CN106898012A - CT images thoracic cavity profile automated detection method - Google Patents

CT images thoracic cavity profile automated detection method Download PDF

Info

Publication number
CN106898012A
CN106898012A CN201710039589.XA CN201710039589A CN106898012A CN 106898012 A CN106898012 A CN 106898012A CN 201710039589 A CN201710039589 A CN 201710039589A CN 106898012 A CN106898012 A CN 106898012A
Authority
CN
China
Prior art keywords
images
image
profile
thoracic cavity
detection method
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.)
Pending
Application number
CN201710039589.XA
Other languages
Chinese (zh)
Inventor
于金辉
吴凯琳
张秉炜
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.)
Zhejiang University ZJU
Original Assignee
Zhejiang University ZJU
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 Zhejiang University ZJU filed Critical Zhejiang University ZJU
Priority to CN201710039589.XA priority Critical patent/CN106898012A/en
Publication of CN106898012A publication Critical patent/CN106898012A/en
Pending legal-status Critical Current

Links

Classifications

    • 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
    • G06T2207/10—Image acquisition modality
    • G06T2207/10016—Video; Image sequence
    • 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
    • G06T2207/10—Image acquisition modality
    • G06T2207/10072—Tomographic images
    • G06T2207/10081—Computed x-ray tomography [CT]
    • 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
    • G06T2207/30008—Bone

Landscapes

  • Apparatus For Radiation Diagnosis (AREA)

Abstract

The invention discloses a kind of CT images thoracic cavity profile automated detection method, the method is that a series of cross section scanning computed tomography images are detected automatically from top to bottom to the human body of user input, remove redundant information, obtain an intrathoracic closed curve, find the Internal periphery in thoracic cavity to external expansion using active contour model as initial profile using the closed curve, automatic decision goes out thoracic cavity profile.Method of the present invention combination torso model feature carries out automatic detection, strong interference immunity, and the profile of the method extraction is continuous, convenient and swift, simple to operate, and a series of CT images can be processed, and has efficiency higher compared to traditional determination methods.

