CN106600609A - Spine partition method and system in medical image - Google Patents
Spine partition method and system in medical image Download PDFInfo
- Publication number
- CN106600609A CN106600609A CN201611082125.9A CN201611082125A CN106600609A CN 106600609 A CN106600609 A CN 106600609A CN 201611082125 A CN201611082125 A CN 201611082125A CN 106600609 A CN106600609 A CN 106600609A
- Authority
- CN
- China
- Prior art keywords
- image
- medical image
- spinal
- spinal column
- line
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
-
- 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
- G06T2207/10088—Magnetic resonance imaging [MRI]
-
- 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
- G06T2207/10104—Positron emission tomography [PET]
-
- 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
- G06T2207/30012—Spine; Backbone
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
Abstract
The invention discloses a spine partition method in a medical image. The spine partition method comprises the steps of inputting the medical image, and partitioning a lung tissue region in the medical image; positioning a spinal cord line in the medical image; respectively determining an initial piece layer and an end piece layer which are corresponding to the spine in a vertical plane of the medical image according to the spinal cord line and the lung tissue region; determining a first side boundary of the spine in a region encircled by the initial piece layer and the end piece layer according to the spinal cord line, and determining a second side boundary of the spine according to the spinal cord line and the lung tissue region; and determining a spine vertebral body region according to the initial piece layer, the end piece layer, the first side boundary and the second side boundary. By the spine partition method, the extraction accuracy of the vertebral body region and an intervertebral disc region can be improved. Moreover, the invention also proposes a spine partition system in the medical image.
Description
Technical field
The present embodiments relate in technical field of medical image processing, more particularly to medical image spinal column dividing method
And system.
Background technology
Prostate specific antigen instrument (Emission Computed Tomography, ECT) mainly includes monochromatic light
Sub- emission computed tomography (SPECT) and PET-Positron emission computed tomography scanning (PET), wherein PET is used as current
High-level nuclear medicine technology, it has also become tumor, the heart, the indispensable important method of disease of brain diagnosis.Nuclear magnetic resonance
(Magnetic Resonance Imaging, MRI) can provide the anatomic form and physiological function information at imaging position, in reaction
It is preferably right that there is unrivaled superiority, particularly soft tissue MR images to present for anatomic form and physiological function information approach
It is than effect and radiationless.The combination that PET and MR are checked has the advantages that sensitivity is good, accuracy is high.
The positron radionuclide of PET imagings can occur positron decay, the negative electricity of the positron and surrounding of generation in the tissue
Son buries in oblivion a pair in opposite direction, each gammaphotons with 511keV energy of generation, and gammaphoton reaches detection through object
Photoelectric effect and Compton scattering can occur with electronics in object with certain probability during device, therefore all cannot be detected
Arrive, this is the root of ray attenuation.Correction for attenuation is the important step of PET imagings.But MR imagings are imaged different from CT, MR imagings
What is reflected is density rather than the tissue density of hydrone, is closed without mapping in PET-MR systems between MR images and PET attenuation quotients
System, it is impossible to for obtaining the decay pattern picture of PET.
At present, method conventional for the correction for attenuation of PET image in PET-MR systems is tissue typing's method:Decay is special
The different tissue of property and organ are divided into different regions, such as air, pulmonary, fat, muscle and skeleton etc., segmentation of then reentrying
Under the 511keV of good zones of different, the attenuation quotient of respective organization, carries out correction for attenuation.But, the skeletal tissue in MR images
Often there is similar gray value with the corresponding voxel such as air, the separation accuracy of skeleton and air voxel has in actual classification
Treat further to improve.
In consideration of it, being necessary to be improved existing medical image cutting method.
The content of the invention
The technical problem to be solved is to propose a kind of method that can effectively realize spine segmentation in medical image.
The present invention solve the technical scheme that adopted of above-mentioned technical problem for:The segmentation side of spinal column in a kind of medical image
Method, including:
Input medical image, and split lung tissue region in the medical image;
Spinal cord line is positioned in the medical image;
According to the spinal cord line and the lung tissue region, determine that spinal column institute is right respectively in the sagittal plane of the medical image
The starting sheet layer answered and termination lamella;
In the starting sheet layer and termination lamella area encompassed, the first side of spinal column is determined according to the spinal cord line
Border, and the second side circle of spinal column is determined according to the spinal cord line and lung tissue region;
Spinal vertebral region is determined based on the starting sheet layer, termination lamella, first side circle and second side circle.
Alternatively, positioning spinal cord line in the medical image includes:
Skin line segmentation result is obtained in the sagittal plane of the medical image;
Linear enhancement process is carried out to the sagittal plane of the medical image, and in the linearly corresponding doctor of enhanced sagittal plane
Image each lamella is learned, the first connected domain and the second connected domain is obtained;
First connected domain and the second connected domain are projected, projected image is obtained;
According to the skin line segmentation result, the position of spinal cord line is positioned in the projected image.
Alternatively, the position that spinal cord line is positioned in the projected image includes:
Retain the pixel point set in set point apart from the skin line in the projected image, obtain pre-selected zone
Domain;
The candidate point of each image line is obtained in the preselected area, the candidate point is the pixel jump of each image line
Point;
According to candidate's point location spinal cord line.
Alternatively, the first side circle of the spinal column obtains as follows:
The starting sheet layer and the lamella terminated between lamella for being pointed to the spinal column is filtered process;
Obtain the spinal column gray reference value of each image line in the lamella after processing after filtering;
According to the spinal column gray reference value, the first side of the spinal column of each image line is obtained on the right side of the spinal cord line
Border.
Alternatively, the second side circle of the spinal column obtains as follows:
In the range of the starting sheet layer and termination topsheet areas of the spinal column, the right hand edge in the lung tissue region is determined;
The spinal cord line is moved to the right hand edge in the lung tissue region, to determine the second side circle of the spinal column.
Alternatively, also include:
The maximum lamella corresponding to lung tissue region is extracted in the coronalplane of the medical image;
The left end point of intervertebral disc is determined in the maximum lamella;
Profile of the intervertebral disc in coronalplane is obtained according to the left end point of the intervertebral disc;
According to intervertebral disc in the profile of coronalplane, the left margin of intervertebral disc is obtained in sagittal plane;
Spine segmentation result is obtained according to the left margin of the spinal vertebral region and the intervertebral disc.
Alternatively, the left end point for determining intervertebral disc includes:
Pre-selection connected domain is obtained in the maximum lamella according to coronalplane lung tissue region;
Target connected domain is determined in the pre-selection connected domain, the pixel of the target connected domain is met in image row direction
First given threshold, meets the second given threshold in image column direction;
Left end point of the left end point of the target connected domain for intervertebral disc.
According to another aspect of the invention, it is proposed that in a kind of medical image spinal column segmenting system, including spinal vertebral carries
Device is taken, the spinal vertebral extraction element includes:
Segmentation of lung parenchyma unit, for splitting lung tissue region in medical image;
Spinal cord line positioning unit, for spinal cord line is positioned in the medical image;
Border determining unit, for according to the spinal cord line and the lung tissue region, in the sagittal of the medical image
Face determines the starting sheet layer corresponding to spinal column respectively and terminates lamella;And,
In the starting sheet layer and termination lamella area encompassed, the first side of spinal column is determined according to the spinal cord line
Border, and the second side circle of spinal column is determined according to the spinal cord line and lung tissue region;
Vertebral body area acquisition unit, for based on the starting sheet layer, termination lamella, first side circle and second side circle
Determine spinal vertebral region.
Alternatively, the spinal cord line positioning unit includes:
Skin line obtains subelement, obtains skin line segmentation result for the sagittal plane in the medical image;
Connected domain obtains subelement, carries out linear enhancement process for the sagittal plane in medical image, and is linearly strengthening
Corresponding each lamella of sagittal plane afterwards, obtains the first connected domain and the second connected domain;
Projected image obtains subelement, for projecting to first connected domain and the second connected domain, obtains projection
Image;
Locator unit, for according to the skin line segmentation result, positioning the position of spinal cord line in the projected image.
