CN116205920B - Method and system for generating key region detection model based on lung scanning data - Google Patents
Method and system for generating key region detection model based on lung scanning data Download PDFInfo
- Publication number
- CN116205920B CN116205920B CN202310493252.1A CN202310493252A CN116205920B CN 116205920 B CN116205920 B CN 116205920B CN 202310493252 A CN202310493252 A CN 202310493252A CN 116205920 B CN116205920 B CN 116205920B
- Authority
- CN
- China
- Prior art keywords
- extraction
- personnel
- setting
- areas
- scanning
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/149—Segmentation; Edge detection involving deformable models, e.g. active contour models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/64—Analysis of geometric attributes of convexity or concavity
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30061—Lung
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Abstract
The invention discloses a method and a system for generating a key region detection model based on lung scanning data, and relates to the technical field of lung scanning analysis; the system comprises a scanning database, a personnel dividing module and a key area generating module; the lung scanning data of a plurality of healthy persons are stored in the scanning database; the personnel division module is configured with a personnel division extraction strategy, and the personnel division extraction strategy comprises: dividing according to the personnel characteristics, corresponding the divided personnel characteristics into a scanning database, and acquiring lung scanning data corresponding to the personnel characteristics from the scanning database; according to the invention, the data in the existing database is analyzed to obtain the models of the lung detection areas of different people, so that the area detection accuracy in the self-help screening process is improved, and the problems of insufficient pertinence and low screening accuracy of the autonomous screening of the existing lung scanning data are solved.
Description
Technical Field
The invention relates to the technical field of lung scanning analysis, in particular to a method and a system for generating a key region detection model based on lung scanning data.
Background
The lung is located in the chest, one of the left and right sides, the left lung is closely adjacent to the heart, the front and back sides are protected by ribs, the left lung is divided into two leaves, and the right lung is three leaves. The surface of the lung is covered with a layer of soft membrane called as a dirty layer pleura, one side which is clung to the chest wall is called as a parietal pleura, the pleural cavity is negative pressure, when the respiratory muscle is contracted, the thoracic cavity is expanded, the lung expands to suck fresh air, otherwise, when the respiratory muscle is expanded, the thoracic cavity is elastically retracted, and waste gas is exhaled; when the lung is detected, the conventional examination technology of computer layer imaging of the lung is adopted, no contrast agent is injected, the conventional spiral transverse scanning is adopted, the scanning range is 2-3 cm below the diaphragm angle of the lung tip to the lower side rib, and the X-ray is utilized to irradiate the human body, so that some images of organs and tissues of the human body are obtained.
In the prior art, when lung detection is carried out, a doctor usually observes and judges a scanned medical image, and problem screening is carried out through experience of the doctor, in the existing method capable of automatically screening the lung image, the abnormal region is marked by extracting the contrast based on image comparison of tissues, but in the automatic screening process, a method for dividing different regions of the lung is lacking, so that pertinence judgment in different regions is lacking in judging, and the problems of high error rate and insufficient accuracy of automatic screening can occur.
Disclosure of Invention
The invention aims to solve at least one of the technical problems in the prior art to a certain extent, and can obtain models of lung detection areas of different people by analyzing the data in the existing database, thereby being beneficial to improving the accuracy of area detection in the self-help screening process so as to solve the problems of insufficient pertinence and lower screening accuracy of the autonomous screening of the problems of the existing lung scanning data.
In order to achieve the above object, a first aspect of the present invention provides a key region detection model generating system based on lung scan data, including a scan database, a personnel dividing module, and a key region generating module; the lung scanning data of a plurality of healthy persons are stored in the scanning database;
the personnel division module is configured with a personnel division extraction strategy, and the personnel division extraction strategy comprises: dividing according to the personnel characteristics, corresponding the divided personnel characteristics into a scanning database, and acquiring lung scanning data corresponding to the personnel characteristics from the scanning database;
the key region generation module comprises a gray level generation unit, a contour generation unit and a comprehensive generation unit, wherein the gray level generation unit is used for generating a gray level section of a key region based on analysis of acquired lung scanning data, the contour generation unit is used for generating a contour section of the key region based on analysis of the acquired lung scanning data, and the comprehensive generation unit is used for integrating the generated gray level section and the contour section of the key region to obtain a detection model of the key region.
Further, the staff member division extraction policy further includes a staff member division sub-policy, and the staff member division sub-policy includes: classifying the person features into gender features, age features and weight features, the gender features including men and women;
the age characteristics comprise a first age interval, and the first age interval is divided into a plurality of age sub-intervals according to the first age interval;
the weight characteristic comprises a first weight interval, and the first weight interval is divided into a plurality of weight sub-intervals according to the first weight interval;
and combining according to the classified sex characteristics, age characteristics and weight characteristics to obtain a plurality of personnel characteristic groups, and setting the number of the personnel characteristic groups as the personnel classification number.