Description

CT images thoracic cavity profile automated detection method
Technical field
The invention belongs to medical image picture Processing Technique field, it is related to the thoracic cavity contour detecting of medical image and its auxiliary A kind of diagnosis, and in particular to CT images thoracic cavity profile automated detection method.
Background technology
With the development of image processing techniques, area of computer aided CT picture processings are applied to doctor more and more widely In treatment.Medical diagnosis at this stage is checked at each position human body commonly using CT images, it will usually cross-section using chest The CT images in face are diagnosed to internal body organs.Normal breast CT aspects are more, the image that each deck structure is showed Difference, internal is strong to the interference that thorax is detected.
In the related image procossing of computer, there are many methods to the detection of profile.Traditional contour extraction method is main There are edge detection algorithm and mathematics morphology.The simple speed of edge detection method is fast, but the profile for extracting is not necessarily continuous, Multiple profiles can be found in the case that interference is more, is unsuitable to apply in the CT images of many disturbing factors.Mathematics morphology Strong interference immunity, but the profile for extracting is sometimes discontinuous.All it is not suitable for the thoracic cavity contour detecting of the CT images in cross section.
The content of the invention
It is an object of the invention to provide a series of a kind of thoracic cavity wheel in human body from top to bottom cross section scanning computed tomography images The method that exterior feature carries out Aulomatizeted Detect, with it, user only needs to be input into a series of cross section scanning computed tomography image, just may be used To be detected to CT images automatically, form a prompt judgement out each thoracic cavity profile of CT images.
To achieve these goals, technical solution of the invention is, by a series of cross sections to user input Scanning computed tomography image is detected that remove redundant information, one preliminary thoracic cavity closed curve of acquisition is adopted as initial profile automatically The profile in thoracic cavity is found to external expansion with active contour model (active contour model), so as to reach automatic decision chest The purpose of chamber profile.
CT images thoracic cavity profile automated detection method of the invention, specifically includes following steps：
1) a series of CT images in user input human body chest cross sections from top to bottom；
2) each image is pre-processed, removes the text information in CT images, and switch to gray-scale map；
3) image is processed, increases the discrimination of rib part and other parts；
4) rib region is extracted, the barycenter in each region is sought；
5) spinal region is determined, with its barycenter as origin, the barycenter in each region that origin is obtained with upper step is connected, and takes The midpoint of each line, closed curve profile is in turn connected to form by all of midpoint；
6) with closed curve profile as initial profile, obtained using active contour model (active contour model) The Internal periphery of trunk.
In above-mentioned technical proposal, step 3) described in image is processed, increase the area of torso portion and other parts Indexing, comprises the following specific steps that：
1) threshold value is rule of thumb taken, retains part of the gray scale more than threshold value, remove unnecessary information；
2) average value processing is carried out to image, the histogram of calculating input image, histogram normalization calculates histogram product Point, carry out histogram equalization；
3) contrast of image is strengthened.
Step 5) described in determination spinal region, specially：To first CT image, found in CT images the latter half The maximum bony areas of area, are recorded as the prime area searched next CT image；To other CT images, initial Areas adjacent is searched, and finds the bony areas of area maximum in each image, that is, obtain vertebra region in all images.
Described step 6) it is specially：
Using the closed curve profile of acquisition as initial curve, energy equation is built, with the elastic energy of curve and bending Energy is self-energy, calculates the shade of gray of image as outer energy, and to external expansion, continuous iteration causes energy side to setting curve Journey reaches minimum value, at this moment Internal periphery of the curve convergence to trunk.
The present invention is advantageous in that：
1. a series of cross section scanning computed tomography images of user input can automatically be detected, every image is first obtained An intrathoracic closed curve, the profile in thoracic cavity is found as initial profile to external expansion, to reach automatic decision thoracic cavity wheel Exterior feature, the profile that the method is extracted is continuous, and the method combination torso model feature is detected, strong interference immunity.
2. this method is convenient and swift, simple to operate, has efficiency higher compared to traditional determination methods.
Brief description of the drawings
Fig. 1 is profile Aulomatizeted Detect flow chart in CT images thoracic cavity of the invention.
Specific embodiment
The present invention is further illustrated below in conjunction with accompanying drawing.
CT images thoracic cavity profile Aulomatizeted Detect flow of the invention is as shown in figure 1, the flow has the following steps successively：
1) user input chest cross section a series of CT images A from top to bottom；
2) for i ＆ lt treatment, i-th image A is taken outi, to image AiPre-processed, removed word letter in CT images Breath, obtains removing the image B of text informationi；
3) to image BiProcessed, increased the discrimination of rib part and other parts, obtained image Ci；
4) rib region is extracted, the barycenter in each region is sought；
5) determine spinal region, with vertebra barycenter as origin, with horizontal direction to the right be x-axis positive direction, vertical direction to Above for y-axis positive direction sets up plane right-angle coordinate, the barycenter of each bony areas that connection origin and step 4 are obtained, in taking Point connects the midpoint of acquirement to form closed curve profile；
6) Internal periphery of trunk is positioned using active contour model (active contour model)；
7) return again to step 2 to next CT image to continue executing with, until all vertebra cross section CT image procossings are finished.
Above-mentioned steps 2) image is pre-processed, specifically include following steps：
1) unnecessary text information part in CT images is removed, leaving needs part to be processed, i.e., each in image Individual pixel P, the value of before processing is (r, g, b), then the value of P can be determined by below equation after treatment：
2) picture after treatment is converted into gray-scale map, obtains removing the image B of redundant informationi；
Step 3) in image is processed, increase the discrimination of rib part and other parts, specifically include following step Suddenly：
1) threshold value is rule of thumb taken, for example, takes threshold value 160, leave part of the gray scale more than 160, remove unnecessary information, I.e. for image BiEach pixel X in 1, the value of before processing is x, then the value of X can be determined by below equation after treatment：
To image BiEach pixel in 1 is processed, and obtains image Bi2；
2) to image Bi2 carry out average value processing, and the histogram of calculating input image, histogram normalization calculates histogram Integration, carries out histogram equalization, obtains the image B after average value processingi3；
3) strengthen the contrast of image, first, the mapping from [0,255] to [0,1], mapping are carried out to the pixel in image Pixel after penetrating does square treatment, then again to square after the pixel mapping that carries out from [0,1] to [0,255] so that numerical value Numerical value is smaller after small processes pixel, bigger after the big processes pixel of numerical value, increases the discrimination between the pixel of different numerical value, i.e., For image BiEach pixel X in 4, the value of before processing is x, then the value of x can be determined by below equation after treatment：
4) the image C of the discrimination for increasing rib part and other parts is finally giveni；
CT image middle rib bone parts are extracted, following steps are specifically included：
1) threshold value is rule of thumb taken, threshold value 180 is such as taken, part of the gray scale more than 180 is left, removal gray scale is less than 180 Part, i.e., for each pixel X in image Ci, the value of before processing is 255, then the value of X can be by below equation after treatment It is determined that：
2) edge of bone portion is extracted using edge detection algorithm, the edge to extracting is filled, and obtains multiple Bony areas, seek the barycenter of each bony areas.
Determine spinal region, specially：To first CT image, the maximum bone of area is found in CT images the latter half Region, is recorded as the prime area searched next CT image；To other CT images, looked near prime area Look for, find the bony areas of area maximum in each image, that is, obtain vertebra region in all images.
The Internal periphery of trunk is positioned using active contour model, specially：
Using step 5) in the closed curve profile that obtains as initial curve, energy equation is built, with the elasticity of curve Energy and flexional are self-energy, calculate the shade of gray of image as outer energy, setting curve to external expansion, continuous iteration So that energy equation reaches minimum value, at this moment Internal periphery of the curve convergence to trunk.

