CN112784922A - Extraction and classification method of intelligent cloud medical images - Google Patents

Extraction and classification method of intelligent cloud medical images Download PDF

Info

Publication number
CN112784922A
CN112784922A CN202110166942.7A CN202110166942A CN112784922A CN 112784922 A CN112784922 A CN 112784922A CN 202110166942 A CN202110166942 A CN 202110166942A CN 112784922 A CN112784922 A CN 112784922A
Authority
CN
China
Prior art keywords
image
gray
abnormal
image data
medical
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.)
Withdrawn
Application number
CN202110166942.7A
Other languages
Chinese (zh)
Inventor
郭付国
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.)
Chongqing Yukang Pharmaceutical Technology Co ltd
Original Assignee
Chongqing Yukang Pharmaceutical 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 Chongqing Yukang Pharmaceutical Technology Co ltd filed Critical Chongqing Yukang Pharmaceutical Technology Co ltd
Priority to CN202110166942.7A priority Critical patent/CN112784922A/en
Publication of CN112784922A publication Critical patent/CN112784922A/en
Withdrawn legal-status Critical Current

Links

Images

Classifications

    • 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
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
    • G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V2201/00—Indexing scheme relating to image or video recognition or understanding
    • G06V2201/03—Recognition of patterns in medical or anatomical images

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)
  • Multimedia (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)
  • Apparatus For Radiation Diagnosis (AREA)

Abstract

The invention discloses an extraction and classification method of intelligent cloud medical images, which comprises the following steps: s1, acquiring medical image data through the cloud server, performing gray scale distribution model calculation on the medical image data, and performing abnormal feature point extraction on the image calculated by the gray scale distribution model; s2, mining abnormal image data of different gray scales according to the relation data model by the extracted abnormal feature points; grading the abnormal image data in a gray distribution mode; and S3, setting a judgment model of gray scale distribution mode grading, classifying the abnormal feature points of the medical images according to the judgment model, and uploading the abnormal feature points to a cloud server.

