CN114724188B - Vein identification method and device based on gray level co-occurrence matrix - Google Patents

Vein identification method and device based on gray level co-occurrence matrix Download PDF

Info

Publication number
CN114724188B
CN114724188B CN202210561152.3A CN202210561152A CN114724188B CN 114724188 B CN114724188 B CN 114724188B CN 202210561152 A CN202210561152 A CN 202210561152A CN 114724188 B CN114724188 B CN 114724188B
Authority
CN
China
Prior art keywords
image
matrix
formula
vein
gray level
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
CN202210561152.3A
Other languages
Chinese (zh)
Other versions
CN114724188A (en
Inventor
高旭
赵国栋
李学双
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.)
Beijing Shengdian Cloud Information Technology Co ltd
Original Assignee
Beijing Shengdian Cloud Information 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 Beijing Shengdian Cloud Information Technology Co ltd filed Critical Beijing Shengdian Cloud Information Technology Co ltd
Priority to CN202210561152.3A priority Critical patent/CN114724188B/en
Publication of CN114724188A publication Critical patent/CN114724188A/en
Application granted granted Critical
Publication of CN114724188B publication Critical patent/CN114724188B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • 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
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T3/00—Geometric image transformations in the plane of the image
    • G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
    • G06T3/4007—Scaling of whole images or parts thereof, e.g. expanding or contracting based on interpolation, e.g. bilinear interpolation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/10—Image enhancement or restoration using non-spatial domain filtering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/20—Image enhancement or restoration using local operators
    • 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
    • G06T5/94—Dynamic range modification of images or parts thereof based on local image properties, e.g. for local contrast enhancement
    • 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
    • G06T2207/20048—Transform domain processing
    • G06T2207/20056—Discrete and fast Fourier transform, [DFT, FFT]

Landscapes

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

Abstract

The invention discloses a vein identification method and a vein identification device based on a gray level co-occurrence matrix, which belong to the technical field of vein identification and processing and comprise the following steps: 1) carrying out scale normalization pretreatment on the vein image; 2) filtering the vein image after the normalization pretreatment by using homomorphic filtering; 3) enhancing the filtered vein image by using a local contrast enhancement method; 4) and identifying the enhanced vein image based on the gray level co-occurrence matrix. According to the invention, the gray level co-occurrence matrix is introduced to carry out vein image matching identification, so that the calculated amount is reduced, and the image identification speed is obviously improved; the gray level correction method based on the combination of the improved homomorphic filtering and the local contrast enhancement improves the influence of the illumination nonuniformity on the vein image, not only can keep the low-frequency component, but also can enhance the high-frequency component, removes redundant noise, enhances the dark details of the vein image, and improves the recognition rate of vein recognition.