Further, the personnel division extraction policy further includes an extraction sub-policy, the extraction sub-policy including: comparing the personnel characteristic group with personnel characteristics in a scanning database to obtain the quantity which can be matched with the personnel characteristics of the personnel characteristic group, and setting the quantity as the extractable quantity;
acquiring the total number of lung scanning data in a scanning database, setting the total number as the extractable total number, and dividing the extractable total number by the personnel dividing number to obtain the extraction reference number;
comparing the extractable quantity with the extraction reference quantity to obtain an extractable proportion, and deleting the personnel feature group corresponding to the extractable quantity when the extractable proportion is smaller than or equal to a first proportion threshold value;
setting a personnel characteristic group obtained after deletion processing as a personnel characteristic extraction group;
multiplying the first proportional threshold value by the extraction reference number to obtain a minimum extraction value, multiplying the second proportional threshold value by the extraction reference number to obtain a maximum extraction value, and setting a range from the minimum extraction value to the maximum extraction value as an extraction reference range;
and performing personnel feature matching on the personnel feature groups and the scanning database, extracting the maximum number of lung scanning data according to the extraction reference range, and setting the number of lung scanning data acquired in each personnel feature group as the scanning extraction number.
Further, the gradation generation unit is configured with a gradation generation policy including: generating a key region gray scale model of each group one by one according to the personnel characteristic groups;
acquiring a lung scan image from the lung scan data for each of the sets of human features;
setting the pixel proportion of the lung scanning image, and setting the set pixel proportion as a first pixel proportion;
setting a first extraction circle, randomly extracting a plurality of first extraction circles from the key area, setting the first extraction circles as a plurality of extraction areas, calculating the average value of gray values of a plurality of pixel points in the extraction areas, setting the average value as first extraction gray, calculating the average value of the first extraction gray corresponding to the plurality of extraction areas, and setting the average value as first extraction reference gray;
calculating the average value of a plurality of first extraction reference gray scales, and setting the average value as a second extraction reference gray scale;
obtaining minimum values and maximum values in a plurality of first extraction reference gray scales, respectively setting the minimum extraction reference gray scales and the maximum extraction reference gray scales, setting a section between the minimum extraction reference gray scales and the maximum extraction reference gray scales as an extraction reference gray scale section, dividing the extraction reference gray scale section into a plurality of extraction reference gray scale subsections, establishing a distribution histogram according to the plurality of extraction reference gray scale subsections, and setting the extraction reference gray scale subsection where the second extraction reference gray scale is located as an extraction gray scale reference section;
the extracted gradation reference section is set as a gradation section of the key region.
Further, the gray scale generation strategy further includes a region extraction sub-strategy, the region extraction sub-strategy including: obtaining the area of a key region, multiplying the area of the key region by a first extraction proportion to obtain the area of a first extraction circle, dividing the area of the key region by the area of the first extraction circle to obtain a first reference number, multiplying the first reference number by a second extraction proportion to obtain a first extraction numerical value, and setting the integer bit of the first extraction numerical value as the first extraction reference number;
and extracting a corresponding number of first extraction circles in the key area according to the first extraction reference number.
Further, the contour generation unit is configured with a contour generation policy, the contour generation policy comprising: extracting key areas of a plurality of lung scanning images according to gray intervals of the key areas, acquiring outlines of the extracted key areas, and setting the outlines as extraction outlines;
setting basic contours of key areas, and comparing a plurality of extracted contours with the basic contours to obtain convex areas and concave areas between each extracted contour and the basic contour;
overlapping the plurality of extraction contours with the basic contour, setting the number of overlapping convex areas as the number of overlapping convex areas, and setting the number of overlapping concave areas as the number of overlapping concave areas;
the method comprises the steps of obtaining a protruding proportion by comparing protruding superposition quantity with scanning extraction quantity, obtaining a recessed proportion by comparing recessed superposition quantity with scanning extraction quantity, performing protruding correction on a protruding region corresponding to a basic contour when the protruding proportion is larger than or equal to a first protruding quantity threshold value, performing recessed correction on a recessed region corresponding to the basic contour when the recessed proportion is larger than or equal to a first recessed quantity threshold value, and setting the corrected basic contour as a contour reference region;
the contour reference region is set as a contour section of the key region.
Further, the contour generation strategy further comprises a contour syndrome strategy, the contour syndrome strategy comprising: when the protruding proportion is greater than or equal to a first protruding quantity threshold value, setting a plurality of protruding areas as protruding correction areas; when the recess ratio is greater than or equal to a first recess number threshold, setting a plurality of recess areas as recess correction areas;
calculating the average value of the areas of a plurality of convex correction areas, setting the average value as a convex area reference area, setting a convex correction area with the smallest difference value with the convex area reference area as a convex correction selection area, and adding a convex correction selection area to a basic contour to carry out convex correction;
the average value of the areas of a plurality of concave correction areas is calculated, the average value is set as a concave area reference area, the concave correction area with the smallest difference value with the concave area reference area is set as a concave correction selection area, and the basic contour is added with the concave correction selection area to carry out concave correction.