Description

Extraction and classification method of intelligent cloud medical images
Technical Field
The invention relates to the field of digital medical image extraction, in particular to an intelligent cloud medical image extraction and classification method.
Background
Since the digital medical images are uploaded to the internet platform, and the data volume of the image atlas of different tissues and organs is huge in the process of extracting the medical images, it is not easy to obtain classified images of different tissues and organs, and the extraction operation of abnormal characteristic points is needed in the classified images, so that screening and correction are needed in the process of extracting the medical images, and a person skilled in the art is urgently needed to solve corresponding technical problems.
Disclosure of Invention
The invention aims to at least solve the technical problems in the prior art, and particularly creatively provides an intelligent cloud medical image extracting and classifying method.
In order to achieve the above object, the present invention provides an extraction and classification method for intelligent cloud medical images, comprising the following steps:
s1, acquiring medical image data through the cloud server, performing gray scale distribution model calculation on the medical image data, and performing abnormal feature point extraction on the image calculated by the gray scale distribution model;
s2, mining abnormal image data of different gray scales according to the relation data model by the extracted abnormal feature points; grading the abnormal image data in a gray distribution mode;
and S3, setting a judgment model of gray scale distribution mode grading, classifying the abnormal feature points of the medical images according to the judgment model, and uploading the abnormal feature points to a cloud server.
Preferably, the S1 includes:
s1-1, after medical image data are obtained, an image decoding process is carried out, stability judgment is carried out on image features of different color areas, and binarization resolution extraction operation is carried out according to gray values;
s1-2, according to the acquired medical image data, an image center point is taken, a circle area formed by an extension radius R is set, pixel points are subjected to alternate sampling to form a sampling sequence, weighting processing is carried out according to a sampling gray value, and gray scale distribution model calculation is completed;
Figure BDA0002937885660000021
n is (1,255), the value range of the weighting coefficient alpha (i, j) is that the transverse pixel point i carries out the size invariance on the image block GiAnd the vertical pixel point j performs image block G according to size invariancejThe degree of image correlation between alpha (i, j) E [1,8 ∈]，HaIs the gaussian distribution of the matrix set standard deviation a;
Figure BDA0002937885660000022
is a dot product operation, and epsilon is a gray characteristic factor;
preferably, the S1 further includes:
s1-3, in the abnormal feature point extraction process, the same pixel points are set through the reference image set and the extracted image set, and feature comparison is carried out after compression;
comparing the image features in the extracted image set with the reference image set, wherein the brightness of the image features in each extracted image set is different; a plurality of image frames in the reference image set are image frames of medical image data in an intact state, and the matching efficiency of the extracted image set is determined according to the updating time of the image frames in the reference image set;
carrying out the preliminary screening of abnormal feature points through transmission parameters, and setting an extraction model for extracting an image set:
Figure BDA0002937885660000023
wherein the content of the first and second substances,
Figure BDA0002937885660000024
in order to extract the image set gray level mean value, image subblocks with transverse pixel points i unchanged according to size
Figure BDA0002937885660000025
Image sub-block with vertical pixel point j unchanged according to size
Figure BDA0002937885660000026
After the transposition T is carried out, the spatial distribution value M of the abnormal characteristic point image of the extraction image set is multiplied, the distribution coefficient beta is used for adjusting,
the image frame transmission rate of the image centralized cache is extracted in the process of calculating the gray characteristic points;
Figure BDA0002937885660000027
wherein, the image block G is invariable according to the size by the transverse pixel point iiSet of (2)
Figure BDA0002937885660000031
And image block G with vertical pixel point j unchanged according to sizejSet of (2)
Figure BDA0002937885660000032
FstandardA frequency is invoked for a reference image frame of a reference image set.
Preferably, the S1 further includes:
s1-4, after comprehensive evaluation of the extraction model and the image frame transmission rate, the extraction of the abnormal feature points is performed rapidly, the extraction quantity is determined according to the reference image frame quantity of the reference image set and the quantity of the extracted image frames in the extracted image frames, and according to the resolution I and the contrast J of the extracted image frames, the extraction rule of the abnormal feature points is defined as:
Figure BDA0002937885660000033
wherein U E [1,128 ]]U is the influence factor of adjacent sampled image frames in the medical image data, and since 8 orientations are adopted, 8 is multiplied,
Figure BDA0002937885660000034
is a desizing estimation value of x, y resolution of adjacent sampling image frame coordinates in medical image data,
Figure BDA0002937885660000035
is the desizing estimated value of the x, y contrast of adjacent sampling image frame coordinates in the medical image data, v is the influence factor of the sampling image frame of the interval point in the medical image data,
Figure BDA0002937885660000036
for sampling the dessicant estimated value of the image frame coordinate x, y resolution at intervals in the medical image data,
Figure BDA0002937885660000037
for sampling the dessicant estimated value of image frame coordinate x, y contrast at intervals in medical image data,w is a texture factor of all sampled image frames in the medical image data, and L is a texture gray level correlation value of all sampled image frames in the medical image data;