Description

Vein identification method and device based on gray level co-occurrence matrix
Technical Field
The invention relates to the technical field of vein identification and processing, in particular to a vein identification method and device based on a gray level co-occurrence matrix.
Background
Vein recognition is a new infrared biological recognition technology, which is to use an infrared camera to shoot the distribution diagram of veins in vivo (back of the hand, back of the finger, abdomen of the finger, palm and wrist) according to the characteristic that hemoglobin in the blood of veins in the human body absorbs near infrared or radiates far infrared rays of the human body.
The gray level correction is an image detail enhancement method, is used for enhancing a magnetic resonance image in the early stage, has a good enhancement effect on the image quality, and is also applied to vein recognition at present to realize detail enhancement of the vein image. For example, chinese patent CN112560808B discloses a method and a device for in vivo vein recognition based on gray scale information, wherein the vein recognition method includes the following steps: 1) carrying out scale normalization pretreatment on the in-vivo vein image; 2) adopting a gray level correction method combining regional variance transformation and single-scale Retinex to perform enhancement processing on the vein image after normalization processing; 3) carrying out rough image matching based on an improved gray difference curved surface method; 4) and carrying out fine image matching based on a correlation coefficient method.
When the vein image is enhanced by the vein identification, the vein image is filtered by a Gaussian filter, and after the vein image is filtered by the traditional Gaussian filter, a lot of low-frequency information is lost in the image, so that a smooth area of the image basically disappears, and the identification rate is influenced; when the filtered vein image is subjected to enhancement processing, some high-frequency parts may have an over-enhancement phenomenon, so that noise enhancement is caused, and the feature extraction is inaccurate; in addition, the conventional vein recognition method has a problem of large calculation amount, and the image recognition speed is affected.
Disclosure of Invention
The invention aims to provide a vein identification method and device based on a gray level co-occurrence matrix, and aims to solve the problems that in the prior art, the vein image enhancement effect is poor, and the identification effect is poor due to inaccurate feature extraction.
In order to achieve the purpose, the technical scheme provided by the invention is as follows:
the invention relates to a vein identification method based on a gray level co-occurrence matrix, which comprises the following steps:
1) carrying out scale normalization pretreatment on the vein image;
2) filtering the vein image after the normalization pretreatment by using homomorphic filtering;
3) enhancing the filtered vein image by using a local contrast enhancement method;
4) and identifying the enhanced vein image based on the gray level co-occurrence matrix.
Preferably, in the step 1), a bilinear interpolation method is adopted to perform scale normalization processing on the vein image.
Preferably, the specific way of performing the filtering processing on the vein image after the normalization preprocessing by using the homomorphic filtering in the step 2) is as follows:
2.1) vein image after scale normalization processing
Figure 735086DEST_PATH_IMAGE001
Carrying out logarithmic transformation, and leading:
Figure 312698DEST_PATH_IMAGE002
in the formula,
Figure 392912DEST_PATH_IMAGE003
is a vein image after the logarithmic transformation,xandyrespectively a row coordinate and a column coordinate of the vein image matrix;
2.2) taking Fourier transform on two sides of the above formula, and enabling:
Figure 30566DEST_PATH_IMAGE004
in the formula,
Figure 898028DEST_PATH_IMAGE005
for the logarithmically transformed vein image
Figure 154697DEST_PATH_IMAGE003
The fourier transformed image of (a) is,uandvrespectively x and y are variables obtained by Fourier transform,
Figure 587953DEST_PATH_IMAGE006
is Fourier transform;
2.3) passing Gaussian filter
Figure 912622DEST_PATH_IMAGE007
Filtering the above formula to obtain a filtered frequency domain image
Figure 673904DEST_PATH_IMAGE008
The calculation formula is as follows:
Figure 593319DEST_PATH_IMAGE009
wherein, the Gaussian filter
Figure 881081DEST_PATH_IMAGE007
The calculation formula of (2) is as follows:
Figure 860538DEST_PATH_IMAGE010
in the formula:
Figure 843538DEST_PATH_IMAGE011
is a high-frequency amplification factor, has a value range of more than 1 and less than 2,
Figure 68108DEST_PATH_IMAGE012
the offset is a parameter with the value range of more than 0 and less than 1 and controlling the slope of the filter
Figure 210376DEST_PATH_IMAGE014
Is the center of the filter and is,
Figure 360735DEST_PATH_IMAGE015
is the mid-point in the frequency domain, is the high frequency gain,
Figure 831030DEST_PATH_IMAGE016
in order to gain at a low frequency, the gain is,cfor the sharpening constant, the value range is
Figure 357827DEST_PATH_IMAGE017
；
2.4) carrying out inverse Fourier transform on the filtered vein image to obtain an image with a frequency domain converted into a space domain
Figure 853137DEST_PATH_IMAGE018
That is, the final filtered image is obtained, and the formula of the inverse fourier transform is:
Figure 174397DEST_PATH_IMAGE019
in the formula,
Figure 131988DEST_PATH_IMAGE020
representing the inverse fourier transform.
Preferably, the step 3) of performing enhancement processing on the filtered vein image by using a local contrast enhancement method specifically includes:
3.1) filtering the processed image
Figure 462476DEST_PATH_IMAGE018
Performing exponential operation to obtain homomorphic filtered image
Figure 579336DEST_PATH_IMAGE021
：
Figure 946864DEST_PATH_IMAGE022
3.2) setting
Figure 17850DEST_PATH_IMAGE023
Is an image
Figure 886449DEST_PATH_IMAGE021
The gray value of a certain point in (b),iandjrespectively being said images
Figure 592237DEST_PATH_IMAGE021
Row and column coordinates of the matrix, and
Figure 396245DEST_PATH_IMAGE024
as a center, the window size is (2)n+1)*(2nThe region of +1) is a local region, whereinnCalculating a local average value for a positive integer
Figure 187483DEST_PATH_IMAGE025
And local variance
Figure 364167DEST_PATH_IMAGE026
，
Local mean value
Figure 924462DEST_PATH_IMAGE025
The calculation formula of (c) is:
Figure 899371DEST_PATH_IMAGE027
the calculation formula of the local variance is as follows:
Figure 177906DEST_PATH_IMAGE028
in the formula,
Figure 388307DEST_PATH_IMAGE025
is the average of the local areas and is,
Figure 570152DEST_PATH_IMAGE029
the local standard deviation is taken as the local standard deviation,aandbrespectively are horizontal and vertical coordinates in a local area;
