CN110443868A - PET image acquisition methods, device, computer equipment and storage medium - Google Patents
PET image acquisition methods, device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN110443868A CN110443868A CN201910713405.2A CN201910713405A CN110443868A CN 110443868 A CN110443868 A CN 110443868A CN 201910713405 A CN201910713405 A CN 201910713405A CN 110443868 A CN110443868 A CN 110443868A
- Authority
- CN
- China
- Prior art keywords
- image
- equipment
- deep learning
- learning model
- pet
- 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
- 238000000034 method Methods 0.000 title claims abstract description 57
- 238000013136 deep learning model Methods 0.000 claims abstract description 78
- 238000012549 training Methods 0.000 claims abstract description 55
- 238000012636 positron electron tomography Methods 0.000 claims description 42
- 238000012879 PET imaging Methods 0.000 claims description 40
- 238000004590 computer program Methods 0.000 claims description 28
- 230000004323 axial length Effects 0.000 claims description 10
- 238000005457 optimization Methods 0.000 abstract description 4
- 238000003759 clinical diagnosis Methods 0.000 abstract description 3
- 230000008569 process Effects 0.000 description 20
- 230000035945 sensitivity Effects 0.000 description 8
- 238000013135 deep learning Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 5
- 230000033001 locomotion Effects 0.000 description 5
- 230000029058 respiratory gaseous exchange Effects 0.000 description 4
- 238000002591 computed tomography Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 238000012831 peritoneal equilibrium test Methods 0.000 description 2
- 238000002600 positron emission tomography Methods 0.000 description 2
- 238000012877 positron emission topography Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000000241 respiratory effect Effects 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- 238000010408 sweeping Methods 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- WQZGKKKJIJFFOK-GASJEMHNSA-N Glucose Natural products OC[C@H]1OC(O)[C@H](O)[C@@H](O)[C@@H]1O WQZGKKKJIJFFOK-GASJEMHNSA-N 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 210000000746 body region Anatomy 0.000 description 1
- 230000000747 cardiac effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 150000001875 compounds Chemical class 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 235000014113 dietary fatty acids Nutrition 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 229930195729 fatty acid Natural products 0.000 description 1
- 239000000194 fatty acid Substances 0.000 description 1
- 150000004665 fatty acids Chemical class 0.000 description 1
- 239000008103 glucose Substances 0.000 description 1
- 238000000338 in vitro Methods 0.000 description 1
- 238000001727 in vivo Methods 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 230000004060 metabolic process Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000009659 non-destructive testing Methods 0.000 description 1
- 238000009206 nuclear medicine Methods 0.000 description 1
- 102000039446 nucleic acids Human genes 0.000 description 1
- 108020004707 nucleic acids Proteins 0.000 description 1
- 150000007523 nucleic acids Chemical class 0.000 description 1
- 102000004169 proteins and genes Human genes 0.000 description 1
- 108090000623 proteins and genes Proteins 0.000 description 1
- 238000011524 similarity measure Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 238000012876 topography Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—2D [Two Dimensional] image generation
- G06T11/003—Reconstruction from projections, e.g. tomography
- G06T11/005—Specific pre-processing for tomographic reconstruction, e.g. calibration, source positioning, rebinning, scatter correction, retrospective gating
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
- G06T3/4038—Image mosaicing, e.g. composing plane images from plane sub-images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
-
- 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
- G06T2207/10101—Optical tomography; Optical coherence tomography [OCT]
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
- Nuclear Medicine (AREA)
Abstract
This application involves a kind of PET image acquisition methods, device, computer equipment and storage mediums.The described method includes: obtaining the first image scanned based on the first equipment, obtain the second image scanned based on the second equipment, and the first image and the second image composition training set are trained into deep learning model, the target image scanned based on the first equipment is obtained again, the target image is inputted into the housebroken deep learning model, obtains PET image.The image of the low equipment acquisition of collecting efficiency can be obtained can be directly used for the image of clinical diagnosis using this method by housebroken deep learning model optimization.