Alternatively, also include including intervertebral disc extraction element, and the intervertebral disc extraction element:
Maximum lamella extraction unit, extracts the maximum corresponding to lung tissue region for the coronalplane in the medical image
Lamella;
Intervertebral disc contours extract unit, for the left end point for determining intervertebral disc in the maximum lamella, and according to described
The left end point of intervertebral disc obtains profile of the intervertebral disc in coronalplane;
According to intervertebral disc in the profile of coronalplane, the left margin of intervertebral disc is obtained in sagittal plane, and according to the left side
Extract vertebral body region in boundary.
Compared with prior art, beneficial effects of the present invention are:According to the continuity features of myeloid tissue in medical image,
Spinal cord line is positioned in medical image;With reference to the morphological characteristic of spinal column, obtain the initial layers of spinal vertebral, terminate based on spinal cord line
Layer and left and right border, and the initial layers according to spinal vertebral, stop layer and left and right border determine spinal vertebral region, carry
The high precision of vertebral body extracted region;Shape of the gradient information and spinal column based on intervertebral disc in the image of different directions lamella
Shape and half-tone information, effectively realize the full-automatic dividing of spinal column, and in achievable many bed medical images scannings, spinal column is full-automatic
Segmentation.
Description of the drawings
Magic magiscan structural representations of the Fig. 1 for the embodiment of the present invention;
MRI scan apparatus structure schematic diagrams of the Fig. 2 for the embodiment of the present invention;
Spine segmentation system structure diagrams of the Fig. 3 for the embodiment of the present invention;
Spinal vertebral extraction element structural representations of the Fig. 4 for the embodiment of the present invention;
Spinal cord line positioning unit structural representations of the Fig. 5 for the embodiment of the present invention;
Medical image spine segmentation method flow diagrams of the Fig. 6 for the embodiment of the present invention;
Spinal cord line localization method flow charts of the Fig. 7 for the embodiment of the present invention;
Fig. 8 is the spinal cord line result schematic diagram that the embodiment of the present invention is obtained;
Fig. 9 extracts disc area method flow diagram from magnetic resonance image (MRI) for the embodiment of the present invention;
Figure 10 a are the intervertebral disc result schematic diagram on the sagittal plane that the embodiment of the present invention is obtained;
Figure 10 b are the intervertebral disc result schematic diagram on the coronalplane that the embodiment of the present invention is obtained;
Figure 11 is the sagittal plane view that the embodiment of the present invention does not carry out spine segmentation before processing magnetic resonance image (MRI);
Figure 12 carries out the sagittal plane view of the magnetic resonance image (MRI) after spine segmentation process for the embodiment of the present invention;
Figure 13 is the spinal column result figure of embodiment of the present invention segmentation.
Specific embodiment
In order to be illustrated more clearly that the technical scheme of embodiments of the invention, below will be to making needed for embodiment description
Accompanying drawing is briefly described.It should be evident that drawings in the following description are only some examples of the present invention or enforcement
Example, for one of ordinary skill in the art, on the premise of not paying creative work, can be with according to these accompanying drawings
Apply the present invention to other similar scenes.Unless obviously or separately explained from language environment, identical label generation in figure
The identical structure of table or operation.
As shown in the present invention and claims, unless context clearly points out exceptional situation, " one ", " one ", " one
The word such as kind " and/or " being somebody's turn to do " not refers in particular to odd number, may also comprise plural number.It is, in general, that term " including " only points out bag with "comprising"
Include the step of clearly identifying and element, and these steps and element do not constitute one it is exclusive enumerate, method or equipment
It is likely to comprising the step of other or element.
Although the present invention is made that various references to the certain module of the data processing of embodiment, however, any amount of
Disparate modules can by using and operate on a client being connected with the system by network and/or server.It is described
Module is merely illustrative, and the different aspect of the system and method can use disparate modules.
In medical image or data handling procedure, " image segmentation ", " image zooming-out ", " image classification " or " image
Divide " mutually can convert, express and the image for meeting certain condition is chosen from extensive area.In certain embodiments, cure
Learning image processing system can include one or more form.The medical image system can be single mode system such as:Magnetic is altogether
Shake imaging (MRI), magnetic resonance angiography (MRA) etc., or the multimodal systems comprising nuclear magnetic resonance, and such as numeral subtracts
Shadow angiography (DSA), ultrasonic scanning (US), positron emission tomography (PET), single photon emission computed tomography
One or more of DSA-MR, US-MR, PET-MR, SPECT-MR that scanning (SPECT) is combined to form with MR imaging apparatus etc.
Combination.
In certain embodiments, the region of image scanning can be the one kind such as organ, body, object, damage location, tumor
Or the region that various combinations is located.The region of image scanning can be cervical region, thoracic cavity, abdominal part, organ, extremity, skeleton, blood vessel
In the region that one or more of combination is located.The region of scanning can be the region that the tissue at one or more positions is located.
In certain embodiments, image can be two dimensional image and/or 3-D view.In two dimensional image, most trickle distinguishable unit
Element can be pixel (pixel).In 3-D view, most trickle resolvable elements can be voxel (voxel), and in three-dimensional
In image, image can be by a series of two dimension slicing, two-dimensional slice or two-dimensional slice image construction.In certain embodiments, piece
Layer refers to a certain two dimension slicing of position or 3 d medical images of the two dimension slicing in 3 d medical images.In other embodiments
In, lamella and slice image mutually can be converted, two-dimensional medical images of the expression comprising multiple pixels.
It should be noted that below for the description of magic magiscan, only for convenience of description, can not be this
It is bright be limited in illustrated embodiment within the scope of.It is appreciated that for a person skilled in the art, in the original for understanding the system
After reason, in the case of without departing substantially from this principle, modules are carried out with combination in any, or constitutes subsystem and other modules
Connection, the various amendments and change to implementing said method and in systematic difference field form and details.
Fig. 1 show the magic magiscan structural representation of one embodiment of the invention.The Medical Image Processing system
System 100 may include medical image scan device, control system 120, spine segmentation system 130, input/output device 140 and
Transmission network 150, in the present embodiment, pending medical image may be selected magnetic resonance image (MRI), and medical image scan device can be with
For nuclear magnetic resonance (Magnetic Resonance Imaging, MRI) scanning means 110.
It should be noted that each module of the magic magiscan of the present invention or unit can connect each other, the company
Connecing can be wireless network connection or finite element network connection.Wherein, cable network can include utilization, optical fiber, metallic cable,
The mode of one or more combination such as compound cable, one or more interfaces.Wireless network can be included using bluetooth, region office
Domain net (LAN), wide local area network (WAN), near source field communication (Near Field Communication, NFC) etc. are a kind of or many
Plant the mode of combination.
MRI scan device 110 can be MRI scan instrument or scanning device.It is illustrated in figure 2 the present invention one real
Apply 110 structural representation of MRI scan device of example, including magnet unit 210, Gradient Unit 220, radio frequency unit 230, spectrometer
Other supplementary modules such as 240 and controller 250 or component, wherein, magnet unit 210 is used to produce main field, Gradient Unit
220 mainly include gradient current amplifier (AMP), gradient coil;Radio frequency unit 230 mainly includes radio-frequency sending coil and radio frequency
Receiving coil;Spectrometer 240 mainly includes pulse-series generator, gradient waveform generator, transmitter and receiver etc., and controls
Device 250 is used to control the operation of MRI scan device and is ultimately imaged.
MRI scan device 110 imaging general procedure be:
Controller 250 stores and sends the instruction of the scanning sequence (scan sequence) for needing to perform, and pulse train is sent out
Raw device is controlled to gradient waveform generator and transmitter according to scanning sequence instruction, and gradient waveform generator output is with pre-
The gradient pulse signal of timing sequence and waveform;
Gradient pulse signal through Gx, Gy and Gz gradient current amplifier, then by three independences in Gradient Unit 220
Passage Gx, Gy, Gz, each gradient amplifier excite a corresponding gradient line in the gradient coils set of composition gradient unit 220
Circle, produces for generating the gradient fields of additional space encoded signal, to carry out space orientation to magnetic resonance signal;
Pulse-series generator in spectrometer 240 also performs scanning sequence, and output includes the radio-frequency pulse of radio-frequency transmissions
The timing of the data such as timing, intensity, shape and radio frequency reception and the length of data acquisition window to transmitter, while transmitter
The body transmitting coil that respective radio-frequency pulsing to radio frequency unit 230 is included produces B1 fields;
The reception line that the signal that the atomic nucleus being excited under B1 field actions in patient body send is included by radio frequency unit 230
Circle is perceived, and is then transferred to preamplifier by sending/receiving switch, the magnetic resonance signal of amplification is demodulated, filter,
The digitized processing such as AD conversion are then communicated to the storage module of controller 250, and when storing, module one group of original k- of acquisition is empty
Between after data, the end of scan.Original K space data is rearranged into will be the image being reconstructed corresponding single with each
K space data group, the K space data group are imported into array processor, after Fourier transformation (FFT) carries out image reconstruction
With reference to magnetic resonance signal, one group of view data, the i.e. magnetic resonance image (MRI) of scanning area are formed.