3.3) is provided with
Figure 981542DEST_PATH_IMAGE030
Is composed of
Figure 747372DEST_PATH_IMAGE023
Corresponding enhanced pixel value, enhanced pixelThe calculation formula of the value is:
Figure 761465DEST_PATH_IMAGE031
in the formula,
Figure 171718DEST_PATH_IMAGE032
is a gain constant, satisfies
Figure 613063DEST_PATH_IMAGE033
And order:
Figure 364725DEST_PATH_IMAGE034
then the
Figure 916929DEST_PATH_IMAGE035
In the formula,Cis a constant value with a value range of more than 0 and less than 1,
Figure 447268DEST_PATH_IMAGE036
is the global mean square error.
Preferably, the step 4) of identifying the enhanced vein image based on the gray level co-occurrence matrix specifically includes:
4.1) carrying out feature compression on the image;
4.2) solving the gray level co-occurrence matrix
Figure 59515DEST_PATH_IMAGE037
Gray level co-occurrence matrix
Figure 534358DEST_PATH_IMAGE037
Is shown as
Figure 31199DEST_PATH_IMAGE038
In the direction and at a distance
Figure 42142DEST_PATH_IMAGE039
Has a gray value respectively
Figure 825291DEST_PATH_IMAGE040
And
Figure 787430DEST_PATH_IMAGE041
probability of occurrence, then pixel pair
Figure 556803DEST_PATH_IMAGE042
And
Figure 451947DEST_PATH_IMAGE043
gray level co-occurrence matrix in four directions
Figure 898672DEST_PATH_IMAGE037
The calculation formula of (2) is as follows:
Figure 223474DEST_PATH_IMAGE044
wherein,
Figure 921172DEST_PATH_IMAGE045
；
in the formula,
Figure 405243DEST_PATH_IMAGE046
is a point
Figure 530194DEST_PATH_IMAGE047
Is determined by the gray-scale value of (a),
Figure 342292DEST_PATH_IMAGE048
is a point
Figure 345145DEST_PATH_IMAGE049
Is determined by the gray-scale value of (a),Lis a gray scale level of the image,
Figure 683723DEST_PATH_IMAGE050
and
Figure 854941DEST_PATH_IMAGE051
representing the number of lines and rows of the vein image;
4.3) co-occurrence matrix according to gray level
Figure 278969DEST_PATH_IMAGE037
Taking a distance
Figure 318469DEST_PATH_IMAGE039
Is 1, angle
Figure 10088DEST_PATH_IMAGE038
0 degree, 45 degrees, 90 degrees and 135 degrees respectively, and then the gray level co-occurrence matrix is aligned
Figure 352208DEST_PATH_IMAGE037
And (4) carrying out normalization, wherein the calculation formula is as follows:
Figure 263532DEST_PATH_IMAGE052
in the formula,
Figure 372303DEST_PATH_IMAGE053
in order to normalize the co-occurrence matrix,
Figure 295259DEST_PATH_IMAGE054
for the normalized constant, the calculation formula is:
Figure 932914DEST_PATH_IMAGE055
in the formula,Mis the size of the co-occurrence matrix, typically 8, 16, 32;
4.4) obtaining each texture characteristic, namely, the gray level co-occurrence matrix
Figure 832999DEST_PATH_IMAGE037
4 texture parameters such as energy, entropy, moment of inertia and correlation are calculated, wherein the energy ASMThe calculation formula is as follows:
Figure 214302DEST_PATH_IMAGE056
entropy of the entropy
Figure 257344DEST_PATH_IMAGE057
The calculation formula of (2) is as follows:
Figure 65900DEST_PATH_IMAGE058
moment of inertia
Figure 217396DEST_PATH_IMAGE059
The calculation formula of (2) is as follows:
Figure 12177DEST_PATH_IMAGE060
correlation
Figure 538754DEST_PATH_IMAGE061
The calculation formula of (2) is as follows:
Figure 783790DEST_PATH_IMAGE062
in the formula,
Figure 891424DEST_PATH_IMAGE063
is the mean value of the gray levels,
Figure 489895DEST_PATH_IMAGE064
in order to smooth the average value of the average,
Figure 632163DEST_PATH_IMAGE065
as a standard deviation of the gray scale,
Figure 549566DEST_PATH_IMAGE066
in order to smooth out the standard deviation of the standard deviation,
Figure 19862DEST_PATH_IMAGE063
、
Figure 546658DEST_PATH_IMAGE064
、
Figure 543433DEST_PATH_IMAGE065
、
Figure 864693DEST_PATH_IMAGE066
the calculation methods of (A) are respectively as follows:
Figure 179874DEST_PATH_IMAGE067
connecting the mean values and standard deviations of the energy, entropy, moment of inertia and correlation 4 texture parameters on four angles in series to serve as a template characteristic matrix of the image;
4.5) comparing the characteristics, namely subtracting the characteristic matrix to be matched from the characteristic matrix of the template to obtain a difference matrix A, and then calculating the Euclidean norm of the matrix A, wherein the norm of the matrix A can be realized by solving the maximum singular value of the matrix, and the calculation formula is as follows:
Figure 120148DEST_PATH_IMAGE068
in the formula,
Figure 971430DEST_PATH_IMAGE069
representing the singular value of the difference matrix A, and T represents the maximum singular value;
4.6) selecting a threshold value by using the size of the maximum singular value T, and carrying out image identification according to the threshold value.
Preferably, the characteristic compression of the image in the step 4.1) is to compress the gray level of the original image, i.e. to quantize the gray level 256 to 16 levels and convert the gray level 0-255 to the gray level 0-15.
The invention also relates to a vein recognition device based on the gray level co-occurrence matrix, which comprises the following components:
1) the normalization preprocessing module is used for carrying out scale normalization preprocessing on the vein image;
2) the filtering processing module is used for carrying out filtering processing on the vein image after the normalization preprocessing by using homomorphic filtering;
3) the enhancement processing module is used for enhancing the filtered vein image by using a local contrast enhancement method;
4) and the image identification module is used for identifying the enhanced vein image based on the gray level co-occurrence matrix.
Compared with the prior art, the technical scheme provided by the invention has the following beneficial effects:
1. the vein identification method based on the gray level co-occurrence matrix introduces the gray level co-occurrence matrix to carry out vein image matching identification, in the identification process, firstly, the image is subjected to feature compression, the gray level co-occurrence matrix of the compressed image is calculated, then, the mean value and the standard deviation of 4 texture parameters are connected in series to serve as the feature matrix of the image, then, the template feature and the feature to be matched are subtracted, then, the maximum singular value of the template feature is calculated, finally, the appropriate threshold value is selected by utilizing the size of the maximum singular value, whether veins belong to the same class or not is judged according to the threshold value, the calculated amount is reduced, and the image identification speed is obviously improved.
2. The vein identification method based on the gray level co-occurrence matrix provides a new gray level correction calculation method for carrying out filtering processing on the vein image, namely the filtering processing is mainly carried out based on improved homomorphic filtering.