In a second aspect, the present invention further provides a method for generating a key region detection model based on lung scan data, including:
acquiring lung scanning data of a plurality of healthy persons from a scanning database;
dividing according to the personnel characteristics, corresponding the divided personnel characteristics into a scanning database, and acquiring lung scanning data corresponding to the personnel characteristics from the scanning database;
based on analysis of the acquired lung scanning data, generating a gray scale section of the key region;
generating a contour section of the key region based on analysis of the acquired lung scan data;
and integrating according to the generated gray scale interval and contour interval of the key region to obtain a detection model of the key region.
The invention has the beneficial effects that: according to the method, the lung scanning data of a plurality of healthy people are obtained from the scanning database, the data base can be provided for the establishment of the key area detection model, the division is carried out according to the personnel characteristics, the corresponding personnel characteristics are corresponding to the scanning database after the division, the lung scanning data of the corresponding personnel characteristics are obtained from the scanning database, and the fine degree of the data extraction can be improved by subdividing the personnel characteristics, so that the accuracy of different people in key area comparison is improved;
the gray scale interval of the key area is generated based on analysis of the acquired lung scanning data; generating a contour section of the key region based on analysis of the acquired lung scan data; the detection model of the key region is obtained by integrating the generated gray scale region and the generated outline region of the key region, and the design can be helpful for improving the accuracy of lung region division in the process of computer automatic screening, and can accurately divide different regions of the lung when the focus appears in the lung.
Additional features and advantages of the application will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the application. The objectives and other advantages of the application will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
Drawings
FIG. 1 is a schematic block diagram of a system of the present invention;
FIG. 2 is a flow chart of method steps of the present invention;
FIG. 3 is a schematic diagram of a first circle extraction performed in a key region according to the present invention;
fig. 4 is a schematic diagram of the comparison of the extracted profile and the base profile according to the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Referring to fig. 1, the present invention provides a key region detection model generation system based on lung scan data, which can obtain models of lung detection regions of different people by analyzing data in an existing database, thereby helping to improve region detection accuracy in a self-help screening process, and the key region detection model generation system comprises a scan database, a personnel dividing module and a key region generation module; lung scanning data of a plurality of healthy persons are stored in the scanning database;
the personnel division module is configured with personnel division extraction strategies, and the personnel division extraction strategies comprise: dividing according to the personnel characteristics, corresponding the divided personnel characteristics into a scanning database, and acquiring lung scanning data corresponding to the personnel characteristics from the scanning database; the personnel division extraction strategy further comprises a personnel division sub-strategy, and the personnel division sub-strategy comprises: classifying the person features into gender features, age features and weight features, the gender features including men and women;
the age characteristics comprise a first age interval, and the first age interval is divided into a plurality of age sub-intervals according to the first age interval; in specific implementation, the first age interval is set to 0-100, and the first age interval is set to 5; if an area which cannot be covered appears in the dividing process, for example, when the age is higher than 100, dividing according to the closest area;
the weight characteristic comprises a first weight interval, and the first weight interval is divided into a plurality of weight sub-intervals according to the first weight interval; the first body weight interval is set to be 0-150kg, the first body weight interval is set to be 10, and if an uncovered interval appears in the dividing process, for example, when the body weight is higher than 150, the dividing is carried out according to the nearest interval;
and combining according to the classified sex characteristics, age characteristics and weight characteristics to obtain a plurality of personnel characteristic groups, and setting the number of the personnel characteristic groups as the personnel classification number.
The personnel division extraction policy further includes an extraction sub-policy, the extraction sub-policy including: comparing the personnel characteristic group with personnel characteristics in a scanning database to obtain the quantity which can be matched with the personnel characteristics of the personnel characteristic group, and setting the quantity as the extractable quantity;
acquiring the total number of lung scanning data in a scanning database, setting the total number as the extractable total number, and dividing the extractable total number by the personnel dividing number to obtain the extraction reference number; the extraction reference number is the average number of the total extractable numbers averaged into each of the human feature groups;
comparing the extractable quantity with the extraction reference quantity to obtain an extractable proportion, deleting a personnel feature group corresponding to the extractable quantity when the extractable proportion is smaller than or equal to a first proportion threshold, setting the extractable proportion to be 5%, and when the extractable proportion is smaller, obtaining less data of the group, wherein the effectiveness is lacking and the reference can be omitted;
setting a personnel characteristic group obtained after deletion processing as a personnel characteristic extraction group;
multiplying the first proportional threshold value by the extraction reference number to obtain a minimum extraction value, multiplying the second proportional threshold value by the extraction reference number to obtain a maximum extraction value, setting a range from the minimum extraction value to the maximum extraction value as an extraction reference range, and setting the second proportional threshold value to be between 10% and 30%;
and performing personnel feature matching on the personnel feature groups and the scanning database, extracting the maximum number of lung scanning data according to the extraction reference range, and setting the number of lung scanning data acquired in each personnel feature group as the scanning extraction number.