Description
Technical field
This application involves medical image technical fields, obtain more particularly to a kind of PET image based on deep learning model
Take method, apparatus, computer equipment and storage medium.
Background technique
Positron emission computed tomography (Positron Emission Tomography, PET) is a kind of utilization
Inside to organism inject positron radioactivity isotope labelling compound, and measure in vitro they spatial distribution and when
Between characteristic three-dimensional imaging non-destructive testing technology, have the characteristics that accuracy is good, accurate positioning.
However, there is inefficiency in current traditional PET imaging device acquisition scans data method.Generally
When traditional PET imaging device is scanned object to be scanned, by way of continuous more beds scannings, acquire it is same to
Scan the multi-group data of the different zones of object.But this mode will obtain clinical diagnosable since collecting efficiency is low
Image needs to guarantee that certain acquisition time could obtain enough sample datas.
Summary of the invention
Based on this, it is necessary in view of the above technical problems, provide a kind of PET image acquisition side based on deep learning model
Method, device, computer equipment and storage medium.
A kind of PET image acquisition methods, which comprises
Obtain the first image scanned based on the first equipment;
Obtain the second image scanned based on the second equipment;
Deep learning model is established, by the first image and the second image composition training set training deep learning mould
Type obtains housebroken deep learning module;
Obtain the target image scanned based on the first equipment;
The target image is inputted into the housebroken deep learning model, obtains PET image.
In one of the embodiments, it is characterized in that, the method also includes:
The first image and the second image are being formed into training set, it, will be described before training the deep learning model
First image and second image carry out image registration, and described image registration includes Rigid Registration and non-rigid registration.
It is described short in one of the embodiments, it is characterized in that, first equipment is short wheelbase PET imaging device
The axial length of wheelbase PET imaging device is between 10cm between 50cm.
In one of the embodiments, it is characterized in that, second equipment is long wheelbase PET imaging device, the length
The axial length of wheelbase PET imaging device is between 50cm between 230cm.
In one of the embodiments, it is characterized in that, the first image and the second image form training set, training is deep
Spending learning model includes:
Using the first image of same sweep object and the second image as one group of image sequence;
Obtain the multiple groups described image sequence of multiple sweep objects;
Using the first image in image sequence as input, using the second image in same image group as output, training
The deep learning model.
In one of the embodiments, it is characterized in that, the first image is the more beds obtained based on the first equipment
Stitching image.
In one of the embodiments, it is characterized in that, second image is the single bed obtained based on the second equipment
Image.
A kind of PET image acquisition device, described device include:
First obtains module, for obtaining the first image scanned based on the first equipment;
Second obtains module, for obtaining the second image scanned based on the second equipment;
Training module establishes deep learning model, for training the first image and the second image composition training set
The deep learning model obtains housebroken deep learning model;
Third obtains module, for obtaining the target image scanned based on the first equipment;
Module is obtained, for the target image to be inputted the housebroken deep learning model, obtains PET image.
A kind of computer equipment, including memory and processor, the memory are stored with computer program, the processing
Device performs the steps of when executing the computer program
Obtain the first image scanned based on the first equipment;
Obtain the second image scanned based on the second equipment;
Deep learning model is established, by the first image and the second image composition training set training deep learning mould
Type obtains housebroken deep learning model;
Obtain the target image scanned based on the first equipment;
The target image is inputted into the housebroken deep learning model, obtains PET image.
A kind of computer readable storage medium, is stored thereon with computer program, and the computer program is held by processor
It is performed the steps of when row
Obtain the first image scanned based on the first equipment;
Obtain the second image scanned based on the second equipment;
Deep learning model is established, by the first image and the second image composition training set training deep learning mould
Type obtains housebroken deep learning model;
Obtain the target image scanned based on the first equipment;
The target image is inputted into the housebroken deep learning model, obtains PET image.