3. The vein identification method based on the gray level co-occurrence matrix provided by the invention can avoid the oscillation generated at the position where the gray level of the output image is changed violently, avoid the enhancement of noise, and further contribute to improving the identification rate of vein identification.
Drawings
FIG. 1 is a flow chart of a vein identification method based on a gray level co-occurrence matrix;
fig. 2 is a structural framework diagram of vein recognition based on a gray level co-occurrence matrix.
Detailed Description
For further understanding of the present invention, the present invention will be described in detail with reference to examples, which are provided for illustration of the present invention but are not intended to limit the scope of the present invention.
Example 1
Referring to fig. 1, the present invention relates to a vein identification method based on gray level co-occurrence matrix, which comprises the following steps:
1) carrying out scale normalization processing on the vein image by adopting a bilinear interpolation method;
2) filtering the vein image after the normalization pretreatment by using homomorphic filtering, which comprises the following specific steps:
2.1) vein image after normalization processing
Figure 729170DEST_PATH_IMAGE001
Carrying out logarithmic transformation, and leading:
Figure 298692DEST_PATH_IMAGE070
in the formula,
Figure 777078DEST_PATH_IMAGE003
is a vein image after the logarithmic transformation,xandyrespectively a row coordinate and a column coordinate of the vein image matrix;
2.2) taking Fourier transform on two sides of the above formula, and enabling:
Figure 249910DEST_PATH_IMAGE071
in the formula,
Figure 912972DEST_PATH_IMAGE005
is composed of
Figure 969790DEST_PATH_IMAGE003
The fourier transformed image of (a) is,uandvare respectively asx，yThe variables obtained by the Fourier transform are used as the variables,
Figure 517446DEST_PATH_IMAGE006
is Fourier transform;
2.3) passing Gaussian filter
Figure 343319DEST_PATH_IMAGE007
Filtering the above formula to obtain a filtered frequency domain image
Figure 958976DEST_PATH_IMAGE008
The calculation formula is as follows:
Figure 112876DEST_PATH_IMAGE072
conventional gaussian filter
Figure 57699DEST_PATH_IMAGE007
The calculation formula is as follows:
Figure 3658DEST_PATH_IMAGE073
in the formula:
Figure 8523DEST_PATH_IMAGE074
c is a constant value for the cut-off frequency,
Figure 915299DEST_PATH_IMAGE075
and
Figure 165277DEST_PATH_IMAGE016
respectively a high-frequency and a low-frequency gain,
Figure 965743DEST_PATH_IMAGE076
is a point
Figure 407089DEST_PATH_IMAGE077
To the origin of the frequency plane
Figure 535582DEST_PATH_IMAGE078
The calculation formula of (c) is:
Figure 87786DEST_PATH_IMAGE079
since the image after the traditional gaussian high-pass filtering loses much low-frequency information, the smooth area of the image basically disappears, and in order to adapt to the vein image, the gaussian filter in the embodiment
Figure 975714DEST_PATH_IMAGE007
The calculation formula of (2) is as follows:
Figure 587961DEST_PATH_IMAGE080
in the formula:
Figure 203750DEST_PATH_IMAGE081
is a high-frequency amplification factor, has a value range of more than 1 and less than 2,
Figure 559645DEST_PATH_IMAGE082
the offset is a parameter with the value range of more than 0 and less than 1 and controlling the slope of the filter
Figure 69124DEST_PATH_IMAGE083
Is the center of the filter and is,
Figure 88158DEST_PATH_IMAGE015
is the mid-point in the frequency domain,
Figure 191243DEST_PATH_IMAGE084
in order to gain the gain at high frequency,
Figure 85249DEST_PATH_IMAGE016
in order to gain the gain at low frequencies,cfor the sharpening constant, the value range is
Figure 714814DEST_PATH_IMAGE085
；
Compared with the traditional Gaussian filter, after the filtering processing is carried out by adopting the calculation method of the Gaussian filter, not only can the low-frequency component be reserved, but also the high-frequency component can be enhanced, so that the details of the dark part of the vein image are enhanced;
2.4) carrying out inverse Fourier transform on the filtered vein image to obtain an image with a frequency domain converted into a space domain
Figure 403284DEST_PATH_IMAGE018
That is, the final filtered image is obtained, and the formula of the inverse fourier transform is as follows:
Figure 993666DEST_PATH_IMAGE019
in the formula,
Figure 195758DEST_PATH_IMAGE086
representing an inverse fourier transform;
3) the method for enhancing the vein image after the filtering processing by using the local contrast enhancement method comprises the following specific steps:
3.1) filtering the processed image
Figure 414249DEST_PATH_IMAGE018
Performing exponential operation to obtain homomorphic filtered image
Figure 680146DEST_PATH_IMAGE021
：
Figure 616878DEST_PATH_IMAGE087
3.2) setting
Figure 118266DEST_PATH_IMAGE023
Is an image
Figure 958308DEST_PATH_IMAGE021
The gray value of a certain point in (b),iandjrespectively being said images
Figure 129527DEST_PATH_IMAGE021
Row and column coordinates of the matrix to
Figure 553555DEST_PATH_IMAGE024
As a center, the window size is (2)n+1)*(2nThe region of +1) is a local region, whereinnCalculating a local average value for a positive integer
Figure 593055DEST_PATH_IMAGE025
And local variance
Figure 786139DEST_PATH_IMAGE026
，
Local mean value
Figure 128259DEST_PATH_IMAGE025
The calculation formula of (2) is as follows:
Figure 538118DEST_PATH_IMAGE027
the calculation formula of the local variance is as follows:
Figure 115730DEST_PATH_IMAGE088
in the formula,
Figure 569845DEST_PATH_IMAGE025
is the average of the local areas and is,
Figure 207500DEST_PATH_IMAGE029
the local standard deviation is taken as the local standard deviation,aandbrespectively are horizontal and vertical coordinates in a local area;
3.3) is provided with
Figure 606120DEST_PATH_IMAGE030
Is composed of
Figure 488888DEST_PATH_IMAGE023
And corresponding to the enhanced pixel value, wherein the calculation formula of the enhanced pixel value is as follows:
Figure 531930DEST_PATH_IMAGE089
in the formula,
Figure 340486DEST_PATH_IMAGE032
is a gain constant, and only needs to satisfy the requirement when adopting the traditional calculation method
Figure 960823DEST_PATH_IMAGE033
In order to solve the problem that some high frequency parts may be over-enhanced, the embodiment further comprises:
Figure 145817DEST_PATH_IMAGE090
then
Figure 43366DEST_PATH_IMAGE091
In the formula,Cis a constant value with a value range of more than 0 and less than 1,
Figure 515499DEST_PATH_IMAGE036
is the global mean square error;