Figure BDA0002937885660000038
preferably, the S2 includes:
s2-1, performing complete conversion on the extracted image set through the extraction rule of the abnormal feature points, importing the extracted image set into a relational data model for abnormal feature point analysis, judging similarity through simplifying calculated amount in the complete conversion process, setting a similarity value range, and performing non-maximum inhibition thresholding and edge reconstruction on an image frame extracted through the abnormal feature points in the medical image data to obtain a medical image data closed shape histogram;
wherein the relational data model is:
Figure BDA0002937885660000041
calculating weighting factors of longitudinal texture spectrums of medical image data divided according to image frame resolution
Figure BDA0002937885660000042
And weighting factors of transverse texture spectrums of medical image data divided according to image frame resolution
Figure BDA0002937885660000043
W is more than 1 and less than 255, d is an image frame noise coefficient, and lambda is an image frame filtering parameter; calculating weighting factors of longitudinal texture spectrums of medical image data divided according to image frame contrast
Figure BDA0002937885660000044
Weighting factors of transverse texture spectrum divided by medical image data according to image frame contrast
Figure BDA0002937885660000045
To the accumulation function of W, the image block of the transverse pixel point i according to the size invariance is GiThe image block of the vertical pixel point j with invariance according to the size is Gj；
Thereby obtaining a weighted texture histogram vector C of horizontal and vertical resolution and contrast1(k) And C2(k)；
S2-2, establishing an extracted image of the abnormal feature points through a relational data model by the medical image data, then carrying out refined gray distribution mode division processing, uniformly cutting the extracted abnormal feature point image to obtain an image, uniformly cutting the image into uniform image blocks, and processing each uniform image block; and (5) monitoring the frame rate of the equal image blocks one by one, and detecting abnormal characteristic points when dividing the gray distribution mode.
Preferably, the S2 further includes:
s2-3, when dividing a gray distribution mode, extracting abnormal feature points in the processed medical image data in the abnormal feature point extraction image at the x and y positions of the sampling image frame coordinate, acquiring an abnormal feature point outline, dividing a target object boundary of the abnormal feature point extraction image based on the abnormal feature point outline, performing image analysis in the x and y directions to obtain a transverse axis analysis abnormal point image set and a longitudinal axis analysis abnormal point image set, merging the connected areas of the transverse axis analysis abnormal point image set and the longitudinal axis analysis abnormal point image set, and performing smooth operation on the merged connected areas to acquire a gray distribution map of the extracted image; determining a gray scale change base point of the gray scale distribution map, acquiring abnormal feature point contours from the medical image data at the base point position for contour segmentation, and dividing different gray scale distributions into corresponding classification contour images;
s2-4, in the process of cutting the classification contour image, the cutting shape of the extracted image is a rectangle or a positive direction, then the gray distribution of each extracted image is analyzed one by one, and the gray distribution is marked and graded;
Figure BDA0002937885660000051
Figure BDA0002937885660000052
extracting gray level color variation degree E of image seti,jComparing with the average gray color change degree F of the reference image set, setting a judgment threshold value Q, extracting an image frame set divided by a gray distribution mode, and entering a grading stage according to the judgment threshold value; otherwise, analyzing the gray distribution of the next extracted image; wherein i is a horizontal pixel point of the image frame, j is a longitudinal pixel point of the image frame, q is a gray characteristic point sampling influence weight, f (i, j) is a value function of the horizontal pixel point and the longitudinal pixel point, and the average value O of the gray levels of the previous image frame is obtainedavgAnd the gray average value P of the next image frameavgThe square of the sum of the average value of gray-scale colors between the two and the effective value O of gray-scale of the last image framevalidAnd the effective value P of the gray scale of the next image framevalidThe square of the sum of the effective values of gray-scale colors between the two, s is the suppression threshold of the gray-scale image, r is the weight of the gradient of the gray-scale image,
Figure BDA0002937885660000053
is a dot product operation.
Preferably, the S3 includes:
s3-1, according to the gray distribution after the grading operation, screening the abnormal characteristic points of the image frame in the image frame circle area formed along the extension radius R,
Figure BDA0002937885660000054
σ is a weight of the gradation distribution, and an index η of the gradation distribution in the x direction in the area of the image frame circle formed with the radius RxAnd a gray scale distribution index mu in the y direction in the area of the image frame circle formed with a radius Ry，
S3-2, classifying the abnormal characteristic points of the medical image data through a judgment model, wherein the judgment model is as follows:
Figure BDA0002937885660000055
minimizing a function A (sigma) under the limitation condition of abnormal characteristic points; dividing the image frames of a transverse pixel point i and a longitudinal pixel point j into characteristic factors Q under the limiting conditioni,jAnd carrying out weighting operation with the condition of the abnormal classification so as to divide the abnormal characteristic points.
In summary, due to the adoption of the technical scheme, the invention has the beneficial effects that:
after the abnormal feature points of the medical image data are extracted, the gray level distribution mode is divided through the relational data model, the image uniformity and the image abnormal detection process are unified, and the error rate of resolution and contrast in the abnormal feature point extraction process is reduced, so that the medical image data can be more conveniently and finely positioned.