Claims (4)

1.CT images thoracic cavity profile automated detection method, it is characterised in that comprise the following steps：
1) a series of CT images in user input human body chest cross sections from top to bottom；
2) each image is pre-processed, removes the text information in CT images, and switch to gray-scale map；
3) image is processed, increases the discrimination of rib part and other parts；
4) rib region is extracted, the barycenter in each region is sought；
5) spinal region is determined, with its barycenter as origin, the barycenter in each region that origin is obtained with upper step is connected, and takes each company The midpoint of line, closed curve profile is in turn connected to form by all of midpoint；
6) with closed curve profile as initial profile, the Internal periphery of trunk is obtained using active contour model.
2. CT images thoracic cavity profile automated detection method according to claim 1, it is characterised in that described to image Processed, increased the discrimination of torso portion and other parts, comprised the following steps：
1) threshold value is rule of thumb taken, retains part of the gray scale more than threshold value, remove unnecessary information；
2) average value processing is carried out to image, the histogram of calculating input image, histogram normalization calculates histogram integration, enters Column hisgram is equalized；
3) contrast of image is strengthened.
3. CT images thoracic cavity profile automated detection method according to claim 1, it is characterised in that described step 5) Middle determination spinal region, specially：To first CT image, the maximum bony areas of area are found in CT images the latter half, It is recorded as the prime area searched next CT image；To other CT images, searched near prime area, sought The maximum bony areas of area in each image are looked for, that is, obtains vertebra region in all images.
4. CT images thoracic cavity profile automated detection method according to claim 1, it is characterised in that described step 6) Specially：
Using the closed curve profile of acquisition as initial curve, energy equation is built, with the elastic energy and flexional of curve It is self-energy, calculates the shade of gray of image as outer energy, to external expansion, continuous iteration causes that energy equation reaches to setting curve To minimum value, at this moment Internal periphery of the curve convergence to trunk.
CN201710039589.XA 2017-01-19 2017-01-19 CT images thoracic cavity profile automated detection method Pending CN106898012A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710039589.XA CN106898012A (en) 2017-01-19 2017-01-19 CT images thoracic cavity profile automated detection method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710039589.XA CN106898012A (en) 2017-01-19 2017-01-19 CT images thoracic cavity profile automated detection method

Publications (1)

Publication Number Publication Date
CN106898012A true CN106898012A (en) 2017-06-27

Family

ID=59198197

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710039589.XA Pending CN106898012A (en) 2017-01-19 2017-01-19 CT images thoracic cavity profile automated detection method

Country Status (1)

Country Link
CN (1) CN106898012A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111861984A (en) * 2020-06-08 2020-10-30 上海依智医疗技术有限公司 Method and device for determining lung region, computer equipment and storage medium
CN114271909A (en) * 2021-12-13 2022-04-05 杭州堃博生物科技有限公司 Information processing method, device, system, equipment and medium for chest puncture

