CN103955610B - Medical image computer-aided analysis method - Google Patents

Medical image computer-aided analysis method Download PDF

Info

Publication number
CN103955610B
CN103955610B CN201410172988.XA CN201410172988A CN103955610B CN 103955610 B CN103955610 B CN 103955610B CN 201410172988 A CN201410172988 A CN 201410172988A CN 103955610 B CN103955610 B CN 103955610B
Authority
CN
China
Prior art keywords
data
image
medical image
voxel
sample
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201410172988.XA
Other languages
Chinese (zh)
Other versions
CN103955610A (en
Inventor
韩燕�
董蒨
常晓峰
魏宾
张鲲鹏
牛海涛
朱呈瞻
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.)
Affiliated Hospital of University of Qingdao
Original Assignee
Affiliated Hospital of University of Qingdao
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 Affiliated Hospital of University of Qingdao filed Critical Affiliated Hospital of University of Qingdao
Priority to CN201410172988.XA priority Critical patent/CN103955610B/en
Publication of CN103955610A publication Critical patent/CN103955610A/en
Application granted granted Critical
Publication of CN103955610B publication Critical patent/CN103955610B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Landscapes

  • Measuring And Recording Apparatus For Diagnosis (AREA)
  • Image Processing (AREA)

Abstract

The invention provides a medical image computer-aided analysis method. The medical image computer-aided analysis method comprises the following steps: carrying out primary scanning on a medical image through an infrared light source and an infrared imager, and carrying out image division operation on volume data of the image through the primary scanning so as to obtain primary image data; carrying out secondary scanning on the medical image through a visible light source and a photosensitive element; carrying out image division operation on volume data of the image through the secondary scanning so as to obtain secondary image data; binding the primary image data with the secondary image data, and contrasting with an existing pathological sample in a pathological database, wherein the contrast process specifically comprises mutually contrasting the primary image data and mutually contrasting the secondary image data.