Additional aspects and advantages of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
The above and/or additional aspects and advantages of the present invention will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:
FIG. 1 is a general flow diagram of the present invention;
FIG. 2 is a diagram illustrating the effect of the present invention;
fig. 3 is a diagram of the effect of histogram implementation of the present invention.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the accompanying drawings are illustrative only for the purpose of explaining the present invention, and are not to be construed as limiting the present invention.
As shown in fig. 1 to 3, the invention discloses an extraction and classification method for intelligent cloud medical images, which comprises the following steps:
s1, acquiring medical image data through the cloud server, performing gray scale distribution model calculation on the medical image data, and performing abnormal feature point extraction on the image calculated by the gray scale distribution model;
s2, mining abnormal image data of different gray scales according to the relation data model by the extracted abnormal feature points; grading the abnormal image data in a gray distribution mode;
and S3, setting a judgment model of gray scale distribution mode grading, classifying the abnormal feature points of the medical images according to the judgment model, and uploading the abnormal feature points to a cloud server.
The S1 includes:
s1-1, after medical image data are obtained, an image decoding process is carried out, stability judgment is carried out on image features of different color areas, and binarization resolution extraction operation is carried out according to gray values;
s1-2, according to the acquired medical image data, an image center point is taken, a circle area formed by an extension radius R is set, pixel points are subjected to alternate sampling to form a sampling sequence, weighting processing is carried out according to a sampling gray value, and gray scale distribution model calculation is completed;
Figure BDA0002937885660000071
n is (1,255), the value range of the weighting coefficient alpha (i, j) is that the transverse pixel point i carries out the size invariance on the image block GiAnd the vertical pixel point j performs image block G according to size invariancejThe degree of image correlation between alpha (i, j) E [1,8 ∈]，HaIs the gaussian distribution of the matrix set standard deviation a;
Figure BDA0002937885660000072
is a dot product operation, and epsilon is a gray characteristic factor;
s1-3, in the abnormal feature point extraction process, the same pixel points are set through the reference image set and the extracted image set, and feature comparison is carried out after compression;
comparing the image features in the extracted image set with the reference image set, wherein the brightness of the image features in each extracted image set is different; a plurality of image frames in the reference image set are image frames of medical image data in an intact state, and the matching efficiency of the extracted image set is determined according to the updating time of the image frames in the reference image set;
carrying out the preliminary screening of abnormal feature points through transmission parameters, and setting an extraction model for extracting an image set:
Figure BDA0002937885660000073
wherein the content of the first and second substances,
Figure BDA0002937885660000074
in order to extract the image set gray level mean value, image subblocks with transverse pixel points i unchanged according to size
Figure BDA0002937885660000081
Image sub-block with vertical pixel point j unchanged according to size
Figure BDA0002937885660000082
After the transposition T is carried out, the spatial distribution value M of the abnormal characteristic point image of the extraction image set is multiplied, the distribution coefficient beta is used for adjusting,
the image frame transmission rate of the image centralized cache is extracted in the process of calculating the gray characteristic points;
Figure BDA0002937885660000083
wherein, the image block G is invariable according to the size by the transverse pixel point iiSet of (2)
Figure BDA0002937885660000084
And image block G with vertical pixel point j unchanged according to sizejSet of (2)
Figure BDA0002937885660000085
FstandardCalling a frequency for a reference image frame of a reference image set;
s1-4, after comprehensive evaluation of the extraction model and the image frame transmission rate, the extraction of the abnormal feature points is performed rapidly, the extraction quantity is determined according to the reference image frame quantity of the reference image set and the quantity of the extracted image frames in the extracted image frames, and according to the resolution I and the contrast J of the extracted image frames, the extraction rule of the abnormal feature points is defined as:
Figure BDA0002937885660000086
wherein U E [1,128 ]]U is the influence factor of adjacent sampled image frames in the medical image data, and since 8 orientations are adopted, 8 is multiplied,
Figure BDA0002937885660000087
is a desizing estimation value of x, y resolution of adjacent sampling image frame coordinates in medical image data,
Figure BDA0002937885660000088
is the desizing estimated value of the x, y contrast of adjacent sampling image frame coordinates in the medical image data, v is the influence factor of the sampling image frame of the interval point in the medical image data,
Figure BDA0002937885660000089
for sampling the dessicant estimated value of the image frame coordinate x, y resolution at intervals in the medical image data,
Figure BDA00029378856600000810
sampling the coordinate x and y of the image frame for every other point in the medical image data, removing the dryness estimation value of the contrast, wherein w is the texture factor of all the sampling image frames in the medical image data, and L is the texture gray level correlation value of all the sampling image frames in the medical image data;
Figure BDA00029378856600000811
the S2 includes:
s2-1, performing complete conversion on the extracted image set through the extraction rule of the abnormal feature points, importing the extracted image set into a relational data model for abnormal feature point analysis, judging similarity through simplifying calculated amount in the complete conversion process, setting a similarity value range, and performing non-maximum inhibition thresholding and edge reconstruction on an image frame extracted through the abnormal feature points in the medical image data to obtain a medical image data closed shape histogram;