The key region generation module comprises a gray level generation unit, a contour generation unit and a comprehensive generation unit, wherein the gray level generation unit is used for generating a gray level section of the key region based on analysis of acquired lung scanning data, the gray level generation unit is configured with a gray level generation strategy, and the gray level generation strategy comprises: generating a key region gray scale model of each group one by one according to the personnel characteristic groups;
acquiring a lung scan image from the lung scan data for each of the sets of human features;
setting the pixel proportion of the lung scanning image, and setting the set pixel proportion as a first pixel proportion; in practical implementation, the first pixel ratio is set to 1280 by 720;
referring to fig. 3, a first extraction circle is set, a plurality of first extraction circles are randomly extracted from a key region, the first extraction circles are set as a plurality of extraction regions, an average value of gray values of a plurality of pixel points in the extraction regions is calculated, the first extraction gray is set, an average value of first extraction gray corresponding to the plurality of extraction regions is obtained, and the first extraction reference gray is set;
calculating the average value of a plurality of first extraction reference gray scales, and setting the average value as a second extraction reference gray scale;
obtaining minimum values and maximum values in a plurality of first extraction reference gray scales, respectively setting the minimum extraction reference gray scales and the maximum extraction reference gray scales, setting a section between the minimum extraction reference gray scales and the maximum extraction reference gray scales as an extraction reference gray scale section, dividing the extraction reference gray scale section into a plurality of extraction reference gray scale subsections, establishing a distribution histogram according to the plurality of extraction reference gray scale subsections, and setting the extraction reference gray scale subsection where the second extraction reference gray scale is located as an extraction gray scale reference section; for example, when the gray value is between 100 and 120, obtaining a plurality of extracted gray reference subintervals according to the dividing interval of 2;
the extracted gradation reference section is set as a gradation section of the key region.
The gray level generation strategy further includes a region extraction sub-strategy including: obtaining the area of a key region, multiplying the area of the key region by a first extraction proportion to obtain the area of a first extraction circle, dividing the area of the key region by the area of the first extraction circle to obtain a first reference number, multiplying the first reference number by a second extraction proportion to obtain a first extraction numerical value, and setting the integer bit of the first extraction numerical value as the first extraction reference number; the first extraction ratio is set between 5% and 10%; the second extraction ratio is set between 10% and 20%;
and extracting a corresponding number of first extraction circles in the key area according to the first extraction reference number.
The contour generation unit is used for generating a contour section of the key region based on analysis of the acquired lung scanning data, and is configured with a contour generation strategy, wherein the contour generation strategy comprises: extracting key areas of a plurality of lung scanning images according to gray intervals of the key areas, acquiring outlines of the extracted key areas, and setting the outlines as extraction outlines;
referring to fig. 4, a basic contour is obtained by referring to an existing lung simulation drawing, or according to an existing medical lung model, the basic contour is only used for basic comparison, and subsequently, calibration can be performed according to actual data, a basic contour of a key region is set, and a plurality of extracted contours are compared with the basic contour to obtain a convex region and a concave region between each extracted contour and the basic contour;
overlapping the plurality of extraction contours with the basic contour, setting the number of overlapping convex areas as the number of overlapping convex areas, and setting the number of overlapping concave areas as the number of overlapping concave areas;
the method comprises the steps of obtaining a protruding proportion by comparing protruding superposition quantity with scanning extraction quantity, obtaining a recessed proportion by comparing recessed superposition quantity with scanning extraction quantity, performing protruding correction on a protruding region corresponding to a basic contour when the protruding proportion is larger than or equal to a first protruding quantity threshold value, performing recessed correction on a recessed region corresponding to the basic contour when the recessed proportion is larger than or equal to a first recessed quantity threshold value, and setting the corrected basic contour as a contour reference region; the first protruding quantity threshold value is equal to the first recessed quantity threshold value, and is set to be 50% when the first protruding quantity threshold value is specifically set;
the contour reference region is set as a contour section of the key region.
The contour generation strategy further includes a contour syndrome strategy, the contour syndrome strategy comprising: when the protruding proportion is greater than or equal to a first protruding quantity threshold value, setting a plurality of protruding areas as protruding correction areas; when the recess ratio is greater than or equal to a first recess number threshold, setting a plurality of recess areas as recess correction areas;
calculating the average value of the areas of a plurality of convex correction areas, setting the average value as a convex area reference area, setting a convex correction area with the smallest difference value with the convex area reference area as a convex correction selection area, and adding a convex correction selection area to a basic contour to carry out convex correction;
the average value of the areas of a plurality of concave correction areas is calculated, the average value is set as a concave area reference area, the concave correction area with the smallest difference value with the concave area reference area is set as a concave correction selection area, and the basic contour is added with the concave correction selection area to carry out concave correction.
The comprehensive generation unit is used for integrating the gray scale interval and the contour interval of the generated key region to obtain a detection model of the key region.