because the improved contrast gain mu is spatially adaptive and inversely proportional to the local mean square error gamma _ f (i, j), the local mean square error is larger at the edge of the vein image or other places with severe changes, so the gain is smaller, and the oscillation generated at the places with severe changes of the gray scale of the output image can be avoided. In a smooth region, the local mean square error is small, so that the value of μ is large, and the constant C and the global mean square error δ are used for controlling the degree of high-frequency enhancement to avoid causing noise enhancement.
4) Recognizing the enhanced vein image based on the gray level co-occurrence matrix, and specifically comprising the following steps of:
4.1) performing characteristic compression on the image, namely compressing the gray level of the original image to reduce the calculated amount, and quantizing the gray level into 16 levels at 256 levels. Since the vein image gray scale is generally distributed in a narrow range, if the gray scale of the original vein image is directly compressed to 16 levels, the image definition is reduced. However, in step 2), the local contrast enhancement is already performed on the image, and the dynamic range of the gray value is increased, so that the overall contrast effect of the image is increased. So here the enhanced vein image gray scale can be directly divided by 16 and rounded, converting 0-255 gray scale to 0-15 gray scale;
4.2) solving the gray level co-occurrence matrix
Figure 154290DEST_PATH_IMAGE037
Gray level co-occurrence matrix
Figure 752762DEST_PATH_IMAGE037
Is shown as
Figure 629451DEST_PATH_IMAGE038
In the direction and at a distance
Figure 45389DEST_PATH_IMAGE039
Respectively having a gray value
Figure 141783DEST_PATH_IMAGE040
And
Figure 543946DEST_PATH_IMAGE041
probability of occurrence, then pixel pair
Figure 275141DEST_PATH_IMAGE042
And
Figure 861981DEST_PATH_IMAGE043
gray level co-occurrence matrix in four directions
Figure 944206DEST_PATH_IMAGE037
The calculation formula of (c) is:
Figure 884480DEST_PATH_IMAGE092
wherein,
Figure 234297DEST_PATH_IMAGE045
；
in the formula,
Figure 992037DEST_PATH_IMAGE046
is a point
Figure 295980DEST_PATH_IMAGE047
Is determined by the gray-scale value of (a),
Figure 39945DEST_PATH_IMAGE048
is a point
Figure 11312DEST_PATH_IMAGE049
Is measured in a predetermined time period, and the gray value of (b),Lis a gray scale level of the image,
Figure 175839DEST_PATH_IMAGE050
and
Figure 232657DEST_PATH_IMAGE051
representing the number of lines and rows of the vein image;
4.3) co-occurrence matrix according to gray level
Figure 514734DEST_PATH_IMAGE037
Taking a distance
Figure 340607DEST_PATH_IMAGE039
Is an angle of 1
Figure 440150DEST_PATH_IMAGE038
0 degree, 45 degrees, 90 degrees and 135 degrees respectively, and then the gray level co-occurrence matrix is aligned
Figure 859630DEST_PATH_IMAGE037
And (4) carrying out normalization, wherein the calculation formula is as follows:
Figure 308847DEST_PATH_IMAGE093
in the formula,
Figure 989227DEST_PATH_IMAGE053
in order to normalize the co-occurrence matrix,
Figure 525251DEST_PATH_IMAGE054
for the normalized constants, the calculation formula is:
Figure 25502DEST_PATH_IMAGE094
where M is the size of the co-occurrence matrix, typically 8, 16, 32.
4.4) obtaining each texture characteristic, namely, the gray level co-occurrence matrix
Figure 649381DEST_PATH_IMAGE037
4 texture parameters such as energy, entropy, moment of inertia, correlation and the like are calculated, wherein the calculation formula of the energy ASM is as follows:
Figure 951312DEST_PATH_IMAGE095
entropy of the entropy
Figure 127078DEST_PATH_IMAGE057
The calculation formula of (2) is as follows:
Figure 380205DEST_PATH_IMAGE058
moment of inertia
Figure 807775DEST_PATH_IMAGE059
The calculation formula of (2) is as follows:
Figure 728327DEST_PATH_IMAGE096
correlation
Figure 839109DEST_PATH_IMAGE061
The calculation formula of (2) is as follows:
Figure 454898DEST_PATH_IMAGE062
in the formula, the average value of the gray scale,
Figure 545214DEST_PATH_IMAGE064
in order to smooth the average value of the average,
Figure 54693DEST_PATH_IMAGE065
as a standard deviation of the gray scale,
Figure 837841DEST_PATH_IMAGE066
in order to smooth out the standard deviation of the standard deviation,
Figure 940926DEST_PATH_IMAGE063
、
Figure 336398DEST_PATH_IMAGE064
、
Figure 434804DEST_PATH_IMAGE065
、
Figure 388853DEST_PATH_IMAGE066
the calculation methods of (A) are respectively as follows:
Figure 103868DEST_PATH_IMAGE097
connecting the mean values and standard deviations of the energy, entropy, moment of inertia and correlation 4 texture parameters on four angles in series to serve as a template characteristic matrix of the image;
4.5) comparing the characteristics, namely subtracting the characteristic matrix to be matched from the characteristic matrix of the template to obtain a difference matrix A, and then calculating the Euclidean norm of the matrix A, wherein the norm of the matrix A can be realized by solving the maximum singular value of the matrix, and the calculation formula is as follows:
Figure 411353DEST_PATH_IMAGE098
in the formula,
Figure 411537DEST_PATH_IMAGE069
representing singular values of the difference matrix A, and T represents a maximum singular value;
4.6) selecting a threshold value by utilizing the size of the maximum singular value T, and carrying out image identification according to the threshold value.
Example 2
Referring to fig. 2, the present invention relates to a vein recognition apparatus based on a gray-scale co-occurrence matrix, which includes:
1) a normalization preprocessing module, configured to perform scale normalization preprocessing on the vein image, where the normalization preprocessing module is configured to implement the function of step 1) in embodiment 1.
2) And the filtering processing module is used for performing filtering processing on the vein image subjected to the normalization preprocessing by using homomorphic filtering, and the filtering processing module is used for realizing the function of the step 2) in the embodiment 1.
3) And the enhancement processing module is used for performing enhancement processing on the filtered vein image by using a local contrast enhancement method, and is used for realizing the function of the step 3) in the embodiment 1.
4) An image recognition module, configured to recognize the enhanced vein image based on the gray level co-occurrence matrix, where the image recognition module is configured to implement the function in step 4) in embodiment 1.
Obviously, the vein recognition apparatus based on the gray-scale co-occurrence matrix of the present embodiment can implement the vein recognition method described in embodiment 1.
The present invention has been described in detail with reference to the embodiments, but the description is only for the preferred embodiments of the present invention and should not be construed as limiting the scope of the present invention. All equivalent changes and modifications made within the scope of the present invention shall fall within the scope of the present invention.