Control system 120 can be used to control MRI scan device 110, input/output 140, and/or spine segmentation system
System 130.In certain embodiments, control system 120 can receive the picture signal or and target from MRI scan device 110
The related data of object.Control system 120 can send picture signal or related to target object to spine segmentation system 130
Data, can receive from spine segmentation system 130 treated data or segmentation after image, it is also possible to input/
Output system 140 sends the image of treated data or segmentation.In certain embodiments, control system 120 can include
Computer, program, algorithm, software, storage device and interface etc..Interface can include MRI scan device 110, input/output system
The interface between other modules or unit in system 140, spine segmentation system 130 and/or imaging system.
Control system 120 can include central processing unit (Central Processing Unit, CPU), specialized application collection
Into circuit (Application Specific Integrated Circuit, ASIC), dedicated instruction processor
(Application Specific Instruction Set Processor, ASIP), concurrent physical processor (Physics
Processing Unit, PPU), digital signal processor (Digital Processing Processor, DSP), scene can
Programmed logic gate array (Field-Programmable Gate Array, FPGA), PLD
One kind or several in (Programmable Logic Device, PLD), processor, microprocessor, controller, microcontroller etc.
The combination planted.
Spine segmentation system 130 can process the image information or data letter from control system 120 and/or network 150
Breath.In one embodiment, spine segmentation system 130 can process relevant abdominal part or thoracic cavity view data.For example, spine segmentation dress
Putting and spinal vertebral region or/and disc area being partitioned into from thoracic cavity data, spine segmentation device can be with from abdominal part number
According to extraction disc area.
Spine segmentation system 130 structural representations of the Fig. 3 for one embodiment of the invention.The spine segmentation system 130 includes
Spinal vertebral extraction element 310 and intervertebral disc extraction element 320.Wherein, spinal vertebral extraction element 310 can be from medical image
(such as magnetic resonance image (MRI)) extracts the vertebral body region for belonging to spinal column, such as near the vertebral body region of pulmonary;Intervertebral disc extraction element 320 can
Belong to the disc area of spinal column, such as lumbar disc region from medical science image zooming-out.
In certain embodiments, spine segmentation system 130 can be with central processing unit (Central Processing
Unit, CPU), it is specialized application integrated circuit (Application Specific Integrated Circuit, ASIC), special
Instruction processing unit (Application Specific Instruction Set Processor, ASIP), concurrent physical processor
(Physics Processing Unit, PPU), digital signal processor (Digital Processing Processor,
DSP), field programmable gate array (Field-Programmable Gate Array, FPGA), PLD
One kind or several in (Programmable Logic Device, PLD), processor, microprocessor, controller, microcontroller etc.
The combination planted, for performing aforesaid operations.Spine segmentation system 130 may also include the memorizer with store function, the storage
Device includes, but are not limited to floppy disk, CD, CD-ROM (compact-disc-read only memory), magneto-optic disk, ROM (read only memory), RAM
(random access memory), EPROM (Erasable Programmable Read Only Memory EPROM), EEPROM (the read-only storages of electrically erasable
Device), magnetic or optical card, flash memory or be suitable to store machine-executable instruction other kinds of medium/machine readable media.
Input/output 140 may include the human-computer interaction devices such as display, mouse, keyboard.In one embodiment,
Patient's physiologic informations such as the height of person under inspection, body weight, age, imaging position can be input into by mouse or keyboard;By display
Working condition, the imaging section of the physiologic informations such as the height of person under inspection, body weight, age, imaging position and scanning device can be shown
Medical image or other images of position etc..The type of display can be cathode ray tube (CRT) display, liquid crystal display
(LCD), the combination of one or more in OLED (OLED), plasma display etc..
It is illustrated in figure 4 the spinal vertebral extraction element structural representation of the embodiment of the present invention.Spinal vertebral extraction element
310 may include that segmentation of lung parenchyma unit 410, spinal cord line positioning unit 420, border determining unit 430 and vertebral body region obtain list
Unit 440.
Segmentation of lung parenchyma unit 410 can be in segmentation lung tissue region in medical image (such as magnetic resonance image (MRI)).In a reality
Apply in example, segmentation of lung parenchyma unit 410 can adopt Threshold segmentation, pixel segmentation, Range Image Segmentation, rim detection or be based on mould
The methods such as paste collection determine the demarcation line of skin line and background area in magnetic resonance image (MRI), and the region included in skin line is further
Confirmation belongs to the region of lung tissue.Exemplarily, lung tissue region is obtained using the method based on Threshold segmentation, should be based on threshold value
The method of segmentation can adopt global threshold method, adaptive threshold method or optimal threshold method in one or more.
In another embodiment, lung tissue region is obtained using the method for edge segmentation:Detection gray level or structure have mutation
Place, and the gray value for often showing as edge pixel where being mutated in the picture is discontinuous, can determine that by derivation
This obvious edge.
Spinal cord line positioning unit 420, can position spinal cord line in medical image.Border determining unit 430, can be according to spinal cord
Line and lung tissue region, determine the starting sheet layer corresponding to spinal column respectively in the sagittal plane of medical image and terminate lamella;From
Starting sheet layer determines the first side circle of spinal column, Yi Jigen according to spinal cord line to all slice image regions for terminating lamella
The second side circle of spinal column is determined according to spinal cord line and lung tissue region.Vertebral body area acquisition unit 440, based on starting sheet layer, end
Limited step layer, first side circle and second side circle determine spinal vertebral region.
It is illustrated in figure 5 the spinal cord line positioning unit structural representation of the embodiment of the present invention.Spinal cord line positioning unit 420 is wrapped
Include skin line and obtain subelement 510, connected domain acquisition subelement 520, projected image acquisition subelement 530 and locator unit
540。
Skin line obtains subelement 510, can obtain skin line segmentation result in the sagittal plane of medical image.Implement at one
In example, can be according to the background area of medical image and imaging region discontinuous feature between the two, using based on differential operator
Edge detection method determine background area and the imaging region comprising body part image pixel between the two respectively.
In another embodiment, will also carry out the disposal of gentle filter to image before edge being detected using differential operator.
Connected domain obtains subelement 520, can carry out linear enhancement process in the sagittal plane of medical image, and linearly strengthen
On rear corresponding each lamella of sagittal plane or slice image, the first connected domain and the second connected domain are obtained.Exemplarily, by skin
The pixel of skin line can primarily determine that the region that may include lung tissue, the sagittal connected domain of detection medical image, according to even
The length in logical domain determines the first connected domain and the second connected domain in each lamella or slice image, wherein the first connected domain is the piece
The maximum connected domain of layer length, the second connected domain are only second to the enclosed region of the first connected domain for length.
Projected image obtains subelement 530, and the first connected domain and the second connected domain can be projected, and obtains projection
Picture.In one embodiment, in each lamella or slice image of medical image sagittal view picture comprising the first connected domain and
Second connected domain, projects to sagittal each slice image along sagittal axis (depth) direction, can obtain by multiple projections
The projected image of region composition.
Locator unit 540, can position the position of spinal cord line in projected image according to skin line segmentation result.At one
In embodiment, locator unit 540 is may be selected with spinal cord line degree of association highest skin line as reference line, and according to the reference
Line primarily determines that the preselected area of spinal cord line on projection images;The picture included by each image line is obtained in the preselected area
In vegetarian refreshments, first pixel value is 1 pixel, makes the candidate point that the pixel is correspondence image row, all image line candidate points
Collection be combined into spinal cord line.In another embodiment, preselected area can also include the projection of muscle of back organization edge image, positioning
Subelement only can start the pixel that pixel value is changed into 0 from 1 from low order end in preselected area retains each image line, and delete
On each image line, other are changed into 0 pixel from 1, to eliminate impact of the muscle of back tissue to the detection of spinal cord line.
