image set of B1 n1 (x, y) gray characteristic index lable =1;
VA1 n1 is a first feature and VB1 n1 is a second feature;
the step S7 includes:
step S71: optionally, one of the A2 n2 (x, y) and the B2 n2 (x, y) corresponding to the A2 n2 (x, y) result in a similarity degree E2 n2 for each pair of sub-block images:
where sum represents a summation operation, and · x represents a dot product operation;
Step S72: summing all the E2 n2 to obtain the second similarity index E2: ；
The A2 n2 (x, y) is any one of a first sub-block image set { A2 n (x, y) |n=1, …, N2} of the first sub-block image set having the gray characteristic index lable =2;
Any one of a second sub-block image set { B2 n (x, y) |n=1, …, N2} of the second sub-block image set of B2 n2 (x, y) gray characteristic index lable =2;
GA2 n2 (x, y) is a first gradient amplitude, GB2 n2 (x, y) is a second gradient amplitude, PA2 n2 is a third feature, PB2 n2 is a fourth feature;
The step S8 specifically includes: the similarity E' is the sum of the first similarity index and the second similarity index, and the specific formula is as follows: e' =e1+e2.
2. The gray characteristic-based image quality assessment method according to claim 1, wherein any one of said first sub-block images a n1 (x, y) is selected, said step S2 comprising:
Step S21: 10 points are randomly selected on A n1 (x, y), gray average values are calculated in the field with the diameter of 7mm by taking the 10 points as centers, 10 gray average values { alpha i |i=1, …,10} are obtained, and the difference E of the 10 gray average values is counted, wherein the specific formula is as follows: Wherein/>
Step 22: when the difference E <15, let the gray characteristic index lable =1 of the first sub-block image a n1 (x, y) and the corresponding second sub-block image B n1 (x, y); otherwise lable =2.
3. The image quality evaluation method based on gradation characteristics according to claim 2, wherein in said step S3,
The first category is specifically: a first sub-block image set { A1 n (x, y) |n=1, …, N1} of the gray characteristic index lable =1 and a second sub-block image set { B1 n (x, y) |n=1, …, N1} corresponding to the { A1 n (x, y) |n=1, …, N1} where N1 represents the number of sub-blocks of the gray characteristic index lable =1;
The second category is specifically: the first sub-block image set { A2 n (x, y) |n=1, …, N2} of the gray characteristic index lable =2 and the second sub-block image set { B2 n (x, y) |n=1, …, N2} corresponding to the { A2 n (x, y) |n=1, …, N2} where N2 represents the number of sub-blocks of the gray characteristic index lable =2.
4. The image quality evaluation method based on gray scale characteristics according to claim 3, wherein selecting any one of the first sub-block images A1 n1 (x, y) of { A1 n (x, y) |n=1, …, N1}, the step S4 comprises:
Step S41: dividing the A1 n1 (x, y) into sixteen equal parts, and counting the gray level histogram of the first sub-block image to obtain a sixteen-dimensional vector which is marked as v1 n1;
Step S42: calculating the mean value and variance of the first sub-block images, and storing the mean value and variance as features, so that each first sub-block image obtains a two-dimensional vector, and the vector is marked as v2 n1;
Step S43: combining the v1 n1 and the v2 n1 together to obtain a vector with a first feature of the first sub-block image A1 n1 (x, y) being eighteen dimensions, and finally obtaining the first feature, which is denoted as VA1 n1;
Selecting any second sub-block image B1 n1 (x, y) in { B1 n (x, y) |n=1, …, N1}, wherein the calculation mode of the second feature is consistent with the calculation modes of the step S41, the step S42, and the step S43, and finally obtaining the second feature, which is marked as VB1 n1.
5. The gray-scale-characteristic-based image quality evaluation method according to claim 4, wherein selecting any one of the first sub-block images A2 n2 (x, y) of { A2 n (x, y) |n=1, …, N2}, the step S5 includes:
Step S51: the horizontal gradient information G 1 and the vertical gradient information G 2 of the A2 n2 (x, y) are calculated by sobel operator, respectively: Wherein/> Representing a convolution operation;
Step S52: first gradient magnitude GA2 n2 (x, y) and first gradient direction QA2 n2 (x, y) are calculated using the G 1 and the G 2 as follows:
Step S53: according to the first gradient direction QA2 n2 (x, y), the third feature is finally calculated as PA2 n2 (x, y), in particular:
Selecting any second sub-block image B2 n2 (x, y) in { B2 n (x, y) |n=1, …, N2}, wherein the calculation mode of the fourth feature is consistent with the calculation modes of the step S51, the step S52, and the step S53, and finally obtaining a second gradient amplitude GB2 n2 (x, y) and the fourth feature is PB2 n2 (x, y).
6. An image quality evaluation system based on gray characteristics is characterized by comprising a separation module, a classification module, an extraction module and a calculation module,
The separation module is used for performing blocking processing on the reference image and the image to be evaluated, and dividing the reference image and the image to be evaluated into a first sub-block image and a second sub-block image with preset sizes respectively, wherein the first sub-block image and the second sub-block image are respectively marked as { A n (x, y) |n=1, …, N } and { B n (x, y) |n=1, …, N }, and N represents the number of all sub-blocks after blocking;
the calculation module is used for calculating gray characteristic indexes of each first sub-block image and each second sub-block image;
The classification module is used for classifying the first sub-block image and the second sub-block image into a first category and a second category according to the gray characteristic index;
the extraction module is used for respectively extracting a first feature of each first sub-block image and a second feature of each second sub-block image in the first category, and respectively extracting a third feature of each first sub-block image and a fourth feature of each second sub-block image in the second category;
The calculation module is further configured to calculate a first similarity index of the first class according to the first feature and the second feature, calculate a second similarity index of the second class according to the third feature and the fourth feature, and calculate a final similarity between the reference image and the image to be evaluated according to the first similarity index and the second similarity index;
the first similarity index is obtained in the following manner: optionally, one of the A1 n1 (x, y) and the B1 n1 (x, y) corresponding to the A1 n1 (x, y) result in a similarity degree E1 n1 for each pair of sub-block images:
where sum represents the sum operation; summing all the E1 n1 to obtain the first similarity index E1: /(I) ；
The A1 n1 (x, y) is any one of a first sub-block image set { A1 n (x, y) |n=1, …, N1} of the first sub-block image set having the gray characteristic index lable =1;
Any one of a second sub-block image set { B1 n (x, y) |n=1, …, N1} of the second sub-block image set of B1 n1 (x, y) gray characteristic index lable =1;
VA1 n1 is a first feature and VB1 n1 is a second feature;
The second similar index is obtained by the following steps: optionally, one of the A2 n2 (x, y) and the B2 n2 (x, y) corresponding to the A2 n2 (x, y) result in a similarity degree E2 n2 for each pair of sub-block images:
where sum represents a summation operation, and · x represents a dot product operation; summing all the E2 n2 to obtain the second similarity index E2: /(I) ；
The A2 n2 (x, y) is any one of a first sub-block image set { A2 n (x, y) |n=1, …, N2} of the first sub-block image set having the gray characteristic index lable =2;
Any one of a second sub-block image set { B2 n (x, y) |n=1, …, N2} of the second sub-block image set of B2 n2 (x, y) gray characteristic index lable =2;