wherein the relational data model is:
Figure BDA0002937885660000091
calculating weighting factors of longitudinal texture spectrums of medical image data divided according to image frame resolution
Figure BDA0002937885660000092
And weighting factors of transverse texture spectrums of medical image data divided according to image frame resolution
Figure BDA0002937885660000093
W is more than 1 and less than 255, d is an image frame noise coefficient, and lambda is an image frame filtering parameter; calculating weighting factors of longitudinal texture spectrums of medical image data divided according to image frame contrast
Figure BDA0002937885660000094
Weighting factors of transverse texture spectrum divided by medical image data according to image frame contrast
Figure BDA0002937885660000095
An accumulation function from W to W, wherein an image block of a transverse pixel point i according to size invariance is Gi, and an image block of a longitudinal pixel point j according to size invariance is Gj;
thereby obtaining a weighted texture histogram vector C of horizontal and vertical resolution and contrast1(k) And C2(k)；
As shown in fig. 2 and 3, in S2-2, the medical image data is processed by establishing an extracted image of the abnormal feature point through the relational data model, then performing a refined gray scale distribution pattern division process, uniformly cutting the extracted abnormal feature point image to obtain an image, uniformly cutting the image into uniform image blocks, and processing each uniform image block; and (5) monitoring the frame rate of the equal image blocks one by one, and detecting abnormal characteristic points when dividing the gray distribution mode.
S2-3, when dividing a gray distribution mode, extracting abnormal feature points in the processed medical image data in the abnormal feature point extraction image at the x and y positions of the sampling image frame coordinate, acquiring an abnormal feature point outline, dividing a target object boundary of the abnormal feature point extraction image based on the abnormal feature point outline, performing image analysis in the x and y directions to obtain a transverse axis analysis abnormal point image set and a longitudinal axis analysis abnormal point image set, merging the connected areas of the transverse axis analysis abnormal point image set and the longitudinal axis analysis abnormal point image set, and performing smooth operation on the merged connected areas to acquire a gray distribution map of the extracted image; determining a gray scale change base point of the gray scale distribution map, acquiring abnormal feature point contours from the medical image data at the base point position for contour segmentation, and dividing different gray scale distributions into corresponding classification contour images;
s2-4, in the process of cutting the classification contour image, the cutting shape of the extracted image is a rectangle or a positive direction, then the gray distribution of each extracted image is analyzed one by one, and the gray distribution is marked and graded;
Figure BDA0002937885660000101
Figure BDA0002937885660000102
extracting gray level color variation degree E of image seti,jComparing with the average gray level color variation F of the reference image set, setting a determination threshold Q, and extractingAn image frame set divided by a gray level distribution mode is output, and a grading stage is entered according to a judgment threshold value; otherwise, analyzing the gray distribution of the next extracted image; judging whether to carry out grading operation, wherein i is a transverse pixel point of the image frame, j is a longitudinal pixel point of the image frame, q is a sampling influence weight of the gray characteristic point, f (i, j) is a value function of the transverse pixel point and the longitudinal pixel point, and the average value O of the gray of the previous image frame is obtainedavgAnd the gray average value P of the next image frameavgThe square of the sum of the average value of gray-scale colors between the two and the effective value O of gray-scale of the last image framevalidAnd the effective value P of the gray scale of the next image framevalidThe square of the sum of the effective values of gray-scale colors between the two, s is the suppression threshold of the gray-scale image, r is the weight of the gradient of the gray-scale image,
Figure BDA0002937885660000103
is a dot product operation;
the S3 includes:
s3-1, according to the gray distribution after the grading operation, screening the abnormal characteristic points of the image frame in the image frame circle area formed along the extension radius R,
Figure BDA0002937885660000104
σ is a weight of the gradation distribution, and an index η of the gradation distribution in the x direction in the area of the image frame circle formed with the radius RxAnd a gray scale distribution index mu in the y direction in the area of the image frame circle formed with a radius Ry，
S3-2, classifying the abnormal characteristic points of the medical image data through a judgment model, wherein the judgment model is as follows:
Figure BDA0002937885660000111
minimizing a function A (sigma) under the limitation condition of abnormal characteristic points; dividing the image frames of a transverse pixel point i and a longitudinal pixel point j into characteristic factors Q under the limiting conditioni,jPerforming a weighting operation with the condition of the abnormality classification to thereby classify the abnormalityAnd dividing the feature points.
The beneficial effects are that: after the abnormal feature points of the medical image data are extracted, the gray level distribution mode is divided through the relational data model, the image uniformity and the image abnormal detection process are unified, and the error rate of resolution and contrast in the abnormal feature point extraction process is reduced, so that the medical image data can be more conveniently and finely positioned.
While embodiments of the invention have been shown and described, it will be understood by those of ordinary skill in the art that: various changes, modifications, substitutions and alterations can be made to the embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims (7)