Some embodiments of the invention, are illustrated in figure 6 a kind of exemplary stream of medical image spine segmentation process
Cheng Tu.In certain embodiments, spine segmentation flow process can be performed by spine segmentation system 130, mainly comprised the steps:
Step 601. is input into medical image, and splits lung tissue region from medical image.The medical image can be magnetic
Resonance image, and the imaging position of magnetic resonance image (MRI) can be human body, animal or their a portion, for example, organ,
Tissue, diseased region, tumor locus, or the combination in any at above-mentioned position.The magnetic resonance image (MRI) can be comprising multiple lamellas
Or the view data of slice image, and view data can refer to corresponding to the minimum unit of image, i.e. unit pixel (pixel)
Data.The acquisition of magnetic resonance image (MRI) can be obtained by MRI scan device 110, it is also possible to by transmission network 150 from external data
Source obtains, can be obtaining from input-output apparatus 140.
In one embodiment, magnetic resonance image (MRI) is from MRI scan device 110, and performs in segmentation of lung parenchyma unit 410
Following process:Magnetic resonance image (MRI) is smoothed;Differential process life is carried out to the magnetic resonance image (MRI) after smoothing processing
Into gradient image;Should dividing ridge method obtain from gradient image and belong to the voxel of " lung tissue ", extract lung tissue image.Example
Property, dividing ridge method refers to document Vincent L, Soille P.Watersheds in digital spaces:an
efficient algorithm based on immersion simulations[J].IEEE transactions on
pattern analysis and machine intelligence,1991,13(6):583-598.In another embodiment,
Dividing ridge method can also generate watershed line according to the low image value of gradient image, to generate the lung tissue of segmentation.In another reality
Apply in example, while lung tissue is split, the segmentation result of skin line can be also obtained according to above-mentioned same method.
Step 602. positions spinal cord line in medical image, and the spinal cord line (Spinal cord line, SC lines) is spinal column
Edge line.In one embodiment, position spinal cord line obtain in spinal cord line positioning unit 420, acquisition process may include as
Lower step:The connection that spine regions may be included in each lamella or slice image is extracted in the three-dimensional sagittal plane of magnetic resonance image (MRI)
Domain;The lamella corresponding to spinal column intermediate layer is determined in magnetic resonance image (MRI);And the possibility of spinal cord line is primarily determined that according to the lamella
Pixel;Pixel now may include the pixel that muscle of back tissue etc. is not belonging to spinal cord line, therefore, further adopt
The pixel for belonging to spinal cord line is determined with the method for candidate point.
Step 603. is determined corresponding to spinal column in the sagittal plane of medical image respectively according to spinal cord line and lung tissue region
Starting sheet layer and termination lamella.In one embodiment, the middle sheet of spinal cord on sagittal view picture can be obtained first, divided
Lung tissue cross-sectional image data conversion is cut into three-dimensional sagittal plane view data;From the beginning of the middle sheet of spinal cord, (lamella forward
Number is less than current lamella) find sagittal view picture in split lung tissue area for the first time more than the lamella of setting pixel threshold,
And make the starting sheet layer that the lamella is spinal column;Meanwhile, from the beginning of the middle sheet of the spinal cord for obtaining, (the piece number of plies is more than and works as backward
Topsheet layer) find sagittal view picture in split lung tissue area for the first time more than the lamella of setting pixel threshold, and make the piece
Termination lamella of the layer as spinal column.Alternatively, set pixel threshold and gray value may be selected as 350-450.
In another embodiment, on sagittal view picture, the middle sheet of spinal cord can be obtained as follows:In magnetic resonance
Image sagittal plane determines spinal column middle sheet;For each lamella of 3 layers of scope up and down of spinal column middle sheet, Retention area is most
Two big connected domains;Judge whether the connected domain and the spinal cord line that retain or so 3 pixel width regions intersect, if two region phases
Hand over, then further calculate the area of the connected domain;If two regions are non-intersect, give up the connected domain;Make spinal column middle sheet
The maximum connected domain of 3 layers of scope inner area up and down corresponding to lamella be spinal cord on sagittal view picture middle sheet.With ridge
The middle sheet of marrow is starting, has split lung tissue area for the first time more than setting in finding sagittal view picture respectively forwardly, backward
Determine the lamella (setting pixel threshold may be selected 400) of pixel threshold, order meets the starting sheet layer of the lamella for spinal cord of above-mentioned condition
Or terminate lamella.
Step 604. is in starting sheet layer and terminates, in lamella area encompassed, determining the first side of spinal column according to spinal cord line
Border, and the second side circle of spinal column is determined according to spinal cord line and lung tissue region.
Exemplarily, the first side circle of spinal column obtains as follows:It is pointed to starting sheet layer and the termination of spinal column
Lamella or slice image between lamella is filtered process;Obtain each on lamella or slice image after processing after filtering
The spinal column gray reference value of image line;According to spinal column gray reference value, the spinal column of each image line is obtained on the right side of spinal cord line
First side circle.The second side circle of spinal column obtains as follows:In the piece of the starting sheet layer to termination lamella of spinal column
In tomographic image regional extent, the right hand edge in lung tissue region is determined;Spinal cord line is moved to the right hand edge in lung tissue region, with true
Determine the second side circle of spinal column.
Step 605. determines spinal vertebral region based on starting sheet layer, termination lamella, first side circle and second side circle.
Spinal vertebral region is starting sheet layer, terminates the set of lamella, first side circle and institute of second side circle enclosing region.
In one embodiment, medical image selects magnetic resonance image (MRI), and the positioning of spinal cord line may include following steps:
Step 701. obtains skin line segmentation result in the sagittal plane of magnetic resonance image (MRI), and the segmentation of skin line can refer to lung group
The cutting procedure knitted.
In one embodiment, skin line segmentation result is obtained in skin line obtains subelement 510, and the skin line is obtained
Take subelement 510 and perform following operation:Magnetic resonance image (MRI) is smoothed;To the magnetic resonance image (MRI) after smoothing processing
Differential process are carried out, gradient image is generated;Should dividing ridge method obtain from gradient image and belong to the pixel or voxel of skin line,
To extract the edge of skin line image or skin line.
Step 702. carries out linear enhancement process to the sagittal plane of magnetic resonance image (MRI), and in linear enhanced sagittal plane pair
Each lamella answered, obtains the first connected domain and the second connected domain.
In one embodiment, following operation can be performed in connected domain obtains subelement 520:The three of magnetic resonance image (MRI)
Dimension sagittal plane, processes the sagittal slice image using linear enhanced method;Sagittal view picture after enhancement process is adopted
Binary image is obtained with threshold segmentation method;In the binary image for obtaining, retain most in each slice image or lamella
Long connected domain and length are only second to the secondary connected domain of most long connected domain.Exemplarily, using such as list of references Li Q,
Sone S,Doi K.Selective enhancement filters for nodules,vessels,and airway
walls in two-and three-dimensional CT scans[J].Medical physics,2003,30(8):
Method shown in 2040-2051 carries out enhancement process to sagittal plane slice image;In Threshold segmentation process, gray threshold is set
For 0, for currently processed slice image after enhancement process any pixel, if the gray value of the pixel is more than or equal to setting
Gray threshold, then make the pixel be 1;If the gray value of the pixel is less than setting gray threshold, the pixel is made to be 0.
In another embodiment, the pixel of each lamella sagittal to magnetic resonance image (MRI) or slice image carries out threshold value point
After cutting, multiple connected domains in each lamella or slice image, can be obtained, in one of slice image, include three connected domains S1、
S2And S3, and the length relation in tee joint domain is S1>S2>S3, S can be made according to mentioned above principle1For the first connected domain, S is made2For second
Connected domain, and delete connected domain S3。
In another embodiment, each lamella sagittal to magnetic resonance image (MRI) or slice image are entered after row threshold division,
One of slice image includes four connected domains S1、S2、S3And S4, and the length relation in tee joint domain is S1>S2=S3>S4, root
S can be made according to mentioned above principle1For the first connected domain, S is made2Or S3One of them is the second connected domain, retains the first connected domain and second
Connected domain, while deleting remaining connected domain.