Claims (6)

1. A vein identification method based on a gray level co-occurrence matrix is characterized in that: which comprises the following steps:
1) carrying out scale normalization pretreatment on the vein image;
2) filtering the vein image after the normalization pretreatment by using homomorphic filtering;
3) enhancing the filtered vein image by using a local contrast enhancement method;
4) the vein image after enhancement processing is identified based on the gray level co-occurrence matrix, and the method specifically comprises the following steps:
4.1) carrying out feature compression on the image;
4.2) solving the gray level co-occurrence matrix
Figure DEST_PATH_IMAGE001
Gray level co-occurrence matrix
Figure 179692DEST_PATH_IMAGE001
Is shown as
Figure 276961DEST_PATH_IMAGE002
In the direction and at a distance
Figure DEST_PATH_IMAGE003
Has a gray value respectively
Figure 640946DEST_PATH_IMAGE004
And
Figure DEST_PATH_IMAGE005
probability of occurrence, then pixel pair
Figure 825022DEST_PATH_IMAGE006
And
Figure DEST_PATH_IMAGE007
gray level co-occurrence matrix in four directions
Figure 274458DEST_PATH_IMAGE001
The calculation formula of (2) is as follows:
Figure 440997DEST_PATH_IMAGE008
wherein,
Figure DEST_PATH_IMAGE009
；
in the formula,
Figure 393909DEST_PATH_IMAGE010
is a point
Figure DEST_PATH_IMAGE011
Is determined by the gray-scale value of (a),
Figure 754746DEST_PATH_IMAGE012
is a point
Figure DEST_PATH_IMAGE013
Is determined by the gray-scale value of (a),Lis a gray scale level of the image,
Figure 691478DEST_PATH_IMAGE014
and
Figure DEST_PATH_IMAGE015
representing the number of lines and rows of the vein image;
4.3) co-occurrence matrix according to gray level
Figure 927287DEST_PATH_IMAGE001
Taking a distance
Figure 469127DEST_PATH_IMAGE003
Is an angle of 1
Figure 233821DEST_PATH_IMAGE002
0 degree, 45 degrees, 90 degrees and 135 degrees respectively, and then the gray level co-occurrence matrix is aligned
Figure 362576DEST_PATH_IMAGE001
And (4) carrying out normalization, wherein the calculation formula is as follows:
Figure 605338DEST_PATH_IMAGE016
in the formula,
Figure DEST_PATH_IMAGE017
in order to normalize the co-occurrence matrix,
Figure 532843DEST_PATH_IMAGE018
for the normalized constant, the calculation formula is:
Figure DEST_PATH_IMAGE019
in the formula,Mthe size of the symbiotic matrix is 8, 16 and 32;
4.4) obtaining each texture characteristic,i.e. to gray level co-occurrence matrix
Figure 999596DEST_PATH_IMAGE001
4 texture parameters of energy, entropy, moment of inertia and correlation are calculated, wherein the calculation formula of the energy ASM is as follows:
Figure 114183DEST_PATH_IMAGE020
entropy of the entropy
Figure DEST_PATH_IMAGE021
The calculation formula of (2) is as follows:
Figure 927680DEST_PATH_IMAGE022
moment of inertia
Figure DEST_PATH_IMAGE023
The calculation formula of (2) is as follows:
Figure 240850DEST_PATH_IMAGE024
correlation
Figure DEST_PATH_IMAGE025
The calculation formula of (c) is:
Figure 612926DEST_PATH_IMAGE026
in the formula,
Figure DEST_PATH_IMAGE027
is the mean value of the gray levels,
Figure 745967DEST_PATH_IMAGE028
in order to smooth the average value of the average,
Figure DEST_PATH_IMAGE029
as a standard deviation of the gray scale,
Figure 622875DEST_PATH_IMAGE030
in order to smooth out the standard deviation of the standard deviation,
Figure 993813DEST_PATH_IMAGE027
、
Figure 5632DEST_PATH_IMAGE028
、
Figure 94811DEST_PATH_IMAGE029
、
Figure 483067DEST_PATH_IMAGE030
the calculation methods of (A) are respectively as follows:
Figure DEST_PATH_IMAGE031
connecting the mean values and standard deviations of the energy, entropy, moment of inertia and correlation 4 texture parameters on four angles in series to serve as a template characteristic matrix of the image;
4.5) comparing the characteristics, namely subtracting the characteristic matrix to be matched from the characteristic matrix of the template to obtain a difference matrix A, and then calculating the Euclidean norm of the matrix A, wherein the norm of the matrix A can be realized by solving the maximum singular value of the matrix, and the calculation formula is as follows:
Figure 741135DEST_PATH_IMAGE032
in the formula,
Figure DEST_PATH_IMAGE033
representing the singular values of the difference matrix A, T representing the maximumA large singular value;
4.6) selecting a threshold value by utilizing the size of the maximum singular value T, and carrying out image identification according to the threshold value.
2. The vein identification method based on the gray level co-occurrence matrix according to claim 1, wherein: in the step 1), a bilinear interpolation method is adopted to carry out scale normalization processing on the vein image.
3. The vein identification method based on the gray level co-occurrence matrix according to claim 1, wherein: the specific way of performing filtering processing on the vein image after the normalization preprocessing by using homomorphic filtering in the step 2) is as follows:
2.1) vein image after normalization processing
Figure 720592DEST_PATH_IMAGE034
Carrying out logarithmic transformation, and leading:
Figure DEST_PATH_IMAGE035
in the formula,
Figure 828225DEST_PATH_IMAGE036
is a vein image after the logarithmic transformation,xandyrespectively a row coordinate and a column coordinate of the vein image matrix;
2.2) taking Fourier transform on two sides of the above formula, and enabling:
Figure DEST_PATH_IMAGE037
in the formula,
Figure 285752DEST_PATH_IMAGE038
for the logarithmically transformed vein image
Figure 867168DEST_PATH_IMAGE036
The fourier transformed image of (a) is,uandvrespectively x and y are variables obtained by Fourier transform,
Figure DEST_PATH_IMAGE039
is Fourier transform;
2.3) passing Gaussian filter
Figure 17527DEST_PATH_IMAGE040
Filtering the above formula to obtain a filtered frequency domain image
Figure DEST_PATH_IMAGE041
The calculation formula is as follows:
Figure 612456DEST_PATH_IMAGE042
wherein, the Gaussian filter
Figure 342515DEST_PATH_IMAGE040
The calculation formula of (2) is as follows:
Figure DEST_PATH_IMAGE043
in the formula:
Figure 808131DEST_PATH_IMAGE044
is a high-frequency amplification factor, has a value range of more than 1 and less than 2,
Figure DEST_PATH_IMAGE045
the offset is a parameter with the value range of more than 0 and less than 1 and controlling the slope of the filter
Figure 427593DEST_PATH_IMAGE048
Is the center of the filter and is,
Figure DEST_PATH_IMAGE049
is the mid-point in the frequency domain,
Figure 244240DEST_PATH_IMAGE050
in order to gain the gain at high frequency,
Figure DEST_PATH_IMAGE051
in order to gain the gain at low frequencies,cfor the sharpening constant, the value range is
Figure 43568DEST_PATH_IMAGE052
；
2.4) carrying out inverse Fourier transform on the filtered vein image to obtain an image with a frequency domain converted into a space domain
Figure DEST_PATH_IMAGE053
That is, the final filtered image is obtained, and the formula of the inverse fourier transform is:
Figure 124876DEST_PATH_IMAGE054
in the formula,
Figure DEST_PATH_IMAGE055
representing the inverse fourier transform.