Referring to fig. 2, the present invention further provides a method for generating a key region detection model based on lung scan data, including:
step S1, acquiring lung scanning data of a plurality of healthy persons from a scanning database;
step S2, dividing according to personnel characteristics, corresponding the divided personnel characteristics into a scanning database, and acquiring lung scanning data corresponding to the personnel characteristics from the scanning database; step S2 further includes: step S211, classifying the personnel characteristics into sex characteristics, age characteristics and weight characteristics, wherein the sex characteristics comprise men and women;
step S212, the age characteristic comprises a first age interval, and the first age interval is divided into a plurality of age sub-intervals according to the first age interval;
step S213, the weight characteristic comprises a first weight section, and the first weight section is divided into a plurality of weight sections according to the first weight interval;
step S214, combining according to the classified sex characteristics, age characteristics and weight characteristics to obtain a plurality of personnel characteristic groups, and setting the number of the personnel characteristic groups as the personnel classification number.
Step S2 further includes: step S221, according to the personnel characteristic group and the personnel characteristic in the scanning database, the quantity which can be matched with the personnel characteristic of the personnel characteristic group is obtained and is set as the extractable quantity;
step S222, obtaining the total number of lung scanning data in a scanning database, setting the total number as the extractable total number, and dividing the extractable total number by the personnel dividing number to obtain the extraction reference number;
step S223, comparing the extractable quantity with the extraction reference quantity to obtain an extractable proportion, and deleting the personnel feature group corresponding to the extractable quantity when the extractable proportion is smaller than or equal to a first proportion threshold;
step S224, setting the personnel characteristic group obtained after the deletion process as a personnel characteristic extraction group;
step S225, the first proportional threshold is multiplied by the extraction reference number to obtain a minimum extraction value, the second proportional threshold is multiplied by the extraction reference number to obtain a maximum extraction value, and the range from the minimum extraction value to the maximum extraction value is set as an extraction reference range;
in step S226, the person feature matching is performed between the person feature set and the scan database, the maximum number of lung scan data is extracted according to the extraction reference range, and the number of lung scan data acquired in each person feature set is set as the scan extraction number.
Step S3, based on analysis of the acquired lung scanning data, generating a gray scale section of the key region; step S3 further includes: step S31, generating a key area gray model of each group one by one according to the personnel characteristic groups;
step S32, acquiring a lung scanning image from lung scanning data of each personnel characteristic group;
step S33, setting the pixel proportion of the lung scanning image, and setting the set pixel proportion as a first pixel proportion;
step S34, setting a first extraction circle, randomly extracting a plurality of first extraction circles from the key area, setting the first extraction circles as a plurality of extraction areas, calculating the average value of gray values of a plurality of pixel points in the extraction areas, setting the average value of the first extraction gray values corresponding to the plurality of extraction areas, and setting the average value as a first extraction reference gray value;
step S35, an average value of a plurality of first extraction reference gray scales is obtained, and the average value is set as a second extraction reference gray scale;
step S36, obtaining minimum values and maximum values in a plurality of first extraction reference gray scales, respectively setting the minimum extraction reference gray scale and the maximum extraction reference gray scale, setting a section between the minimum extraction reference gray scale and the maximum extraction reference gray scale as an extraction reference gray scale section, dividing the extraction reference gray scale section into a plurality of extraction reference gray scale subsections, establishing a distribution histogram according to the plurality of extraction reference gray scale subsections, and setting the extraction reference gray scale subsection where the second extraction reference gray scale is located as an extraction gray scale reference section;
step S37 sets the extracted gradation reference section as the gradation section of the key region.
Step S34 further includes: step S341, obtaining the area of a key region, multiplying the area of the key region by a first extraction ratio to obtain the area of a first extraction circle, dividing the area of the key region by the area of the first extraction circle to obtain a first reference number, multiplying the first reference number by a second extraction ratio to obtain a first extraction value, and setting the integer bit of the first extraction value as the first extraction reference number;
in step S342, a corresponding number of first extraction circles are extracted in the key region according to the first extraction reference number.
Step S4, generating a contour section of the key region based on analysis of the acquired lung scanning data; step S4 further includes: step S41, extracting key areas of a plurality of lung scanning images according to gray scale intervals of the key areas, obtaining outlines of the extracted key areas, and setting the outlines as extraction outlines;
step S42, setting basic contours of key areas, and comparing a plurality of extracted contours with the basic contours to obtain convex areas and concave areas between each extracted contour and the basic contour;
step S43, overlapping a plurality of extracted contours with the basic contours, setting the number of mutually overlapped convex areas as the number of mutually overlapped convex areas, and setting the number of mutually overlapped concave areas as the number of mutually overlapped concave areas;
step S44, the protruding superposition quantity is compared with the scanning extraction quantity to obtain a protruding proportion, the recessed superposition quantity is compared with the scanning extraction quantity to obtain a recessed proportion, when the protruding proportion is greater than or equal to a first protruding quantity threshold value, protruding correction is carried out on a protruding region corresponding to the basic contour, when the recessed proportion is greater than or equal to a first recessed quantity threshold value, recess correction is carried out on a recessed region corresponding to the basic contour, and the corrected basic contour is set as a contour reference region;
in step S45, the contour reference area is set as a contour section of the key area.