The first connected domain of step 703. pair and the second connected domain are projected, and obtain projected image.It is basic in step 702
On, the connected domain of two reservations can be obtained in sagittal each lamella or slice image, these connected domains are that spinal cord line may
Distributed areas, the connected domain that each lamella retains can be projected in a certain direction, to obtain projected image.
In one embodiment, projected image obtains subelement 530 and performs following operation:To each slice image of sagittal plane
On the first connected domain, the second connected domain along depth direction carry out projection obtain projected image I1.Projected image I1Skin can be included
Other non-spinal regions such as skin border, lung tissue border.
Step 704. positions the position of spinal cord line in projected image according to skin line segmentation result.Obtaining projection
As on the basis of, locator unit 540 can also carry out following operation:A reference line is selected, distance reference line set point is made
Pixel point set is preselected area I2。
Exemplarily, reference line may be selected the skin line of spinal column middle sheet in magnetic resonance image (MRI), and preselected area I2Can
Being projected image I1Pixel point set of the middle skin line apart from spinal column middle sheet in 30 pixel coverages.In the middle of spinal column
The acquisition of lamella can be carried out in skin line obtains subelement 510, and may include following process:In each piece of magnetic resonance image (MRI)
Skin line and lung tissue region are obtained in layer or slice image;By dividing in each lamella of the transverse section of statistics magnetic resonance image (MRI)
Cut the connected domain number of skin line determination;The middle sheet of spinal column is determined according to the number of connected domain.
In one embodiment, for arbitrary slice image of magnetic resonance image (MRI) transverse section, if connected domain number is more than
2, then the maximum connected domain of Retention area;Obtain in all lamellas or slice image of transverse section and retain connected domain area most
Big level, and obtain the pixel of the largest connected domain low order end and high order end;Calculate the pixel of low order end and high order end
The meansigma methodss of corresponding columns, the meansigma methodss are intermediate layer of the spinal column in sagittal view picture.
In another embodiment, for arbitrary lamella or slice image of magnetic resonance image (MRI) transverse section, if the lamella or
The connected domain number of slice image is less than or equal to 2, then from the beginning of lung tissue zones vanishes layer is split, to splitting skin line area
Domain disappearance layer terminates, and retains level of the connected domain number equal to 1;Find to retain in the connected domain that above-mentioned all lamellas retain and connect
The maximum lamella of logical domain area, and determine the connected domain low order end and the pixel of high order end of the maximum lamella, make low order end and
The meansigma methodss of the pixel correspondence columns of high order end are intermediate layer of the spinal column in sagittal view picture.
In another embodiment, locator unit 540 can obtain the candidate point of each image line, the candidate in preselected area
Pixel jump point of the point for each image line, according to candidate's point location spinal cord line.Exemplarily, the acquisition process bag of candidate point
Include:
First, in preselected area I2Determine that each image line is changed into 0 pixel position from right to left from 1, and make the picture
Vegetarian refreshments is candidate point, if being changed into 0 candidate point number more than 1 from 1 in the image line, is retained first on this image line
Be changed into 0 candidate point all pixels point to the left from 1, this step is repeated until last image line, can by aforesaid operations
Eliminate impact of the muscle of back tissue to the detection of spinal cord line.
Then, there is the image line that pixel value is 1 from first time and the image line that pixel value is 1 occur to last time, often
One image line retains a candidate point, and the testing result of spinal cord line is obtained according to candidate point.Exemplarily, with arbitrary image row
For current image line, the selected candidate point of current image line is obtained from candidate point.The candidate point of selecting of current image line should
Meet following condition:The row coordinate distance that the row coordinate of candidate point has selected candidate point with previous image row on current image line is most
Near candidate point, and ensure that the row coordinate distance of two candidate points is less than or equal to 3 pixels.
Finally, current image line has been selected candidate point and candidate point has been selected as previous image row, repeated this step
Suddenly up to last image line of image;If all candidate points on current image line do not meet above-mentioned condition or present image
No candidate point on row, then the row coordinate and front two image line for determining whether candidate point on front image line have selected candidate point
The nearest candidate point of row coordinate distance, and ensure that the row coordinate distance of two candidate points is less than or equal to 3 pixels, meet above-mentioned bar
The candidate point of part is that current image line has selected candidate point;Otherwise, current image line candidate point column number scale is 0.Need
It is bright, if the columns of candidate point is 0, the columns of candidate point has been selected as current image line candidate with previous image row
The columns of point, the collection of the columns composition of the candidate point of all image lines are combined into spinal cord line.
It is illustrated in figure 8 the spinal cord line positioning result schematic diagram of one embodiment of the invention.According to the knot of the sagittal plane view
Really, spinal cord line (in the figure full curve constituted by the high point of gray value) is positioned at the right side edge of spinal column, and prolongs from neck area
Extend lumbar regions.
In one embodiment, right margin of the first side circle for spinal column, and the right margin of spinal column is obtained as follows
Take:In the range of the spinal column starting sheet layer and image-region determined by termination lamella obtained by step 603, to every on sagittal plane
Individual lamella or slice image carry out mean filter;The spinal cord line that step 602 determines or so 3 pixel areas are found in every image line
Gray scale maximum in domain, makes the maximum be correspondence image row spinal column gray reference value;On the right side of spinal cord line, look into by image line
Look for gray value for the first time more than the pixel position of spinal column gray reference value on the image line, as the right side of spinal column on the image line
Border, repeats this step until last image line, the pixel point set for meeting above-mentioned condition on all image lines is spinal column
Right margin.
Similarly, left margin of the second side circle for spinal column, and the left margin of spinal column is obtained as follows:By step
In the range of the rapid 603 spinal column starting sheet layers for obtaining and image-region determined by termination lamella, retain each lamella on sagittal plane
Or slice image has split the right hand edge line of lung tissue;Determine the rightmost edges point of lung tissue right hand edge line;Step 602 is obtained
Spinal cord line move at rightmost edges point, determine the left margin of spinal column.
It is by the extractable thoracic cavity position of said method or the vertebral body region near lung areas, further of the invention also to carry
Take lumbar vertebra position disc area.As shown in figure 3, the spine segmentation system for magnetic resonance image (MRI) of the present invention also includes intervertebral
Disk extraction element 320, the intervertebral disc extraction element 320 may include maximum lamella extraction unit and intervertebral disc contours extract unit.
Wherein, maximum lamella extraction unit, obtains the maximum lamella corresponding to lung tissue region for the coronalplane in magnetic resonance image (MRI).
Intervertebral disc contours extract unit, for the left end point of intervertebral disc is determined in maximum lamella, and obtains according to the left end point of intervertebral disc
Profile of the intervertebral disc in coronalplane;According to intervertebral disc in the profile of coronalplane, the left margin of intervertebral disc is obtained in sagittal plane;With
And, spine segmentation result is obtained according to the left margin of spinal vertebral region and intervertebral disc.
One embodiment of the invention obtains spine segmentation result mainly to be included:From magnetic resonance image (MRI) extract spinal vertebral region and
Disc area, wherein, spinal vertebral region is extracted from magnetic resonance image (MRI) can refer to the method flow shown in Fig. 6.Such as Fig. 9 institutes
Show, extracting disc area from magnetic resonance image (MRI) mainly includes:
Step 901. is input into magnetic resonance image (MRI), and splits lung tissue region from magnetic resonance image (MRI).Magnetic resonance image (MRI) is obtained
Take and can be obtained by MRI scan device 110, it is also possible to obtained from external data source by transmission network 150, can be with from defeated
Enter/outut device 140 obtains.
Coronalplane of the step 902. in magnetic resonance image (MRI), obtains the maximum lamella corresponding to lung tissue region, coughs referred to as
For maximum slice image, the step can be performed in maximum lamella extraction unit.