GA2 n2 (x, y) is a first gradient amplitude, GB2 n2 (x, y) is a second gradient amplitude, PA2 n2 is a third feature, PB2 n2 is a fourth feature;
The final similarity between the reference image and the image to be evaluated is obtained by the following steps: the similarity E' is the sum of the first similarity index and the second similarity index, and the specific formula is as follows: e' =e1+e2.
7. The gray-scale-characteristic-based image quality evaluation system according to claim 6, wherein any one of the first sub-block images a n1 (x, y) is selected, and the gray-scale-characteristic index is obtained by: 10 points are randomly selected on A n1 (x, y), gray average values are calculated in the field with the diameter of 7mm by taking the 10 points as centers, 10 gray average values { alpha i |i=1, …,10} are obtained, and the difference E of the 10 gray average values is counted, wherein the specific formula is as follows: Wherein the method comprises the steps of
When the difference E <15, let the gray characteristic index lable =1 of the first sub-block image a n1 (x, y) and the corresponding second sub-block image B n1 (x, y); otherwise lable = 2;
The first category is specifically: a first sub-block image set { A1 n (x, y) |n=1, …, N1} of the gray characteristic index lable =1 and a second sub-block image set { B1 n (x, y) |n=1, …, N1} corresponding to the { A1 n (x, y) |n=1, …, N1} where N1 represents the number of sub-blocks of the gray characteristic index lable =1;
The second category is specifically: a first sub-block image set { A2 n (x, y) |n=1, …, N2} of the gray characteristic index lable =2 and a second sub-block image set { B2 n (x, y) |n=1, …, N2} corresponding to the { A2 n (x, y) |n=1, …, N2} where N2 represents the number of sub-blocks of the gray characteristic index lable =2;
Selecting any first sub-block image A1 n1 (x, y) in { A1 n (x, y) |n=1, …, N1}, where the first feature is obtained by: dividing the A1 n1 (x, y) into sixteen equal parts, and counting the gray level histogram of the first sub-block image to obtain a sixteen-dimensional vector which is marked as v1 n1;
calculating the mean value and variance of the first sub-block images, and storing the mean value and variance as features, so that each first sub-block image obtains a two-dimensional vector, and the vector is marked as v2 n1;
Combining the v1 n1 and the v2 n1 together to obtain a vector with a first feature of the first sub-block image A1 n1 (x, y) being eighteen dimensions, and finally obtaining the first feature, which is denoted as VA1 n1;
Selecting any second sub-block image B1 n1 (x, y) in { B1 n (x, y) |n=1, …, N1}, where the second feature is obtained by: the second characteristic is finally obtained in accordance with the acquisition mode of the first characteristic and is marked as VB1 n1;
Selecting any first sub-block image A2 n2 (x, y) of { A2 n (x, y) |n=1, …, N2}, where the third feature is obtained by: the horizontal gradient information G 1 and the vertical gradient information G 2 of the A2 n2 (x, y) are calculated by sobel operator, respectively: Wherein/> Representing a convolution operation;
First gradient magnitude GA2 n2 (x, y) and first gradient direction QA2 n2 (x, y) are calculated using the G 1 and the G 2 as follows:
According to the first gradient direction QA2 n2 (x, y), the third feature is finally calculated as PA2 n2 (x, y), in particular:
Selecting any second sub-block image B2 n2 (x, y) in { B2 n (x, y) |n=1, …, N2}, where the fourth feature is obtained by: and finally obtaining a second gradient amplitude GB2 n2 (x, y) and the fourth characteristic PB2 n2 (x, y) according to the obtaining mode of the third characteristic.
CN202011314950.3A 2020-11-20 2020-11-20 Image quality evaluation method and system based on gray scale characteristics Active CN112330657B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011314950.3A CN112330657B (en) 2020-11-20 2020-11-20 Image quality evaluation method and system based on gray scale characteristics