Step S44 further includes: step S441, when the protruding proportion is greater than or equal to a first protruding amount threshold, setting a plurality of protruding areas as protruding correction areas; when the recess ratio is greater than or equal to a first recess number threshold, setting a plurality of recess areas as recess correction areas;
step S442, an average value of areas of a plurality of convex correction areas is obtained, the average value is set as a convex area reference area, a convex correction area with the smallest difference value with the convex area reference area is set as a convex correction selection area, and the basic outline is added with the convex correction selection area to carry out convex correction;
in step S443, the average value of the areas of the plurality of concave correction areas is obtained, the average value is set as the reference area of the concave area, the concave correction area with the smallest difference value with the reference area of the concave area is set as the concave correction selection area, and the basic contour is added with the concave correction selection area to carry out concave correction.
And S5, integrating according to the generated gray scale interval and contour interval of the key region to obtain a detection model of the key region.
It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media having computer-usable program code embodied therein. The storage medium may be implemented by any type or combination of volatile or nonvolatile Memory devices, such as static random access Memory (Static Random Access Memory, SRAM), electrically erasable Programmable Read-Only Memory (Electrically Erasable Programmable Read-Only Memory, EEPROM), erasable Programmable Read-Only Memory (Erasable Programmable Read Only Memory, EPROM), programmable Read-Only Memory (PROM), read-Only Memory (ROM), magnetic Memory, flash Memory, magnetic disk, or optical disk. These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
In the embodiments provided in the present application, it should be understood that the disclosed apparatus and method may be implemented in other manners. The above-described apparatus embodiments are merely illustrative, for example, the division of the units is merely a logical function division, and there may be other manners of division in actual implementation, and for example, multiple units or components may be combined or integrated into another system, or some features may be omitted, or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be through some communication interface, device or unit indirect coupling or communication connection, which may be in electrical, mechanical or other form.
Claims (5)
1. The key region detection model generation system based on the lung scanning data is characterized by comprising a scanning database, a personnel dividing module and a key region generation module; the lung scanning data of a plurality of healthy persons are stored in the scanning database;
the personnel division module is configured with a personnel division extraction strategy, and the personnel division extraction strategy comprises: dividing according to the personnel characteristics, corresponding the divided personnel characteristics into a scanning database, and acquiring lung scanning data corresponding to the personnel characteristics from the scanning database;
the key region generation module comprises a gray level generation unit, a contour generation unit and a comprehensive generation unit, wherein the gray level generation unit is used for generating a gray level section of a key region based on analysis of acquired lung scanning data, the contour generation unit is used for generating a contour section of the key region based on analysis of the acquired lung scanning data, and the comprehensive generation unit is used for integrating the generated gray level section and the contour section of the key region to obtain a detection model of the key region;
the personnel division extraction strategy further comprises a personnel division sub-strategy, and the personnel division sub-strategy comprises: classifying the person features into gender features, age features and weight features, the gender features including men and women;
the age characteristics comprise a first age interval, and the first age interval is divided into a plurality of age sub-intervals according to the first age interval;
the weight characteristic comprises a first weight interval, and the first weight interval is divided into a plurality of weight sub-intervals according to the first weight interval;
combining according to the classified sex characteristics, age characteristics and weight characteristics to obtain a plurality of personnel characteristic groups, and setting the number of the personnel characteristic groups as the classifying number of personnel;
the gradation generation unit is configured with a gradation generation policy including: generating a key region gray scale model of each group one by one according to the personnel characteristic groups;
acquiring a lung scan image from the lung scan data for each of the sets of human features;
setting the pixel proportion of the lung scanning image, and setting the set pixel proportion as a first pixel proportion;
setting a first extraction circle, randomly extracting a plurality of first extraction circles from the key area, setting the first extraction circles as a plurality of extraction areas, calculating the average value of gray values of a plurality of pixel points in the extraction areas, setting the average value as first extraction gray, calculating the average value of the first extraction gray corresponding to the plurality of extraction areas, and setting the average value as first extraction reference gray;
calculating the average value of a plurality of first extraction reference gray scales, and setting the average value as a second extraction reference gray scale;
obtaining minimum values and maximum values in a plurality of first extraction reference gray scales, respectively setting the minimum extraction reference gray scales and the maximum extraction reference gray scales, setting a section between the minimum extraction reference gray scales and the maximum extraction reference gray scales as an extraction reference gray scale section, dividing the extraction reference gray scale section into a plurality of extraction reference gray scale subsections, establishing a distribution histogram according to the plurality of extraction reference gray scale subsections, and setting the extraction reference gray scale subsection where the second extraction reference gray scale is located as an extraction gray scale reference section;
setting the extracted gray scale reference interval as a gray scale interval of a key area;
the contour generation unit is configured with a contour generation policy comprising: extracting key areas of a plurality of lung scanning images according to gray intervals of the key areas, acquiring outlines of the extracted key areas, and setting the outlines as extraction outlines;
setting basic contours of key areas, and comparing a plurality of extracted contours with the basic contours to obtain convex areas and concave areas between each extracted contour and the basic contour;
overlapping the plurality of extraction contours with the basic contour, setting the number of overlapping convex areas as the number of overlapping convex areas, and setting the number of overlapping concave areas as the number of overlapping concave areas;
the method comprises the steps of obtaining a protruding proportion by comparing protruding superposition quantity with scanning extraction quantity, obtaining a recessed proportion by comparing recessed superposition quantity with scanning extraction quantity, performing protruding correction on a protruding region corresponding to a basic contour when the protruding proportion is larger than or equal to a first protruding quantity threshold value, performing recessed correction on a recessed region corresponding to the basic contour when the recessed proportion is larger than or equal to a first recessed quantity threshold value, and setting the corrected basic contour as a contour reference region;
the contour reference region is set as a contour section of the key region.