Step 903. determines the left end point of intervertebral disc in maximum lamella, and obtains intervertebral disc according to the left end point of intervertebral disc
In the profile of coronalplane.In one embodiment, the left end point acquisition process of intervertebral disc includes:Retain maximum lamella or most large stretch of
Split on tomographic image below lung tissue lower limb, more than lower limb and middle column or so the band of 15 pixels in gray scale
Image;Retain pixel position of the Grad less than -10 along coronal image line direction, and Retention area is more than 20 pixels
Value, and length, more than the left end point of the connected domain of 12 pixels (12 pixels in interval), makes above-mentioned meeting in image column direction
Left end point of the pixel of condition for intervertebral disc, and the collection of multiple pixels is combined into profile of the intervertebral disc in coronalplane.
Step 904. according to intervertebral disc coronalplane profile, sagittal plane obtain intervertebral disc left margin, and according to
Left margin extracts vertebral body region.Exemplarily, a full curve can be obtained as fitting to impact point, coronal image is split
As a result three-dimensional sagittal plane segmentation result is converted into, the left margin that the following spinal column of lung tissue is located on sagittal plane is replaced.According to lung group
The right margin for knitting the left margin and aforementioned spinal column of following spinal column can determine that the corresponding disc area of the lumbar vertebra of abdomen area.As schemed
10a is the result schematic diagram for splitting intervertebral disc on embodiment of the present invention sagittal plane, and Figure 10 b are to divide on embodiment of the present invention coronalplane
Cut the result schematic diagram of intervertebral disc.Spine portion below pulmonary it is divisible go out the first intervertebral disc i1, the second intervertebral disc i2, the 3rd
Intervertebral disc i3 and the 4th intervertebral disc i4.Determine the point of multiple intervertebral disc high order ends, the point of this high order end be fitted to into a curve,
It is exactly the left margin of spinal column.
As shown in Fig. 6,7 or 9, flow chart is employed herein for illustrating at data according to an embodiment of the invention
Operating procedure performed by reason system.It should be appreciated that being displayed in the operating procedure of above or below not necessarily in order
To accurately carry out.Conversely, according to inverted order or while various steps can be processed.It is also possible to other operating procedures are added
During being added to these, or a certain step or number step operation are removed from these processes.
If Figure 11 is the sagittal plane view that one embodiment of the invention does not carry out spine segmentation before processing magnetic resonance image (MRI).Its
In, due to the density of hydrone in MR images response imaging tissue, spine edge and lung tissue edge have same or analogous
Gray value, cannot realize the accurate extraction of spinal column only according to pixel value.Such as Figure 12 is that the magnetic after carrying out spine segmentation process is common
Shake image sagittal plane view.In image after process, spinal column and lung tissue, muscular tissue have clearly demarcation line (in figure
Two higher continuous lines of gray value, hence it is evident that projecting pixel), and two demarcation line extend to lumbar region from neck area
Domain, two demarcation line areas encompassed contain cervical vertebral body, thoracic vertebral body and lumbar intervertebral disc;Can according to two demarcation line
Spine regions are extracted accurately from magnetic resonance image (MRI).As shown in figure 13, it is that the embodiment of the present invention splits the spinal column result figure for obtaining.Point
The spinal column for obtaining is cut comprising vertebral body regions such as cervical vertebra, thoracic vertebra, lumbar vertebra, sacral caudal vertebras, also including the disc area at lumbar vertebra position,
And it is clear-cut, even if remaining to realize the automatic segmentation of spinal column in the case where vertebral body bends.
The magnetic resonance image (MRI) spine segmentation method of the present invention is can be applicable in PET/MR multi-mode imaging systems, is schemed based on MR
The information of picture carries out PET correction for attenuations:By MR image segmentations or several regions are divided into, wherein each region is respectively comprising more
The individual voxel for belonging to same tissue or organ, several regions can correspond to skin, lung tissue, soft tissue and skeleton etc. no respectively
With the voxel of organ or tissue;Distribute different attenuation quotients according to several regions that prior information is segmentation, generate decay
Figure;PET data is rebuild according to decay pattern, and obtains the PET image of scanned position.
In one embodiment, the process of magic magiscan process multi-modality medical image includes:Swept using MRI
Imaging apparatus 110 scan person under inspection's organic region, and obtain correspondence organic region MR images, the MR images include multiple voxels or
Pixel;Using PET scan device (such as PET scanner) scanning same organs region, the PET data of correspondence organic region is obtained;
Segmentation is carried out to MR images and obtains multiple regions, wherein the voxel for belonging to same organization type is divided into the same area, it is the plurality of
Region includes lung tissue region, spine regions, rib region, adipose tissue regions etc., wherein spine regions division (or point
Cut) method extraction spinal vertebral region as shown in Figure 6 is can refer to, the intervertebral disc of spinal column is extracted with reference to method as shown in Figure 9
Region;For each pixel (or voxel) the distribution respective attenuation coefficient after splitting in MR images, the first decay pattern is generated;According to first
Decay pattern rebuilds PET data, obtains the first PET image;First decay pattern is updated according to the first PET image, and produces second and declined
Subtract figure;The first PET image is rebuild based on the second decay pattern, obtain the second PET image;Repeat said process until producing
Final target decay pattern (decay pattern estimation) and final goal PET image (PET image estimation).
In existing magnetic resonance image (MRI), be close to due to osseous tissue and the aeriferous lung tissue gray value of bag, more difficult differentiation, Chang Hui
Cause decay pattern to estimate inaccurate problem, and the present embodiment joint is special using the form of the structural information and different views of spinal column
Levy, be capable of achieving the accurate extraction of spinal tissues, effectively realize accurately distinguishing for osseous tissue and lung tissue;Based on magnetic resonance image (MRI)
Accurately segmentation result, is that each voxel distributes correspondence pad value or attenuation quotient, and the decay pattern for obtaining is also more accurate;Using decay
The structure that figure carries out PET data correction is more accurate, is more met the result of tracer actual distribution.
It should be noted that through the above description of the embodiments, those skilled in the art can be understood that
Part or all of to the present invention can be realized by software and with reference to required general hardware platform.Based on such understanding,
Can be embodied in the form of software product the part that technical scheme is substantially contributed to prior art in other words
Out, the computer software product may include one or more machine readable medias for being stored thereon with machine-executable instruction,
These instructions can be caused when by one or more machines execution such as computer, computer network or other electronic systems should
One or more machine embodiments in accordance with the present invention are performing operation.
Although the present invention is disclosed as above with preferred embodiment, so which is not limited to the present invention, any this area skill
Art personnel, without departing from the spirit and scope of the present invention, when a little modification and perfect, therefore the protection model of the present invention can be made
Enclose when by being defined that claims are defined.
Claims (10)
1. in a kind of medical image spinal column dividing method, it is characterised in that include:
Input medical image, and split lung tissue region in the medical image;
Spinal cord line is positioned in the medical image;
According to the spinal cord line and the lung tissue region, determined corresponding to spinal column respectively in the sagittal plane of the medical image
Starting sheet layer and termination lamella;
In the starting sheet layer and termination lamella area encompassed, the first side of spinal column is determined according to the spinal cord line
Boundary, and the second side circle of spinal column is determined according to the spinal cord line and lung tissue region;
Spinal vertebral region is determined based on the starting sheet layer, termination lamella, first side circle and second side circle.
2. in medical image according to claim 1 spinal column dividing method, it is characterised in that in the medical image
Positioning spinal cord line includes:
Skin line segmentation result is obtained in the sagittal plane of the medical image;
Linear enhancement process is carried out to the sagittal plane of the medical image, and in the linearly corresponding medical science figure of enhanced sagittal plane
As each lamella, the first connected domain and the second connected domain are obtained;
First connected domain and the second connected domain are projected, projected image is obtained;
According to the skin line segmentation result, the position of spinal cord line is positioned in the projected image.
3. in medical image according to claim 2 spinal column dividing method, it is characterised in that in the projected image
The position of positioning spinal cord line includes:
Retain the pixel point set in set point apart from the skin line in the projected image, obtain preselected area;
The candidate point of each image line is obtained in the preselected area, the candidate point is the pixel jump point of each image line;
According to candidate's point location spinal cord line.