4. The vein identification method based on the gray level co-occurrence matrix according to claim 3, wherein: the step 3) of enhancing the filtered vein image by using the local contrast enhancement method specifically comprises the following steps:
3.1) filtering the processed image
Figure 882616DEST_PATH_IMAGE053
Performing exponential operation to obtain homomorphic filtered image
Figure 389821DEST_PATH_IMAGE056
：
Figure DEST_PATH_IMAGE057
3.2) setting
Figure 992841DEST_PATH_IMAGE058
Is an image
Figure 934514DEST_PATH_IMAGE056
The gray value of a certain point in (b),iandjrespectively being said images
Figure 800839DEST_PATH_IMAGE056
Row and column coordinates of the matrix to
Figure DEST_PATH_IMAGE059
As a center, the window size is (2)n+1)*(2nThe region of +1) is a local region, whereinnCalculating a local average value for a positive integer
Figure 592078DEST_PATH_IMAGE060
And local variance
Figure DEST_PATH_IMAGE061
，
The local mean is calculated as:
Figure 733209DEST_PATH_IMAGE062
the calculation formula of the local variance is as follows:
Figure DEST_PATH_IMAGE063
in the formula,
Figure 559082DEST_PATH_IMAGE060
is the average of the local areas and is,
Figure 97773DEST_PATH_IMAGE064
the local standard deviation is taken as the local standard deviation,aandbrespectively are horizontal and vertical coordinates in a local area;
3.3) is provided with
Figure DEST_PATH_IMAGE065
Is composed of
Figure 376308DEST_PATH_IMAGE058
And corresponding to the enhanced pixel value, wherein the calculation formula of the enhanced pixel value is as follows:
Figure 55551DEST_PATH_IMAGE066
in the formula,
Figure DEST_PATH_IMAGE067
is a gain constant, satisfies
Figure 470352DEST_PATH_IMAGE068
And order:
Figure DEST_PATH_IMAGE069
then
Figure 976682DEST_PATH_IMAGE070
In the formula,Cis a constant value with a value range of more than 0 and less than 1,
Figure DEST_PATH_IMAGE071
is the global mean square error.
5. The vein identification method based on the gray level co-occurrence matrix according to claim 1, wherein: the step 4.1) of compressing the image features means compressing the gray scale of the original image, i.e. quantizing 256 gray scales into 16 gray scales, and converting 0-255 gray scales into 0-15 gray scales.
6. A vein recognition device based on a gray level co-occurrence matrix is characterized in that: it includes:
1) the normalization preprocessing module is used for carrying out scale normalization preprocessing on the vein image;
2) the filtering processing module is used for carrying out filtering processing on the vein image after the normalization preprocessing by using homomorphic filtering;
3) the enhancement processing module is used for enhancing the filtered vein image by using a local contrast enhancement method;
4) an image identification module for identifying the enhanced vein image based on the gray level co-occurrence matrix,
the method comprises the following specific steps:
4.1) carrying out feature compression on the image;
4.2) solving the gray level co-occurrence matrix
Figure 476933DEST_PATH_IMAGE072
Gray level co-occurrence matrix
Figure 694288DEST_PATH_IMAGE072
Is shown as
Figure 698016DEST_PATH_IMAGE002
In the direction and at a distance
Figure 77045DEST_PATH_IMAGE003
Has a gray value respectively
Figure 533434DEST_PATH_IMAGE004
And
Figure 288900DEST_PATH_IMAGE005
probability of occurrence, then pixel pair
Figure DEST_PATH_IMAGE073
And
Figure 462916DEST_PATH_IMAGE074
gray level co-occurrence matrix in four directions
Figure 278425DEST_PATH_IMAGE072
The calculation formula of (2) is as follows:
Figure DEST_PATH_IMAGE075
wherein,
Figure 753269DEST_PATH_IMAGE076
；
in the formula,
Figure DEST_PATH_IMAGE077
is a point
Figure 843584DEST_PATH_IMAGE078
Is determined by the gray-scale value of (a),
Figure DEST_PATH_IMAGE079
is a point
Figure 854528DEST_PATH_IMAGE080
Is determined by the gray-scale value of (a),Lis a gray scale level of the image,
Figure 840938DEST_PATH_IMAGE014
and
Figure 537499DEST_PATH_IMAGE015
representing the number of lines and rows of the vein image;
4.3) co-occurrence matrix according to gray level
Figure 634768DEST_PATH_IMAGE072
Taking a distance
Figure 467595DEST_PATH_IMAGE003
Is an angle of 1
Figure 126372DEST_PATH_IMAGE002
0 degree, 45 degrees, 90 degrees and 135 degrees respectively, and then the gray level co-occurrence matrix is aligned
Figure 44649DEST_PATH_IMAGE072
And (4) carrying out normalization, wherein the calculation formula is as follows:
Figure DEST_PATH_IMAGE081
in the formula,
Figure 211188DEST_PATH_IMAGE082
in order to normalize the co-occurrence matrix,
Figure 164101DEST_PATH_IMAGE018
for the normalized constants, the calculation formula is:
Figure DEST_PATH_IMAGE083
in the formula,Mthe size of the symbiotic matrix is 8, 16 and 32;
4.4) obtaining each texture characteristic, namely, the gray level co-occurrence matrix
Figure 23472DEST_PATH_IMAGE072
4 texture parameters of energy, entropy, moment of inertia and correlation are calculated, wherein the calculation formula of the energy ASM is as follows:
Figure 163467DEST_PATH_IMAGE084
entropy of the entropy
Figure DEST_PATH_IMAGE085
The calculation formula of (2) is as follows:
Figure 900741DEST_PATH_IMAGE086
moment of inertia
Figure DEST_PATH_IMAGE087
The calculation formula of (2) is as follows:
Figure 973739DEST_PATH_IMAGE088
correlation
Figure DEST_PATH_IMAGE089
The calculation formula of (2) is as follows:
Figure 4012DEST_PATH_IMAGE090
in the formula, the average value of the gray scale,
Figure 896881DEST_PATH_IMAGE028
in order to smooth the average value of the average,
Figure 369670DEST_PATH_IMAGE029
as a standard deviation of the gray scale,
Figure DEST_PATH_IMAGE091
in order to smooth out the standard deviation of the standard deviation,
Figure 828333DEST_PATH_IMAGE027
、
Figure 498349DEST_PATH_IMAGE028
、
Figure 612936DEST_PATH_IMAGE029
、
Figure 393810DEST_PATH_IMAGE091
the calculation methods of (A) are respectively as follows:
Figure 175821DEST_PATH_IMAGE092
connecting the mean values and standard deviations of 4 texture parameters of energy, entropy, moment of inertia and correlation in series at four angles to serve as a template characteristic matrix of the image;
4.5) comparing the characteristics, namely subtracting the characteristic matrix to be matched from the characteristic matrix of the template to obtain a difference matrix A, and then calculating the Euclidean norm of the matrix A, wherein the norm of the matrix A can be realized by solving the maximum singular value of the matrix, and the calculation formula is as follows:
Figure DEST_PATH_IMAGE093
in the formula,
Figure 49361DEST_PATH_IMAGE094
representing the singular value of the difference matrix A, and T represents the maximum singular value;
4.6) selecting a threshold value by utilizing the size of the maximum singular value T, and carrying out image identification according to the threshold value.
CN202210561152.3A 2022-05-23 2022-05-23 Vein identification method and device based on gray level co-occurrence matrix Active CN114724188B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210561152.3A CN114724188B (en) 2022-05-23 2022-05-23 Vein identification method and device based on gray level co-occurrence matrix