2. The pulmonary scan data-based accent region detection model generation system of claim 1, wherein the staff member partitioning extraction policy further comprises an extraction sub-policy comprising: comparing the personnel characteristic group with personnel characteristics in a scanning database to obtain the quantity which can be matched with the personnel characteristics of the personnel characteristic group, and setting the quantity as the extractable quantity;
acquiring the total number of lung scanning data in a scanning database, setting the total number as the extractable total number, and dividing the extractable total number by the personnel dividing number to obtain the extraction reference number;
comparing the extractable quantity with the extraction reference quantity to obtain an extractable proportion, and deleting the personnel feature group corresponding to the extractable quantity when the extractable proportion is smaller than or equal to a first proportion threshold value;
setting a personnel characteristic group obtained after deletion processing as a personnel characteristic extraction group;
multiplying the first proportional threshold value by the extraction reference number to obtain a minimum extraction value, multiplying the second proportional threshold value by the extraction reference number to obtain a maximum extraction value, and setting a range from the minimum extraction value to the maximum extraction value as an extraction reference range;
and performing personnel feature matching on the personnel feature groups and the scanning database, extracting the maximum number of lung scanning data according to the extraction reference range, and setting the number of lung scanning data acquired in each personnel feature group as the scanning extraction number.
3. The pulmonary scan data-based accent region detection model generation system of claim 2, wherein the gray scale generation strategy further comprises a region extraction sub-strategy comprising: obtaining the area of a key region, multiplying the area of the key region by a first extraction proportion to obtain the area of a first extraction circle, dividing the area of the key region by the area of the first extraction circle to obtain a first reference number, multiplying the first reference number by a second extraction proportion to obtain a first extraction numerical value, and setting the integer bit of the first extraction numerical value as the first extraction reference number;
and extracting a corresponding number of first extraction circles in the key area according to the first extraction reference number.
4. The pulmonary scan data based accent region detection model generation system of claim 3, wherein the contour generation strategy further comprises a contour syndrome strategy comprising: when the protruding proportion is greater than or equal to a first protruding quantity threshold value, setting a plurality of protruding areas as protruding correction areas; when the recess ratio is greater than or equal to a first recess number threshold, setting a plurality of recess areas as recess correction areas;
calculating the average value of the areas of a plurality of convex correction areas, setting the average value as a convex area reference area, setting a convex correction area with the smallest difference value with the convex area reference area as a convex correction selection area, and adding a convex correction selection area to a basic contour to carry out convex correction;
the average value of the areas of a plurality of concave correction areas is calculated, the average value is set as a concave area reference area, the concave correction area with the smallest difference value with the concave area reference area is set as a concave correction selection area, and the basic contour is added with the concave correction selection area to carry out concave correction.
5. A method for adapting a pulmonary scan data based accent region detection model generation system according to any one of claims 1 to 4, comprising:
acquiring lung scanning data of a plurality of healthy persons from a scanning database;
dividing according to the personnel characteristics, corresponding the divided personnel characteristics into a scanning database, and acquiring lung scanning data corresponding to the personnel characteristics from the scanning database;
based on analysis of the acquired lung scanning data, generating a gray scale section of the key region;
generating a contour section of the key region based on analysis of the acquired lung scan data;
and integrating according to the generated gray scale interval and contour interval of the key region to obtain a detection model of the key region.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310493252.1A CN116205920B (en) | 2023-05-05 | 2023-05-05 | Method and system for generating key region detection model based on lung scanning data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310493252.1A CN116205920B (en) | 2023-05-05 | 2023-05-05 | Method and system for generating key region detection model based on lung scanning data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116205920A CN116205920A (en) | 2023-06-02 |
| CN116205920B true CN116205920B (en) | 2023-07-18 |
Family
ID=86515041
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310493252.1A Active CN116205920B (en) | 2023-05-05 | 2023-05-05 | Method and system for generating key region detection model based on lung scanning data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116205920B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103886576A (en) * | 2013-11-22 | 2014-06-25 | 沈阳东软医疗系统有限公司 | Glandular tissue characteristic gray scale detection method and device |