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101819679A (en) * 2010-04-19 2010-09-01 李楚雅 Three-dimensional medical image segmentation method
CN102711626A (en) * 2010-01-07 2012-10-03 株式会社日立医疗器械 Medical image diagnosis device, and method for extracting and processing contour of medical image
CN102842132A (en) * 2012-07-12 2012-12-26 上海联影医疗科技有限公司 CT pulmonary nodule detection method
CN104361597A (en) * 2014-11-23 2015-02-18 长春工业大学 Liver DTI image segmentation method based on active contour model
CN105389821A (en) * 2015-11-20 2016-03-09 重庆邮电大学 Medical image segmentation method based on combination of cloud module and image segmentation
CN105574882A (en) * 2015-12-30 2016-05-11 中国科学院深圳先进技术研究院 Lung segmentation extraction method and system based on CT image of chest cross section
CN105741293A (en) * 2016-01-30 2016-07-06 上海联影医疗科技有限公司 Method for positioning organs in medical image

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102711626A (en) * 2010-01-07 2012-10-03 株式会社日立医疗器械 Medical image diagnosis device, and method for extracting and processing contour of medical image
CN101819679A (en) * 2010-04-19 2010-09-01 李楚雅 Three-dimensional medical image segmentation method
CN102842132A (en) * 2012-07-12 2012-12-26 上海联影医疗科技有限公司 CT pulmonary nodule detection method
CN104361597A (en) * 2014-11-23 2015-02-18 长春工业大学 Liver DTI image segmentation method based on active contour model
CN105389821A (en) * 2015-11-20 2016-03-09 重庆邮电大学 Medical image segmentation method based on combination of cloud module and image segmentation
CN105574882A (en) * 2015-12-30 2016-05-11 中国科学院深圳先进技术研究院 Lung segmentation extraction method and system based on CT image of chest cross section
CN105741293A (en) * 2016-01-30 2016-07-06 上海联影医疗科技有限公司 Method for positioning organs in medical image

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
仇涵等: "利用欧氏距离变换Snake模型分割脊椎CT图像", 《计算机工程与应用》 *
姚沛延等: "基于 CT 图像的脊椎骨组织分割", 《计算机工程与应用》 *

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111861984A (en) * 2020-06-08 2020-10-30 上海依智医疗技术有限公司 Method and device for determining lung region, computer equipment and storage medium
CN111861984B (en) * 2020-06-08 2023-12-22 北京深睿博联科技有限责任公司 Method and device for determining lung region, computer equipment and storage medium
CN114271909A (en) * 2021-12-13 2022-04-05 杭州堃博生物科技有限公司 Information processing method, device, system, equipment and medium for chest puncture

Similar Documents

Publication Publication Date Title
CN111091116B (en) Signal processing method and system for judging arrhythmia
CN104751437B (en) Lung's extraction method based on chest CT image
CN109753997B (en) Automatic accurate robust segmentation method for liver tumor in CT image
CN103310457B (en) A kind of pulmonary parenchyma dividing method based on para-curve correction convex closure
CN102521579B (en) Method for identifying pushing action based on two-dimensional planar camera and system
CN104933709A (en) Automatic random-walk CT lung parenchyma image segmentation method based on prior information
CN104143080A (en) Three-dimensional face recognition device and method based on three-dimensional point cloud
CN104574426B (en) Method and device based on personal signature analysis that is barefoot or wearing sock print image
CN108875586B (en) Functional limb rehabilitation training detection method based on depth image and skeleton data multi-feature fusion
CN103605964A (en) Face detection method and system based on image on-line learning
CN103870808A (en) Finger vein identification method
CN110555380A (en) Finger vein identification method based on Center Loss function
CN106909947A (en) CT image metal artifacts removing method and elimination system based on Mean Shift algorithms
CN110517198A (en) High frequency sensitivity GAN network for LDCT image denoising
CN106778767B (en) Visual image feature extraction and matching method based on ORB and active vision
CN109325955B (en) Retina layering method based on OCT image
CN111145185A (en) Lung parenchyma segmentation method for extracting CT image based on clustering key frame
CN106203255A (en) A kind of pedestrian based on time unifying heavily recognition methods and system
CN106898012A (en) CT images thoracic cavity profile automated detection method
CN106780470A (en) CT image nipple automated detection methods
CN110610502A (en) Automatic aortic arch region positioning and segmentation method based on CT image
CN115588102A (en) Lung cancer multi-modal feature extraction fusion method and system based on depth map
CN111612712B (en) Face correction degree determination method, device, equipment and medium
Yan et al. Automatic detection and localization of pulmonary nodules in ct images based on yolov5
CN114240893A (en) Method for measuring and calculating spinal Cobb angle in external image

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication

Application publication date: 20170627

RJ01 Rejection of invention patent application after publication