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011314950.3A CN112330657B (en) 2020-11-20 2020-11-20 Image quality evaluation method and system based on gray scale characteristics

Publications (2)

Publication Number Publication Date
CN112330657A CN112330657A (en) 2021-02-05
CN112330657B true CN112330657B (en) 2024-06-07

Family

ID=74322035

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011314950.3A Active CN112330657B (en) 2020-11-20 2020-11-20 Image quality evaluation method and system based on gray scale characteristics

Country Status (1)

Country Link
CN (1) CN112330657B (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106709958A (en) * 2016-12-03 2017-05-24 浙江大学 Gray scale gradient and color histogram-based image quality evaluation method
CN108053393A (en) * 2017-12-08 2018-05-18 广东工业大学 A kind of gradient similarity graph image quality evaluation method and device
CN109325550A (en) * 2018-11-02 2019-02-12 武汉大学 Non-reference picture quality appraisement method based on image entropy
CN111507426A (en) * 2020-04-30 2020-08-07 中国电子科技集团公司第三十八研究所 No-reference image quality grading evaluation method and device based on visual fusion characteristics
CN111598837A (en) * 2020-04-21 2020-08-28 中山大学 Full-reference image quality evaluation method and system suitable for visual two-dimensional code

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4006224B2 (en) * 2001-11-16 2007-11-14 キヤノン株式会社 Image quality determination method, determination device, and determination program

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106709958A (en) * 2016-12-03 2017-05-24 浙江大学 Gray scale gradient and color histogram-based image quality evaluation method
CN108053393A (en) * 2017-12-08 2018-05-18 广东工业大学 A kind of gradient similarity graph image quality evaluation method and device
CN109325550A (en) * 2018-11-02 2019-02-12 武汉大学 Non-reference picture quality appraisement method based on image entropy
CN111598837A (en) * 2020-04-21 2020-08-28 中山大学 Full-reference image quality evaluation method and system suitable for visual two-dimensional code
CN111507426A (en) * 2020-04-30 2020-08-07 中国电子科技集团公司第三十八研究所 No-reference image quality grading evaluation method and device based on visual fusion characteristics

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
Color-based structural similarity image quality assessment;Mohammed Ahmed Hassan, Mazen Sheikh Bashraheel;《2017 8th International Conference on Information Technology (ICIT)》;20170517;691-696 *
基于扩展梯度算子的结构相似度图像质量评价方法;邓杰航 等;《科学技术与工程》;20180928;第18卷(第27期);42-47 *
基于视觉感知的无参考图像质量评价;王盛春;《中国优秀硕士学位论文全文数据库 信息科技辑》;20180115(第01期);1-70 *

Also Published As

Publication number Publication date
CN112330657A (en) 2021-02-05

Similar Documents

Publication Publication Date Title
CN108596197B (en) Seal matching method and device
CN104504669B (en) A kind of medium filtering detection method based on local binary patterns
CN110059728B (en) RGB-D image visual saliency detection method based on attention model
CN109948566B (en) Double-flow face anti-fraud detection method based on weight fusion and feature selection
KR20120047853A (en) Histogram methods and systems for object recognition
CN110210492B (en) Stereo image visual saliency detection method based on deep learning
CN109740572A (en) A kind of human face in-vivo detection method based on partial color textural characteristics
CN110490924B (en) Light field image feature point detection method based on multi-scale Harris
CN104537376A (en) A method, a relevant device, and a system for identifying a station caption
CN112184672A (en) No-reference image quality evaluation method and system
Riana et al. Extraction and classification texture of inflammatory cells and nuclei in normal pap smear images
CN111709305A (en) Face age identification method based on local image block
CN111047618A (en) Multi-scale-based non-reference screen content image quality evaluation method
CN108682005B (en) Semi-reference 3D synthetic image quality evaluation method based on covariance matrix characteristics
CN112330657B (en) Image quality evaluation method and system based on gray scale characteristics
CN104394405B (en) A kind of method for evaluating objective quality based on full reference picture
CN112215266B (en) X-ray image contraband detection method based on small sample learning
CN112070714B (en) Method for detecting flip image based on local ternary counting feature
KR101912433B1 (en) Apparatus and method for detecting fake fingerprint
CN112396638A (en) Image processing method, terminal and computer readable storage medium
CN111027573A (en) Image authenticity identification method based on blind evidence obtaining technology
CN115690934A (en) Master and student attendance card punching method and device based on batch face recognition
CN111402189B (en) Video image color cast detection device and method
Hong et al. Saliency-based feature learning for no-reference image quality assessment
CN109886276B (en) Half-word judging method for dial rolling digital character

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