1. An extraction and classification method for intelligent cloud medical images is characterized by comprising the following steps:
s1, acquiring medical image data through the cloud server, performing gray scale distribution model calculation on the medical image data, and performing abnormal feature point extraction on the image calculated by the gray scale distribution model;
s2, mining abnormal image data of different gray scales according to the relation data model by the extracted abnormal feature points; grading the abnormal image data in a gray distribution mode;
and S3, setting a judgment model of gray scale distribution mode grading, classifying the abnormal feature points of the medical images according to the judgment model, and uploading the abnormal feature points to a cloud server.
2. The method for extracting and classifying smart cloud medical images according to claim 1, wherein the step S1 includes:
s1-1, after medical image data are obtained, an image decoding process is carried out, stability judgment is carried out on image features of different color areas, and binarization resolution extraction operation is carried out according to gray values;
s1-2, according to the acquired medical image data, an image center point is taken, a circle area formed by an extension radius R is set, pixel points are subjected to alternate sampling to form a sampling sequence, weighting processing is carried out according to a sampling gray value, and gray scale distribution model calculation is completed;
Figure FDA0002937885650000011
n is (1,255), the value range of the weighting coefficient alpha (i, j) is that the transverse pixel point i carries out the size invariance on the image block GiAnd the vertical pixel point j performs image block G according to size invariancejThe degree of image correlation between alpha (i, j) E [1,8 ∈]，HaIs the gaussian distribution of the matrix set standard deviation a;
Figure FDA0002937885650000012
for dot product operation, ε is the grayscale character factor.
3. The method for extracting and classifying smart cloud medical images according to claim 2, wherein the step S1 further comprises:
s1-3, in the abnormal feature point extraction process, the same pixel points are set through the reference image set and the extracted image set, and feature comparison is carried out after compression;
comparing the image features in the extracted image set with the reference image set, wherein the brightness of the image features in each extracted image set is different; a plurality of image frames in the reference image set are image frames of medical image data in an intact state, and the matching efficiency of the extracted image set is determined according to the updating time of the image frames in the reference image set;
carrying out the preliminary screening of abnormal feature points through transmission parameters, and setting an extraction model for extracting an image set:
Figure FDA0002937885650000021
wherein the content of the first and second substances,
Figure FDA0002937885650000022
in order to extract the image set gray level mean value, image subblocks with transverse pixel points i unchanged according to size
Figure FDA0002937885650000023
Image sub-block with vertical pixel point j unchanged according to size
Figure FDA0002937885650000024
After the transposition T is carried out, the spatial distribution value M of the abnormal characteristic point image of the extraction image set is multiplied, the distribution coefficient beta is used for adjusting,
the image frame transmission rate of the image centralized cache is extracted in the process of calculating the gray characteristic points;
Figure FDA0002937885650000025
wherein, the image block G is invariable according to the size by the transverse pixel point iiSet of (2)
Figure FDA0002937885650000026
And image block G with vertical pixel point j unchanged according to sizejSet of (2)
Figure FDA0002937885650000027
FstandardA frequency is invoked for a reference image frame of a reference image set.
4. The method for extracting and classifying smart cloud medical images according to claim 3, wherein the step S1 further comprises:
s1-4, after comprehensive evaluation of the extraction model and the image frame transmission rate, the extraction of the abnormal feature points is performed rapidly, the extraction quantity is determined according to the reference image frame quantity of the reference image set and the quantity of the extracted image frames in the extracted image frames, and according to the resolution I and the contrast J of the extracted image frames, the extraction rule of the abnormal feature points is defined as:
Figure FDA0002937885650000031
wherein U E [1,128 ]]U is the influence factor of adjacent sampled image frames in the medical image data, and since 8 orientations are adopted, 8 is multiplied,
Figure FDA0002937885650000032
is a desizing estimation value of x, y resolution of adjacent sampling image frame coordinates in medical image data,
Figure FDA0002937885650000033
is the desizing estimated value of the x, y contrast of adjacent sampling image frame coordinates in the medical image data, v is the influence factor of the sampling image frame of the interval point in the medical image data,
Figure FDA0002937885650000034
for sampling the dessicant estimated value of the image frame coordinate x, y resolution at intervals in the medical image data,
Figure FDA0002937885650000035
sampling the coordinate x and y of the image frame for every other point in the medical image data, removing the dryness estimation value of the contrast, wherein w is the texture factor of all the sampling image frames in the medical image data, and L is the texture gray level correlation value of all the sampling image frames in the medical image data;
Figure FDA0002937885650000036
5. the method for extracting and classifying smart cloud medical images according to claim 1, wherein the step S2 includes:
s2-1, performing complete conversion on the extracted image set through the extraction rule of the abnormal feature points, importing the extracted image set into a relational data model for abnormal feature point analysis, judging similarity through simplifying calculated amount in the complete conversion process, setting a similarity value range, and performing non-maximum inhibition thresholding and edge reconstruction on an image frame extracted through the abnormal feature points in the medical image data to obtain a medical image data closed shape histogram;