4. in medical image according to claim 1 spinal column dividing method, it is characterised in that the first side of the spinal column
Border obtains as follows:
The starting sheet layer and the lamella terminated between lamella for being pointed to the spinal column is filtered process;
Obtain the spinal column gray reference value of each image line in the lamella after processing after filtering;
According to the spinal column gray reference value, the first side of the spinal column of each image line is obtained on the right side of the spinal cord line
Boundary.
5. in medical image according to claim 1 spinal column dividing method, it is characterised in that the second side of the spinal column
Border obtains as follows:
In the range of the starting sheet layer and termination topsheet areas of the spinal column, the right hand edge in the lung tissue region is determined;
The spinal cord line is moved to the right hand edge in the lung tissue region, to determine the second side circle of the spinal column.
6. in the medical image according to any one of claim 1-5 spinal column dividing method, it is characterised in that also include:
The maximum lamella corresponding to lung tissue region is extracted in the coronalplane of the medical image;
The left end point of intervertebral disc is determined in the maximum lamella;
Profile of the intervertebral disc in coronalplane is obtained according to the left end point of the intervertebral disc;
According to intervertebral disc in the profile of coronalplane, the left margin of intervertebral disc is obtained in sagittal plane;
Spine segmentation result is obtained according to the left margin of the spinal vertebral region and the intervertebral disc.
7. in medical image according to claim 6 spinal column dividing method, it is characterised in that the determination intervertebral disc
Left end point includes:
Pre-selection connected domain is obtained in the maximum lamella according to coronalplane lung tissue region;
Target connected domain is determined in the pre-selection connected domain, the pixel of the target connected domain meets first in image row direction
Given threshold, meets the second given threshold in image column direction;
Left end point of the left end point of the target connected domain for intervertebral disc.
8. in a kind of medical image spinal column segmenting system, it is characterised in that including spinal vertebral extraction element, the spinal column vertebra
Body extraction element includes:
Segmentation of lung parenchyma unit, for splitting lung tissue region in medical image;
Spinal cord line positioning unit, for spinal cord line is positioned in the medical image;
Border determining unit, for according to the spinal cord line and the lung tissue region, in the sagittal plane point of the medical image
Starting sheet layer that Que Ding be corresponding to spinal column and termination lamella;And,
In the starting sheet layer and termination lamella area encompassed, the first side of spinal column is determined according to the spinal cord line
Boundary, and the second side circle of spinal column is determined according to the spinal cord line and lung tissue region;
Vertebral body area acquisition unit, for being determined based on the starting sheet layer, termination lamella, first side circle and second side circle
Spinal vertebral region.
9. in medical image according to claim 8 spinal column segmenting system, it is characterised in that spinal cord line positioning is single
Unit includes:
Skin line obtains subelement, obtains skin line segmentation result for the sagittal plane in the medical image;
Connected domain obtains subelement, carries out linear enhancement process for the sagittal plane in medical image, and linear enhanced
Corresponding each lamella of sagittal plane, obtains the first connected domain and the second connected domain;
Projected image obtains subelement, for projecting to first connected domain and the second connected domain, obtains projected image;
Locator unit, for according to the skin line segmentation result, positioning the position of spinal cord line in the projected image.
10. in medical image according to claim 8 spinal column segmenting system, it is characterised in that also carry including intervertebral disc
Device is taken, and the intervertebral disc extraction element includes:
Maximum lamella extraction unit, extracts most large stretch of corresponding to lung tissue region for the coronalplane in the medical image
Layer;
Intervertebral disc contours extract unit, for the left end point for determining intervertebral disc in the maximum lamella, and according to the intervertebral
The left end point of disk obtains profile of the intervertebral disc in coronalplane;
According to intervertebral disc in the profile of coronalplane, the left margin of intervertebral disc is obtained in sagittal plane, and carried according to the left margin
Take vertebral body region.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611082125.9A CN106600609B (en) | 2016-11-30 | 2016-11-30 | Spine segmentation method and system in medical image |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611082125.9A CN106600609B (en) | 2016-11-30 | 2016-11-30 | Spine segmentation method and system in medical image |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106600609A true CN106600609A (en) | 2017-04-26 |
CN106600609B CN106600609B (en) | 2020-02-07 |
Family
ID=58594284
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201611082125.9A Active CN106600609B (en) | 2016-11-30 | 2016-11-30 | Spine segmentation method and system in medical image |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106600609B (en) |
Cited By (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107563998A (en) * | 2017-08-30 | 2018-01-09 | 上海联影医疗科技有限公司 | Medical image cardiac image processing method |
CN107844800A (en) * | 2017-07-31 | 2018-03-27 | 深圳联影医疗科技有限公司 | Determination system, the method and device of a kind of optimal sagittal plain of full vertebra |
CN108053400A (en) * | 2017-12-21 | 2018-05-18 | 上海联影医疗科技有限公司 | Image processing method and device |
CN108510507A (en) * | 2018-03-27 | 2018-09-07 | 哈尔滨理工大学 | A kind of 3D vertebra CT image active profile dividing methods of diffusion-weighted random forest |
CN108615250A (en) * | 2018-05-31 | 2018-10-02 | 上海联影医疗科技有限公司 | Image rebuilding method, device, system and computer readable storage medium |
CN109087296A (en) * | 2018-08-07 | 2018-12-25 | 东北大学 | A method of extracting human region in CT image |
CN109360213A (en) * | 2018-12-14 | 2019-02-19 | 浙江工业大学 | A kind of automation centrum recognition methods based on backbone ultrasound coronal image |
CN109493317A (en) * | 2018-09-25 | 2019-03-19 | 哈尔滨理工大学 | The more vertebra dividing methods of 3D based on concatenated convolutional neural network |
CN109615656A (en) * | 2018-11-20 | 2019-04-12 | 哈尔滨理工大学 | A kind of backbone localization method based on pattern search |
CN109978887A (en) * | 2018-12-10 | 2019-07-05 | 深圳市旭东数字医学影像技术有限公司 | Spinal cord automatic division method and its system based on medical image |
CN110599494A (en) * | 2018-06-12 | 2019-12-20 | 中国科学院深圳先进技术研究院 | Rib image reconstruction system and method, terminal and readable storage medium |
CN110675444A (en) * | 2019-09-26 | 2020-01-10 | 东软医疗系统股份有限公司 | Method and device for determining head CT scanning area and image processing equipment |
CN112634218A (en) * | 2020-12-17 | 2021-04-09 | 王跃 | Lumbar muscle measurement method based on transverse position magnetic resonance image |
CN112862830A (en) * | 2021-01-28 | 2021-05-28 | 陕西师范大学 | Multi-modal image segmentation method, system, terminal and readable storage medium |
CN113470004A (en) * | 2021-07-22 | 2021-10-01 | 上海嘉奥信息科技发展有限公司 | Single vertebral body segmentation method, system and medium based on CT |
CN113487591A (en) * | 2021-07-22 | 2021-10-08 | 上海嘉奥信息科技发展有限公司 | CT-based whole spine segmentation method and system |
CN114170114A (en) * | 2021-12-14 | 2022-03-11 | 北京柏惠维康科技有限公司 | Method and device for enhancing spine CT image and spine surgical robot |
CN114187320A (en) * | 2021-12-14 | 2022-03-15 | 北京柏惠维康科技有限公司 | Spine CT image segmentation method and spine imaging identification method and device |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20060251302A1 (en) * | 2005-03-23 | 2006-11-09 | Amer Abufadel | Detection of intervertebral disk orientation in spine images using curve evolution |