Description

A kind of medical image computer-aided analysis method
Technical field
The present invention relates to medical instruments field, more particularly to a kind of medical image computer-aided analysis method.
Background technology
Computer aided detection is using advanced computer software and hardware analysis and process digital radiation image, to find simultaneously Detection characteristics of lesion, its result is referred to as " second suggestion " for diagnostician, helps radiation technician to improve Lesion Detection rate, It is referred to as " second eyes " of radiologist, it can improve diagnostic accuracy and improve the repeatability of diagnosis, shortens and read The piece time, improve operating efficiency.
Conventional computer aided detection method is method for feature analysis and temporal subtraction.
1st, method for feature analysis:Certain variation characteristic is extracted by post processing of image, is exported again with after corresponding pattern match The method of diagnostic result.Method for feature analysis is divided into selection region of interest, extracts the steps such as feature, pattern match and medical diagnosis on disease.Choosing After taking region of interest, using various post processing of image means to selected extracting section feature, according to characteristics of lesion to it is related Disease performance carries out pattern match and medical diagnosis on disease.
2nd, temporal subtraction:Two width images before and after to morbidity are mutually reduced computing, in removing two width images Identical part, retains different piece, the substantially prominent abnormal change under homogeneous background.
Method for feature analysis is to pre-build pathological model, the pathological regions of target image is extracted, by characteristics of lesion and pathology Model is matched.Because pathological model is fixed model, or the model library being made up of multiple fixed models, and pathology is special Levy and vary, diagnosed according to matching degree merely, situation about judging by accident can occur unavoidably.
And temporal subtraction is mutually to be reduced computing for two width images before and after morbidity, rather than to single width figure As being analyzed comparison, if patient only has single image, the method failure.
Above two method, is required for carrying out image procossing and data processing, specifically includes three below step:
(1) intensive treatment of focus shade.At present, mainly being filtered using spatial filtering and spatial frequency makes focus contrast Increase, the part contrast beyond focus is reduced, so that focus is easily detected.
(2) extraction of focus candidate shade.Focus shade is detected according to threshold values process and the characteristic quantity of focus is measured, is judged Whether it is selected focus.The characteristic quantity yardstick of shade, such as effective diameter, the circle about its characteristic of switch amount are determined first Degree, degree of irregularity etc., using circularity is big, degree of irregularity is little shade as the scale for extracting inactivity of yang-qi shadow, then through multiple valve Value process, obtains black and white contrast image.
(3) focus candidate's shaded region (reduction false positive) is reduced.Select to reflect that focus is cloudy using image recognition technology The characteristic quantity of shadow, according to characteristic quantity focus and false positive shade are recognized, characteristic quantity is more, false positive number occurs fewer.As tied 10 features above amounts such as effective diameter, circularity, density, then carry out diagnostic classification to characteristic quantity used in section shadow, commonly use Method artificial neural network method, fuzzy clustering algorithm, linear classification etc..
In above-mentioned steps, the intensive treatment step of focus shade is most important, is following two diagnostic process steps Basis, if deviation occurs in the step, deviation can cause chain deviation in follow-up two steps, and this deviation is even put Greatly, mistaken diagnosis is caused.
Therefore, how to set up a complete pathology storehouse, store pathological model as much as possible, improve the strong of focus shade Change treatment effect, be current problem demanding prompt solution.
The content of the invention
The present invention proposes a kind of medical image computer-aided analysis method, by setting up and constantly updates pathological data Storehouse, medical image to be diagnosed mutually is compared with existing pathology sample, improves the correctness of auxiliary diagnosis, solves existing meter Calculation machine aided diagnosis method pathological model is single, focus shade has error.
The technical scheme is that what is be achieved in that:
A kind of medical image computer-aided analysis method, comprises the following steps:
Step (1), carries out single pass, to single pass figure by infrared light supply and infrared thermography to medical image The volume data of picture carries out image splitting operation, obtains a view data;
Described image splitting operation specifically includes following steps:
Step (a), creates a division region child node set setNodes, and one with the number of volume data formed objects According to region NodeArray, the data area NodeArray is used to depositing the address of each voxel place child node, and by number Sky is set to according to the data content of region NodeArray;
Step (b), travels through successively each tissue points Voxel (xi, yi, zi) in volume data, after traversal is completed, jumps Go to step (g);
Step (c), judges whether Voxel (xi, yi, zi) had been expanded;Determination methods are corresponding to see its Whether the value in NodeArray (xi, yi, zi) is empty;If propagating through, then return to step (b);
Step (d), creates a sub-regions node Nodej, and the original position of the node is Voxel (xi, yi, zi);
Step (e), is extended respectively to the positive direction of tri- reference axis of X, Y, Z, and judge the voxel of new extension with Whether Voxel (xi, yi, zi) has feature consistency;When inconsistent voxel occurs in a direction, then stop the direction Extension;
Step (f), can obtain end position Voxel (xi+m, yi+n, the zi+ of node Nodej after the completion of the step (e) k);The value of NodeArray (xs, ys, zs) is set to corresponding to the voxel Voxel (xs, ys, zs) that first Nodej is included The address of Nodej, then be added in child node set setNodes;
Step (g), in volume data has been traveled through after all voxels, by the data in the NodeArray of ergodic data region, Generate the neighbouring relations of sub-district domain node;
Step (2), rescan is carried out by visible light source and photo-sensitive cell to the medical image;To rescan The volume data of image carries out described image splitting operation, obtains secondary image data;
Step (3), a view data and secondary image data is bound, with existing pathology in pathological data storehouse Sample is mutually compared, and comparison process is specially a view data and compares mutually, and secondary image data is compared mutually;
Step (4), a view data comparison result r1 is multiplied by coefficient of first order k1, and secondary image data comparison result r2 takes advantage of With quadratic coefficients k2, k1+k2=1, and 0.2≤k1≤0.3,0.7≤k1≤0.8;If r1*k1+r2*k2 >=0.5, judge new Pathology sample is pathology sample, and pathological data storehouse described in typing;If 0.05≤r1*k1+r2*k2 < 0.5, new pathology is judged Sample is suspected lesion sample;If r1*k1+r2*k2 < 0.05, judge pathology sample for normal sample.
Alternatively, it is another positioned at the medical image by the inswept medical image of column infrared light supply in the step (1) The infrared thermography of side carries out induction image forming to the infrared signal after the filtration of medical image film, is once swept Tracing picture.
Alternatively, in the step (2), by the inswept medical image of column visible light source, positioned at the medical image The photo-sensitive cell of opposite side carries out induction image forming to the visible light signal after the filtration of medical image film, obtains secondary sweeping Tracing picture.