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210561152.3A CN114724188B (en) 2022-05-23 2022-05-23 Vein identification method and device based on gray level co-occurrence matrix

Publications (2)

Publication Number Publication Date
CN114724188A CN114724188A (en) 2022-07-08
CN114724188B true CN114724188B (en) 2022-08-16

Family

ID=82231716

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210561152.3A Active CN114724188B (en) 2022-05-23 2022-05-23 Vein identification method and device based on gray level co-occurrence matrix

Country Status (1)

Country Link
CN (1) CN114724188B (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115272684B (en) * 2022-09-29 2022-12-27 山东圣点世纪科技有限公司 Method for processing pseudo noise in vein image enhancement process
CN115578760B (en) * 2022-11-15 2023-05-30 山东圣点世纪科技有限公司 Control system and control method based on topographic relief vein recognition
CN115830023B (en) * 2023-02-16 2023-05-09 山东中都机器有限公司 Image processing-based belt conveyor belt defect detection method
CN117392056A (en) * 2023-09-06 2024-01-12 北京长木谷医疗科技股份有限公司 Homomorphic enhancement-based X-ray medical image normalization method and device

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109934231A (en) * 2019-01-08 2019-06-25 河北科技大学 Bar section character picture recognition methods based on multi-direction gray level co-occurrence matrixes

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8244012B2 (en) * 2007-02-05 2012-08-14 Siemens Medical Solutions Usa, Inc. Computer aided detection of pulmonary embolism with local characteristic features in CT angiography
CN101196987B (en) * 2007-12-25 2011-05-25 哈尔滨工业大学 On-line palm print, palm vein image personal identification method and its special capturing instrument
CN109829511B (en) * 2019-02-28 2022-04-08 中国人民解放军国防科技大学 Texture classification-based method for detecting cloud layer area in downward-looking infrared image
CN111507206B (en) * 2020-03-29 2023-04-18 杭州电子科技大学 Finger vein identification method based on multi-scale local feature fusion
CN112381788B (en) * 2020-11-13 2022-11-22 北京工商大学 Part surface defect increment detection method based on double-branch matching network
CN112560808B (en) * 2021-02-19 2021-05-07 北京圣点云信息技术有限公司 In-vivo vein identification method and device based on gray information

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109934231A (en) * 2019-01-08 2019-06-25 河北科技大学 Bar section character picture recognition methods based on multi-direction gray level co-occurrence matrixes

Also Published As

Publication number Publication date
CN114724188A (en) 2022-07-08

Similar Documents

Publication Publication Date Title
CN114724188B (en) Vein identification method and device based on gray level co-occurrence matrix
CN108122213B (en) A kind of soft image Enhancement Method based on YCrCb
Bhutada et al. Edge preserved image enhancement using adaptive fusion of images denoised by wavelet and curvelet transform
CN108932536B (en) Face posture reconstruction method based on deep neural network
CN108288258B (en) Low-quality image enhancement method under severe weather condition
CN105760841B (en) Identity recognition method and system
CN107392866A (en) A kind of facial image local grain Enhancement Method of illumination robust
Agarwal et al. Medical images contrast enhancement using quad weighted histogram equalization with adaptive gama correction and homomorphic filtering
CN107292834B (en) Infrared image detail enhancement method
CN104574293A (en) Multiscale Retinex image sharpening algorithm based on bounded operation
Mustafa et al. Image enhancement technique on contrast variation: a comprehensive review
Priyanka et al. Low-light image enhancement by principal component analysis
Kumar et al. Image Denoising by Wavelet Based Thresholding Method
Khandizod et al. Comparative analysis of image enhancement technique for hyperspectral palmprint images
Sharumathi et al. A survey on various image enhancement techniques for underwater acoustic images
CN113222833B (en) Side-scan sonar image processing method and device
Hou et al. Super-resolution reconstruction algorithm for terahertz images
CN112560808B (en) In-vivo vein identification method and device based on gray information
Budhiraja et al. Effect of pre-processing on MST based infrared and visible image fusion
CN112669239B (en) Industrial X-ray image contrast enhancement method
Zhou et al. Human visual system based automatic underwater image enhancement in NSCT domain
CN115511726A (en) X-ray image unsharp double-mask image enhancement method and system
Jung et al. SQI-based illumination normalization for face recognition based on discrete wavelet transform
CN108537807A (en) One kind enhancing detection method based on the improved edge of No. four satellite images of high score
Guo et al. A threshold segmentation method for non-uniform illumination image based on brightness equalization

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