US20070173820A1 (en) * | 2006-01-13 | 2007-07-26 | Sdgi Holdings, Inc. | Materials, devices, and methods for treating multiple spinal regions including the anterior region |
CN103606148A (en) * | 2013-11-14 | 2014-02-26 | 深圳先进技术研究院 | Method and apparatus for mixed segmentation of magnetic resonance spine image |
CN105405129A (en) * | 2015-10-31 | 2016-03-16 | 上海联影医疗科技有限公司 | Reconstruction method and device of medical image |
CN105678746A (en) * | 2015-12-30 | 2016-06-15 | 上海联影医疗科技有限公司 | Positioning method and apparatus for the liver scope in medical image |
-
2016
- 2016-11-30 CN CN201611082125.9A patent/CN106600609B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20060251302A1 (en) * | 2005-03-23 | 2006-11-09 | Amer Abufadel | Detection of intervertebral disk orientation in spine images using curve evolution |
US20070173820A1 (en) * | 2006-01-13 | 2007-07-26 | Sdgi Holdings, Inc. | Materials, devices, and methods for treating multiple spinal regions including the anterior region |
CN103606148A (en) * | 2013-11-14 | 2014-02-26 | 深圳先进技术研究院 | Method and apparatus for mixed segmentation of magnetic resonance spine image |
CN105405129A (en) * | 2015-10-31 | 2016-03-16 | 上海联影医疗科技有限公司 | Reconstruction method and device of medical image |
CN105678746A (en) * | 2015-12-30 | 2016-06-15 | 上海联影医疗科技有限公司 | Positioning method and apparatus for the liver scope in medical image |
Cited By (29)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107844800A (en) * | 2017-07-31 | 2018-03-27 | 深圳联影医疗科技有限公司 | Determination system, the method and device of a kind of optimal sagittal plain of full vertebra |
CN107844800B (en) * | 2017-07-31 | 2020-08-07 | 上海联影医疗科技有限公司 | System, method and device for determining optimal sagittal position of whole spine |
CN107563998A (en) * | 2017-08-30 | 2018-01-09 | 上海联影医疗科技有限公司 | Medical image cardiac image processing method |
CN107563998B (en) * | 2017-08-30 | 2020-02-11 | 上海联影医疗科技有限公司 | Method for processing heart image in medical image |
US11129582B2 (en) | 2017-12-21 | 2021-09-28 | Shanghai United Imaging Healthcare Co., Ltd. | System and method for medical imaging of intervertebral discs |
US11826191B2 (en) | 2017-12-21 | 2023-11-28 | Shanghai United Imaging Healthcare Co., Ltd. | System and method for medical imaging of intervertebral discs |
CN108053400B (en) * | 2017-12-21 | 2021-06-15 | 上海联影医疗科技股份有限公司 | Image processing method and device |
CN108053400A (en) * | 2017-12-21 | 2018-05-18 | 上海联影医疗科技有限公司 | Image processing method and device |
CN108510507A (en) * | 2018-03-27 | 2018-09-07 | 哈尔滨理工大学 | A kind of 3D vertebra CT image active profile dividing methods of diffusion-weighted random forest |
CN108615250A (en) * | 2018-05-31 | 2018-10-02 | 上海联影医疗科技有限公司 | Image rebuilding method, device, system and computer readable storage medium |
CN110599494B (en) * | 2018-06-12 | 2021-11-02 | 中国科学院深圳先进技术研究院 | Rib image reconstruction system and method, terminal and readable storage medium |
CN110599494A (en) * | 2018-06-12 | 2019-12-20 | 中国科学院深圳先进技术研究院 | Rib image reconstruction system and method, terminal and readable storage medium |
CN109087296A (en) * | 2018-08-07 | 2018-12-25 | 东北大学 | A method of extracting human region in CT image |
CN109087296B (en) * | 2018-08-07 | 2021-08-10 | 东北大学 | Method for extracting human body region in CT image |
CN109493317A (en) * | 2018-09-25 | 2019-03-19 | 哈尔滨理工大学 | The more vertebra dividing methods of 3D based on concatenated convolutional neural network |
CN109615656A (en) * | 2018-11-20 | 2019-04-12 | 哈尔滨理工大学 | A kind of backbone localization method based on pattern search |
CN109978887A (en) * | 2018-12-10 | 2019-07-05 | 深圳市旭东数字医学影像技术有限公司 | Spinal cord automatic division method and its system based on medical image |
CN109978887B (en) * | 2018-12-10 | 2022-11-11 | 深圳市旭东数字医学影像技术有限公司 | Automatic spinal cord segmentation method and system based on medical image |
CN109360213A (en) * | 2018-12-14 | 2019-02-19 | 浙江工业大学 | A kind of automation centrum recognition methods based on backbone ultrasound coronal image |
CN110675444A (en) * | 2019-09-26 | 2020-01-10 | 东软医疗系统股份有限公司 | Method and device for determining head CT scanning area and image processing equipment |
CN110675444B (en) * | 2019-09-26 | 2023-03-31 | 东软医疗系统股份有限公司 | Method and device for determining head CT scanning area and image processing equipment |
CN112634218A (en) * | 2020-12-17 | 2021-04-09 | 王跃 | Lumbar muscle measurement method based on transverse position magnetic resonance image |
CN112634218B (en) * | 2020-12-17 | 2022-08-23 | 王跃 | Lumbar muscle measurement method based on transverse position magnetic resonance image |
CN112862830A (en) * | 2021-01-28 | 2021-05-28 | 陕西师范大学 | Multi-modal image segmentation method, system, terminal and readable storage medium |
CN112862830B (en) * | 2021-01-28 | 2023-12-22 | 陕西师范大学 | Multi-mode image segmentation method, system, terminal and readable storage medium |
CN113470004A (en) * | 2021-07-22 | 2021-10-01 | 上海嘉奥信息科技发展有限公司 | Single vertebral body segmentation method, system and medium based on CT |
CN113487591A (en) * | 2021-07-22 | 2021-10-08 | 上海嘉奥信息科技发展有限公司 | CT-based whole spine segmentation method and system |
CN114170114A (en) * | 2021-12-14 | 2022-03-11 | 北京柏惠维康科技有限公司 | Method and device for enhancing spine CT image and spine surgical robot |
CN114187320A (en) * | 2021-12-14 | 2022-03-15 | 北京柏惠维康科技有限公司 | Spine CT image segmentation method and spine imaging identification method and device |
Also Published As
Publication number | Publication date |
---|---|
CN106600609B (en) | 2020-02-07 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106600609A (en) | Spine partition method and system in medical image | |
CN106558045B (en) | A kind of segmentation of lung parenchyma method, apparatus, magic magiscan | |
US11869202B2 (en) | Method and system for processing multi-modality image | |
US8787648B2 (en) | CT surrogate by auto-segmentation of magnetic resonance images | |
CN100553561C (en) | The method and apparatus of segmenting structure in the CT angiography | |
US6362821B1 (en) | Surface model generation for visualizing three-dimensional objects using multiple elastic surface nets | |
CN107067398B (en) | Completion method and device for missing blood vessels in three-dimensional medical model | |
CN106683090A (en) | Rib positioning method in medical image and system thereof | |
CN106725570A (en) | Imaging method and system | |
CN107599412A (en) | A kind of three-dimensional modeling method based on institutional framework, system and threedimensional model | |
CN105263407A (en) | Detection of bone tissue using magnetic resonance imaging | |
CN103445801B (en) | For performing the method and system of correction for attenuation | |
US9129391B2 (en) | Semi-automated preoperative resection planning | |
CN112076398A (en) | Radiation therapy system, treatment planning support method, and treatment planning method | |
CN106960439B (en) | A kind of vertebrae identification device and method | |
CN103229210A (en) | Image registration apparatus | |
CN107194925A (en) | Image processing method and system | |
CN105339804B (en) | Divided according to the cortex bone of MR Dixon data | |
CN105678711B (en) | A kind of attenuation correction method based on image segmentation | |
CN107273904A (en) | Image processing method and system | |
CN108694007B (en) | Unfolding ribs from magnetic resonance images | |
CN107204045A (en) | Virtual endoscope system based on CT images | |
CN106650734A (en) | Method for identifying sub areas of locating image, method and device for displaying medical images | |
CN106709925A (en) | Method and device for locating vertebral block in medical image | |
Zhao et al. | Research on 3D reconstruction algorithm of medical CT image based on parallel contour |
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 | ||
CP01 | Change in the name or title of a patent holder | ||
CP01 | Change in the name or title of a patent holder |
Address after: 201807 No. 2258 Chengbei Road, Jiading Industrial Zone, Jiading District, Shanghai. Patentee after: Shanghai Lianying Medical Technology Co., Ltd Address before: 201807 No. 2258 Chengbei Road, Jiading Industrial Zone, Jiading District, Shanghai. Patentee before: SHANGHAI UNITED IMAGING HEALTHCARE Co.,Ltd. |