Alternatively, in the step (3), in the step of existing pathology sample is mutually compared in pathological data storehouse, also wrap The step of decompressing to existing pathology sample is included, is decompressed as a view data and secondary image data.
The invention has the beneficial effects as follows:
(1) accurately, focal size measure error is little for lesion localization;
(2) and a set of complete pathological data storehouse is set up, and constantly data storehouse is updated, improve the standard of diagnosis True property.
Description of the drawings
In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing The accompanying drawing to be used needed for having technology description is briefly described, it should be apparent that, drawings in the following description are only this Some embodiments of invention, for those of ordinary skill in the art, on the premise of not paying creative work, can be with Other accompanying drawings are obtained according to these accompanying drawings.
Fig. 1 is the flow chart of medical image computer aided diagnosing method of the present invention;
Fig. 2 is the flow chart of image splitting operation process in medical image computer aided diagnosing method of the present invention.
Specific embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete Site preparation is described, it is clear that described embodiment is only a part of embodiment of the invention, rather than the embodiment of whole.It is based on Embodiment in the present invention, it is every other that those of ordinary skill in the art are obtained under the premise of creative work is not made Embodiment, belongs to the scope of protection of the invention.
Computer-assisted analysis has become reference frame important during diagnosis, but existing area of computer aided Diagnosis is all that image procossing and data processing, meeting after medical image film printing are carried out while medical image shooting process Have and reported with piece, doctor makes final judgement according to report and the observation of oneself.If patient is to other hospital admissions, doctor Can only diagnose by rule of thumb, or patient needs to carry out a filming image again, is all one to the body or economy of patient Plant loss.
The medical image computer-aided analysis method of the present invention, is entered by infrared light and visible ray to medical image film Row twice sweep, then carries out image splitting operation to scan image, obtain twice sweep data respectively with pathology storehouse in disease Reason sample is mutually compared, and comparison result is multiplied by weight and obtains last diagnostic result.Below in conjunction with the accompanying drawings, the method for the present invention is carried out Elaborate.
As shown in figure 1, the medical image computer-aided analysis method of the present invention, comprises the following steps:
Step (1), carries out single pass, to single pass figure by infrared light supply and infrared thermography to medical image The volume data of picture carries out image splitting operation, obtains a view data.Quilt during the infrared light image film of infrared light supply Filter, infrared light is different with the transmitance of other parts in focus, therefore the imaging on infrared thermography is different.Once sweep Retouching can strengthen the positioning to lesions position, and to focal size preliminary surveying is carried out.
As shown in Fig. 2 described image splitting operation specifically includes following steps:
Step (a), creates a division region child node set setNodes, and one with the number of volume data formed objects According to region NodeArray, the NodeArray is used to deposit the address of each voxel place child node, and by NodeArray's Data content is set to sky;
Step (b), travels through successively each tissue points Voxel (xi, yi, zi) in volume data, after traversal is completed, jumps Go to step (g);
Step (c), judges whether Voxel (xi, yi, zi) had been expanded;Determination methods are corresponding to see its Whether the value in NodeArray (xi, yi, zi) is empty;If propagating through, then return to step (b);
Step (d), creates a sub-regions node Nodej, and the original position of the node is Voxel (xi, yi, zi);
Step (e), is extended respectively to the positive direction of tri- reference axis of X, Y, Z, and judge the voxel of new extension with Whether Voxel (xi, yi, zi) has feature consistency;When inconsistent voxel occurs in a direction, then stop the direction Extension;
Step (f), can obtain end position Voxel (xi+m, yi+n, the zi+ of node Nodej after the completion of the step (e) k);The value of NodeArray (xs, ys, zs) is set to corresponding to the voxel Voxel (xs, ys, zs) that first Nodej is included The address of Nodej, then be added in child node set setNodes;
Step (g), in volume data has been traveled through after all voxels, by the data in traversal NodeArray, generates sub-district The neighbouring relations of domain node.
The image splitting operation process of the present invention is only by traveling through twice volume data, it is possible to enter line splitting to volume data, And the region after dividing is not in the situation of Hypersegmentation.
Flow chart shown in Fig. 1 is returned to, the method for the present invention also includes:
Step (2), rescan is carried out by visible light source and photo-sensitive cell to the medical image;To rescan The volume data of image carries out described image splitting operation, obtains secondary image data.Shown in image splitting operation process and Fig. 2 Process is identical, repeats no more here.Filtered during visible light-transmissive image film, it is seen that sharpening of the light at focus edge is imitated It is really good, it is possible to increase focus contrast, image flatness is improved, obtain accurate focal size.
Step (3), a view data and secondary image data is bound, with existing pathology in pathological data storehouse Sample is mutually compared, and comparison process is specially a view data and compares mutually, and secondary image data is compared mutually.
Step (4), a view data comparison result r1 is multiplied by coefficient of first order k1, and secondary image data comparison result r2 takes advantage of With quadratic coefficients k2, k1+k2=1, and k1≤0.3, k2 >=0.7;If r1*k1+r2*k2 >=0.5, judge that new pathology sample is Pathology sample, and typing pathological data storehouse;If 0.05≤r1*k1+r2*k2 < 0.5, judge new pathology sample for suspected lesion Sample;If r1*k1+r2*k2 < 0.05, judge pathology sample for normal sample.
Preferably, it is another positioned at the medical image by the inswept medical image of column infrared light supply in the step (1) The infrared thermography of side carries out induction image forming to the infrared signal after the filtration of medical image film, is once swept Tracing picture.
Preferably, in the step (2), by the inswept medical image of column visible light source, positioned at the medical image The photo-sensitive cell of opposite side carries out induction image forming to the visible light signal after the filtration of medical image film, obtains secondary sweeping Tracing picture.
Preferably, in the step (3), in the step of existing pathology sample is mutually compared in pathological data storehouse, also wrap The step of decompressing to existing pathology sample is included, is decompressed as a view data and secondary image data.
The medical image computer-aided analysis method of the present invention, accurately, focal size measure error is little for lesion localization;And And a set of complete pathological data storehouse is set up, and constantly data storehouse is updated, improve the accuracy of diagnosis.
Presently preferred embodiments of the present invention is the foregoing is only, not to limit the present invention, all essences in the present invention Within god and principle, any modification, equivalent substitution and improvements made etc. should be included within the scope of the present invention.