| CN104083170A (en) * | 2014-07-04 | 2014-10-08 | 深圳安科高技术股份有限公司 | Method and system for automatically extracting magnetic resonance image corpus callosum |
| CN106548173A (en) * | 2016-11-24 | 2017-03-29 | 国网山东省电力公司电力科学研究院 | A kind of improvement no-manned plane three-dimensional information getting method based on classification matching strategy |
| CN106558045A (en) * | 2016-10-20 | 2017-04-05 | 上海联影医疗科技有限公司 | A kind of segmentation of lung parenchyma method, device, magic magiscan |
| CN114119447A (en) * | 2020-08-31 | 2022-03-01 | 上海宽带技术及应用工程研究中心 | Method, system, medium and device for constructing lung nodule multi-attribute classification model |
| CN115049807A (en) * | 2022-06-28 | 2022-09-13 | 上海市胸科医院 | Method and device for establishing pulmonary blood vessel model and server |
| CN115661149A (en) * | 2022-12-26 | 2023-01-31 | 天津医科大学总医院 | Lung image processing system based on lung tissue data |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP5258694B2 (en) * | 2009-07-27 | 2013-08-07 | 富士フイルム株式会社 | Medical image processing apparatus and method, and program |
| US20220180975A1 (en) * | 2019-01-28 | 2022-06-09 | The Broad Institute, Inc. | Methods and systems for determining gene expression profiles and cell identities from multi-omic imaging data |
-
2023
- 2023-05-05 CN CN202310493252.1A patent/CN116205920B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103886576A (en) * | 2013-11-22 | 2014-06-25 | 沈阳东软医疗系统有限公司 | Glandular tissue characteristic gray scale detection method and device |
| CN104083170A (en) * | 2014-07-04 | 2014-10-08 | 深圳安科高技术股份有限公司 | Method and system for automatically extracting magnetic resonance image corpus callosum |
| CN106558045A (en) * | 2016-10-20 | 2017-04-05 | 上海联影医疗科技有限公司 | A kind of segmentation of lung parenchyma method, device, magic magiscan |
| CN106548173A (en) * | 2016-11-24 | 2017-03-29 | 国网山东省电力公司电力科学研究院 | A kind of improvement no-manned plane three-dimensional information getting method based on classification matching strategy |
| CN114119447A (en) * | 2020-08-31 | 2022-03-01 | 上海宽带技术及应用工程研究中心 | Method, system, medium and device for constructing lung nodule multi-attribute classification model |
| CN115049807A (en) * | 2022-06-28 | 2022-09-13 | 上海市胸科医院 | Method and device for establishing pulmonary blood vessel model and server |
| CN115661149A (en) * | 2022-12-26 | 2023-01-31 | 天津医科大学总医院 | Lung image processing system based on lung tissue data |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116205920A (en) | 2023-06-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8634610B2 (en) | System and method for assessing cancer risk | |
| JP4702971B2 (en) | Computer-aided diagnosis system | |
| CN111539944B (en) | Method, device, electronic equipment and storage medium for acquiring statistical attribute of lung focus | |
| US20110019886A1 (en) | Medical image processing apparatus, method, and program | |
| WO2005023086B1 (en) | Systems, methods, and computer program products for analysis of vessel attributes for diagnosis, disease staging, and surgical planning | |
| CN109919254B (en) | Breast density classification method, system, readable storage medium and computer device | |
| CA2613746A1 (en) | Non-invasive joint evaluation | |
| Aresta et al. | Automatic lung nodule detection combined with gaze information improves radiologists’ screening performance | |
| KR102516154B1 (en) | Medical image analysis system and similar case retrieval system using quantified parameters, and method for the same | |
| CN116312986A (en) | Three-dimensional medical image labeling method and device, electronic equipment and readable storage medium | |
| CN115880159A (en) | Method and computer readable storage medium for CT perfusion image parameter map correction | |
| US11935234B2 (en) | Method for detecting abnormality, non-transitory computer-readable recording medium storing program for detecting abnormality, abnormality detection apparatus, server apparatus, and method for processing information | |
| CN116205920B (en) | Method and system for generating key region detection model based on lung scanning data | |
| Soni et al. | Automatic Pulmonary Cancer Detection using Prewitt & Morphological Dilation | |
| US20220036985A1 (en) | Lung analysis and reporting system | |
| CN115399874A (en) | Optimal implantation path planning method and device for pedicle screw and electronic equipment | |
| Vance | The influence of ancestry, sex, and age on the morphology of the frontal sinus in black and white individuals | |
| CN111627554A (en) | Fracture image automatic classification system based on deep convolutional neural network | |
| CN111261285A (en) | Display method, interaction method and storage medium of diagnostic information interface | |
| CN111242930A (en) | Method and device for acquiring pulmonary tuberculosis recognition model, storage medium and processor | |
| CN116612193B (en) | Angiography image processing and calibrating method based on deep learning | |
| CN116958456B (en) | Heart three-dimensional model construction method, system and storage medium based on image registration | |
| Solves Llorens et al. | MRI skin segmentation for the virtual deformation of the breast under mammographic compression | |
| Soni et al. | Atomatic Lung Cancer Detection using Sobel & Morphological Operations | |
| Gong et al. | Trail-Traced Threshold Test (T4) with a Weighted Binomial Distribution for a Psychophysical Test |
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 |