wherein the relational data model is:
Figure FDA0002937885650000037
calculating weighting factors of longitudinal texture spectrums of medical image data divided according to image frame resolution
Figure FDA0002937885650000038
And weighting factors of transverse texture spectrums of medical image data divided according to image frame resolution
Figure FDA0002937885650000039
W is more than 1 and less than 255, d is an image frame noise coefficient, and lambda is an image frame filtering parameter; calculating weighting factors of longitudinal texture spectrums of medical image data divided according to image frame contrast
Figure FDA0002937885650000041
Weighting factors of transverse texture spectrum divided by medical image data according to image frame contrast
Figure FDA0002937885650000042
The cumulative function to W is that the image block of the horizontal pixel point i according to the size invariance is Gi, and the image block of the vertical pixel point j according to the size invariance is Gj；
Thereby obtaining a weighted texture histogram vector C of horizontal and vertical resolution and contrast1(k) And C2(k)；
S2-2, establishing an extracted image of the abnormal feature points through a relational data model by the medical image data, then carrying out refined gray distribution mode division processing, uniformly cutting the extracted abnormal feature point image to obtain an image, uniformly cutting the image into uniform image blocks, and processing each uniform image block; and (5) monitoring the frame rate of the equal image blocks one by one, and detecting abnormal characteristic points when dividing the gray distribution mode.
6. The method for extracting and classifying smart cloud medical images according to claim 5, wherein the step S2 further comprises:
s2-3, when dividing a gray distribution mode, extracting abnormal feature points in the processed medical image data in the abnormal feature point extraction image at the x and y positions of the sampling image frame coordinate, acquiring an abnormal feature point outline, dividing a target object boundary of the abnormal feature point extraction image based on the abnormal feature point outline, performing image analysis in the x and y directions to obtain a transverse axis analysis abnormal point image set and a longitudinal axis analysis abnormal point image set, merging the connected areas of the transverse axis analysis abnormal point image set and the longitudinal axis analysis abnormal point image set, and performing smooth operation on the merged connected areas to acquire a gray distribution map of the extracted image; determining a gray scale change base point of the gray scale distribution map, acquiring abnormal feature point contours from the medical image data at the base point position for contour segmentation, and dividing different gray scale distributions into corresponding classification contour images;
s2-4, in the process of cutting the classification contour image, the cutting shape of the extracted image is a rectangle or a positive direction, then the gray distribution of each extracted image is analyzed one by one, and the gray distribution is marked and graded;
Figure FDA0002937885650000043
Figure FDA0002937885650000051
extracting gray level color variation degree E of image seti,jComparing the average gray level color change degree with the reference image set to set a determination threshold valueQ, extracting an image frame set divided by a gray level distribution mode, and entering a grading stage according to a judgment threshold value; otherwise, analyzing the gray distribution of the next extracted image; wherein i is a horizontal pixel point of the image frame, j is a longitudinal pixel point of the image frame, q is a gray characteristic point sampling influence weight, f (i, j) is a value function of the horizontal pixel point and the longitudinal pixel point, and the average value O of the gray levels of the previous image frame is obtainedavgAnd the gray average value P of the next image frameavgThe square of the sum of the average value of gray-scale colors between the two and the effective value O of gray-scale of the last image framevalidAnd the effective value P of the gray scale of the next image framevalidThe square of the sum of the effective values of gray-scale colors between the two, s is the suppression threshold of the gray-scale image, r is the weight of the gradient of the gray-scale image,
Figure FDA0002937885650000052
is a dot product operation.
7. The method for extracting and classifying smart cloud medical images according to claim 1, wherein the step S3 includes:
s3-1, according to the gray distribution after the grading operation, screening the abnormal characteristic points of the image frame in the image frame circle area formed along the extension radius R,
Figure FDA0002937885650000053
σ is a weight of the gradation distribution, and an index η of the gradation distribution in the x direction in the area of the image frame circle formed with the radius RxAnd a gray scale distribution index mu in the y direction in the area of the image frame circle formed with a radius Ry，
S3-2, classifying the abnormal characteristic points of the medical image data through a judgment model, wherein the judgment model is as follows:
Figure FDA0002937885650000054
minimizing a function A (sigma) under the limitation condition of abnormal characteristic points;dividing the image frames of a transverse pixel point i and a longitudinal pixel point j into characteristic factors Q under the limiting conditioni,jAnd carrying out weighting operation with the condition of the abnormal classification so as to divide the abnormal characteristic points.
CN202110166942.7A 2021-02-07 2021-02-07 Extraction and classification method of intelligent cloud medical images Withdrawn CN112784922A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110166942.7A CN112784922A (en) 2021-02-07 2021-02-07 Extraction and classification method of intelligent cloud medical images