Claims (3)

1. a kind of medical image computer-aided analysis method, it is characterised in that comprise the following steps:
Step (1), carries out single pass, to single pass image by infrared light supply and infrared thermography to medical image Volume data carries out image splitting operation, obtains a view data;
Described image splitting operation specifically includes following steps:
Step (a), creates a division region child node set setNodes, and one with the data field of volume data formed objects Domain NodeArray, the data area NodeArray are used to depositing the address of each voxel place child node, and by data field The data content of domain NodeArray is set to sky;
Step (b), travels through successively each tissue points Voxel (xi, yi, zi) in volume data, after traversal is completed, jumps to Step (g);
Step (c), judges whether Voxel (xi, yi, zi) had been expanded;Determination methods are to see its correspondence NodeArray Whether the value in (xi, yi, zi) is empty;If propagating through, then return to step (b);
Step (d), creates a sub-regions node Nodej, and the original position of the node is Voxel (xi, yi, zi);
Step (e), is extended respectively to the positive direction of tri- reference axis of X, Y, Z, and judges the voxel and Voxel of new extension Whether (xi, yi, zi) has feature consistency;When inconsistent voxel occurs in a direction, then stop the extension of the direction;
Step (f), can obtain the end position Voxel (xi+m, yi+n, zi+k) of node Nodej after the completion of the step (e); The value of NodeArray (xs, ys, zs) is set to Nodej corresponding to the voxel Voxel (xs, ys, zs) that first Nodej is included Address, then be added in child node set setNodes;
Step (g), in volume data has been traveled through after all voxels, by the data in the NodeArray of ergodic data region, generates The neighbouring relations of sub-district domain node;
Step (2), rescan is carried out by visible light source and photo-sensitive cell to the medical image;To rescan image Volume data carry out described image splitting operation, obtain secondary image data;
Step (3), a view data and secondary image data is bound, with existing pathology sample in pathological data storehouse Mutually compare, comparison process is specially a view data and compares mutually, and secondary image data is compared mutually;In pathological data storehouse In the step of existing pathology sample is mutually compared, also include the step of decompressing to existing pathology sample, decompress once to scheme As data and secondary image data;
Step (4), a view data comparison result r1 is multiplied by coefficient of first order k1, and secondary image data comparison result r2 is multiplied by two Ordered coefficients k2, k1+k2=1, and 0.2≤k1≤0.3,0.7≤k1≤0.8;If r1*k1+r2*k2 >=0.5, new pathology is judged Sample is pathology sample, and pathological data storehouse described in typing;If 0.05≤r1*k1+r2*k2 < 0.5, new pathology sample is judged For suspected lesion sample;If r1*k1+r2*k2 < 0.05, judge pathology sample for normal sample.
2. medical image computer-aided analysis method as claimed in claim 1, it is characterised in that in the step (1), leads to The inswept medical image of column infrared light supply is crossed, the infrared thermography positioned at the medical image opposite side is to through medical image Infrared signal after film is filtered carries out induction image forming, obtains single pass image.
3. medical image computer-aided analysis method as claimed in claim 1, it is characterised in that in the step (2), leads to The inswept medical image of column visible light source is crossed, the photo-sensitive cell positioned at the medical image opposite side is to through medical image glue Visible light signal after piece is filtered carries out induction image forming, obtains rescan image.
CN201410172988.XA 2014-04-22 2014-04-22 Medical image computer-aided analysis method Active CN103955610B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201410172988.XA CN103955610B (en) 2014-04-22 2014-04-22 Medical image computer-aided analysis method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201410172988.XA CN103955610B (en) 2014-04-22 2014-04-22 Medical image computer-aided analysis method