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110166942.7A CN112784922A (en) 2021-02-07 2021-02-07 Extraction and classification method of intelligent cloud medical images

Publications (1)

Publication Number Publication Date
CN112784922A true CN112784922A (en) 2021-05-11

Family

ID=75761151

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110166942.7A Withdrawn CN112784922A (en) 2021-02-07 2021-02-07 Extraction and classification method of intelligent cloud medical images

Country Status (1)

Country Link
CN (1) CN112784922A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116957524A (en) * 2023-09-21 2023-10-27 青岛阿斯顿工程技术转移有限公司 Talent information intelligent management method and system in technology transfer process
CN117853334A (en) * 2024-03-07 2024-04-09 中国人民解放军海军青岛特勤疗养中心 Medical image reconstruction method and system based on DICOM image

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107330869A (en) * 2017-06-28 2017-11-07 哈尔滨理工大学 Extraordinary image vegetarian refreshments reconstruct after overlapping cell segmentation
CN108320280A (en) * 2018-01-16 2018-07-24 南京理工大学 The crater image method for detecting abnormality of view-based access control model clarity and contours extract
CN109493343A (en) * 2018-12-29 2019-03-19 上海鹰瞳医疗科技有限公司 Medical image abnormal area dividing method and equipment
CN109870459A (en) * 2019-02-21 2019-06-11 武汉武大卓越科技有限责任公司 The track plates crack detection method of non-fragment orbit
CN111666439A (en) * 2020-05-28 2020-09-15 重庆渝抗医药科技有限公司 Working method for rapidly extracting and dividing medical image big data aiming at cloud environment

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107330869A (en) * 2017-06-28 2017-11-07 哈尔滨理工大学 Extraordinary image vegetarian refreshments reconstruct after overlapping cell segmentation
CN108320280A (en) * 2018-01-16 2018-07-24 南京理工大学 The crater image method for detecting abnormality of view-based access control model clarity and contours extract
CN109493343A (en) * 2018-12-29 2019-03-19 上海鹰瞳医疗科技有限公司 Medical image abnormal area dividing method and equipment
CN109870459A (en) * 2019-02-21 2019-06-11 武汉武大卓越科技有限责任公司 The track plates crack detection method of non-fragment orbit
CN111666439A (en) * 2020-05-28 2020-09-15 重庆渝抗医药科技有限公司 Working method for rapidly extracting and dividing medical image big data aiming at cloud environment

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116957524A (en) * 2023-09-21 2023-10-27 青岛阿斯顿工程技术转移有限公司 Talent information intelligent management method and system in technology transfer process
CN116957524B (en) * 2023-09-21 2024-01-05 青岛阿斯顿工程技术转移有限公司 Talent information intelligent management method and system in technology transfer process
CN117853334A (en) * 2024-03-07 2024-04-09 中国人民解放军海军青岛特勤疗养中心 Medical image reconstruction method and system based on DICOM image
CN117853334B (en) * 2024-03-07 2024-05-14 中国人民解放军海军青岛特勤疗养中心 Medical image reconstruction method and system based on DICOM image

Similar Documents

Publication Publication Date Title
CN114549522B (en) Textile quality detection method based on target detection
CN104990925B (en) One kind is based on gradient multi thresholds optimization defect inspection method
CN107038416B (en) Pedestrian detection method based on binary image improved HOG characteristics
CN110428450B (en) Scale-adaptive target tracking method applied to mine tunnel mobile inspection image
CN109540925B (en) Complex ceramic tile surface defect detection method based on difference method and local variance measurement operator
CN116645367B (en) Steel plate cutting quality detection method for high-end manufacturing
CN114723704A (en) Textile quality evaluation method based on image processing
CN114820625B (en) Automobile top block defect detection method
CN115994907B (en) Intelligent processing system and method for comprehensive information of food detection mechanism
CN112784922A (en) Extraction and classification method of intelligent cloud medical images
CN112862744B (en) Intelligent detection method for internal defects of capacitor based on ultrasonic image
CN109886932A (en) Gear ring of wheel speed sensor detection method of surface flaw based on SVM
CN109543686B (en) Character recognition preprocessing binarization method based on self-adaptive multi-threshold
CN116152242B (en) Visual detection system of natural leather defect for basketball
CN116883408B (en) Integrating instrument shell defect detection method based on artificial intelligence
CN114332081B (en) Textile surface abnormity determination method based on image processing
CN114494318B (en) Cornea contour extraction method based on cornea dynamic deformation video of Ojin algorithm
CN115018790A (en) Workpiece surface defect detection method based on anomaly detection
CN112785589A (en) Method for acquiring digital medical abnormal image through cloud network
CN109829511B (en) Texture classification-based method for detecting cloud layer area in downward-looking infrared image
CN113160166B (en) Medical image data mining working method through convolutional neural network model
CN113034403B (en) Working method for denoising medical image data through Gaussian filtering
CN116542910A (en) Stamping part round hole positioning and deformation detection method based on sub-pixel edge extraction
CN113643290B (en) Straw counting method and device based on image processing and storage medium
CN112784923A (en) Medical abnormal image classification method under big data environment

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
WW01 Invention patent application withdrawn after publication
WW01 Invention patent application withdrawn after publication

Application publication date: 20210511