Publications (2)

Publication Number Publication Date
CN103955610A CN103955610A (en) 2014-07-30
CN103955610B true CN103955610B (en) 2017-04-26

Family

ID=51332885

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201410172988.XA Active CN103955610B (en) 2014-04-22 2014-04-22 Medical image computer-aided analysis method

Country Status (1)

Country Link
CN (1) CN103955610B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105046111B (en) * 2015-09-10 2018-09-25 济南市儿童医院 A kind of Amplitude integrated electroencephalogram result automatic recognition system
CN108109682A (en) * 2018-01-19 2018-06-01 大连外国语大学 A kind of medical image identifying system and its method

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103282790A (en) * 2010-12-21 2013-09-04 皇家飞利浦电子股份有限公司 Fast dual contrast mr imaging
CN103608839A (en) * 2011-03-28 2014-02-26 皇家飞利浦有限公司 Contrast-dependent resolution image

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7058210B2 (en) * 2001-11-20 2006-06-06 General Electric Company Method and system for lung disease detection
US10098563B2 (en) * 2006-11-22 2018-10-16 Toshiba Medical Systems Corporation Magnetic resonance imaging apparatus

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103282790A (en) * 2010-12-21 2013-09-04 皇家飞利浦电子股份有限公司 Fast dual contrast mr imaging
CN103608839A (en) * 2011-03-28 2014-02-26 皇家飞利浦有限公司 Contrast-dependent resolution image

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
基于图像特征检索和识别的医学影像辅助诊断系统;刘俊 等;《中国科技信息》;20080215(第4期);第166-167页 *

Also Published As

Publication number Publication date
CN103955610A (en) 2014-07-30

Similar Documents

Publication Publication Date Title
Yun et al. Improvement of fully automated airway segmentation on volumetric computed tomographic images using a 2.5 dimensional convolutional neural net
CN109741346A (en) Area-of-interest exacting method, device, equipment and storage medium
CN108986891A (en) Medical imaging processing method and processing device, electronic equipment and storage medium
CN103994786B (en) Image detecting method for arc ruler lines of pointer instrument scale
CN106408609A (en) Parallel mechanism end motion pose detection method based on binocular vision
CN111553892B (en) Lung nodule segmentation calculation method, device and system based on deep learning
CN105640577A (en) Method and system automatically detecting local lesion in radiographic image
CN110782489B (en) Image data matching method, device and equipment and computer readable storage medium
DE102007003260A1 (en) System and method for marking and identifying lymph nodes in medical images
Pechaud et al. Extraction of tubular structures over an orientation domain
CN102693353A (en) Method and computer system for automatically generating a statistical model
CN108052909B (en) Thin fiber cap plaque automatic detection method and device based on cardiovascular OCT image
CN111292386A (en) CT projection metal trace completion metal artifact correction method based on U-net
CN110400299A (en) A kind of method and device of lung's pleural effusion detection
CN104732520A (en) Cardio-thoracic ratio measuring algorithm and system for chest digital image
CN115512110A (en) Medical image tumor segmentation method related to cross-modal attention mechanism
CN110327071A (en) CT scan method, apparatus, equipment and storage medium
CN103955610B (en) Medical image computer-aided analysis method
CN113240606B (en) Traditional Chinese medicine inspection image color correction method and system
CN114612318A (en) Three-dimensional modeling method, system and equipment based on cultural relic CT image contour line
CN109087357A (en) Scan orientation method, apparatus, computer equipment and computer readable storage medium
CN112990266A (en) Method, device, equipment and storage medium for processing multi-modal brain image data
US10331936B2 (en) Duplicate pattern reconstructions
US10217248B2 (en) Method for removing streak from detector cell with performance difference
CN107392948A (en) A kind of method for registering images of point of amplitude real-time polarization imaging system

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
CB03 Change of inventor or designer information

Inventor after: Han Yan

Inventor after: Dong San

Inventor after: Chang Xiaofeng

Inventor after: Wei Bin

Inventor after: Zhang Kunpeng

Inventor after: Niu Haitao

Inventor after: Zhu Chengzhan

Inventor before: Han Yan

CB03 Change of inventor or designer information
TA01 Transfer of patent application right

Effective date of registration: 20170321

Address after: 266000 Shandong city of Qingdao province Jiangsu City Road No. 16

Applicant after: Affiliated Hospital of University Of Qingdao

Address before: 266000 Shandong city of Qingdao province Jiangsu City Road No. 16

Applicant before: Han Yan

TA01 Transfer of patent application right
GR01 Patent grant
GR01 Patent grant