Above-mentioned PET image acquisition methods, device, computer equipment and storage medium are swept by obtaining based on the first equipment
The first image retouched, obtains the second image for scanning based on the second equipment, and by the first image and the second figure
As composition training set, training deep learning model, then the target image scanned based on the first equipment is obtained, by the target
Image inputs housebroken deep learning model, obtains PET image.It will scan to obtain based on the first low equipment of collecting efficiency
First image, and depth training pattern is instructed based on the second image that higher second equipment of collecting efficiency scans
Practice.The target image scanned based on the first equipment is inputted into trained deep learning model again, thus acquisition and base
It scans to obtain the consistent PET image of picture quality in higher second equipment of collecting efficiency.Pass through target image and instruction in this way
Deep learning model after white silk, obtains PET image, image is made to can be used for clinical diagnosis.
Detailed description of the invention
Fig. 1 is the flow diagram of PET image acquisition methods in one embodiment;
Fig. 2 is the structural block diagram of PET image acquisition device in one embodiment;
Fig. 3 is the structural block diagram of deep learning model training apparatus in one embodiment;
Fig. 4 is the internal structure chart of computer equipment in one embodiment.
Specific embodiment
It is with reference to the accompanying drawings and embodiments, right in order to which the objects, technical solutions and advantages of the application are more clearly understood
The application is further elaborated.It should be appreciated that specific embodiment described herein is only used to explain the application, not
For limiting the application.
Positron e mission computed tomography (Positron Emission Tomography, PET), is nuclear medicine
The more advanced clinical examination image technology in field.It is the usually necessary substance in biological life metabolism by certain substance,
Such as: glucose, protein, nucleic acid, fatty acid, short-life radionuclide (such as 18F, 11C etc.) on label inject human body
Afterwards, radionuclide releases positive electron in decay process, and a positive electron is in a few tenths of millimeter of advancing to after several millimeters
Bury in oblivion after encountering an electronics, to generate the photon that contrary a pair of of energy is 511KeV.This passes through photon
Super-sensitive detector captures to form data information, these information are scattered through computer and a series of correction such as random
Corrected data is formed, by carrying out reconstruction processing to corrected data, we are available to assemble situation in vivo
3-D image, to achieve the purpose that diagnosis.
Existing PET imaging device limited axial extent, (PET scan is alternatively referred to as a bed to each PET scan
Position, hereafter referred to as bed) during, PET imaging device with respect to patient at rest, so, when progress body scan or large area
When scanning, the scanning for carrying out multiple beds is generally required.According to Patient height and the axial covering model of Solid-state pet detector system
It encloses, general body scan needs to carry out the PET scan of 5-8 bed.Between two neighboring bed, it is scanning bed can be along axial traveling
A distance, while guaranteeing the axial overlap for having certain between two neighboring bed.When the mistake of scanning bed each bed of movement
Cheng Zhong, the data of each bed of PET system scanning collection.However, the scanning obtained by the PET imaging device of limited axial extent
Effective sample data in data is simultaneously inadequate, causes picture quality not high, cannot be directly used to clinical diagnosis.
In one embodiment, as shown in Figure 1, providing a kind of PET image acquisition methods based on deep learning model,
It the described method comprises the following steps:
Step 201, the first image scanned based on the first equipment is obtained.
In step 201, wherein the first equipment is short wheelbase PET imaging device, the axis of the short wheelbase PET imaging device
To length between 10cm between 50cm.Since the axial length of PET imaging device is shorter, during the scanning process, can only obtain
The scan data in a part of region of patient body.When needing to be scanned entire patient's large area region, then need sweeping
During retouching, short wheelbase PET imaging device is made to be axially moveable multiple beds, to obtain patient corresponding with each bed position
The scan data of body region.The data splicing reconstruction of multiple beds is obtained into the first image again.Also, in this way using short wheelbase
Scanning process needs to continue longer time.
Specifically, the first obtained image is the more bed stitching images obtained based on the first equipment.In the first equipment pair
During patient is scanned, the scan data of patient's different zones is obtained respectively by the scanning of multiple and different beds.It is right again
Each bed data carry out image reconstruction respectively, obtain the image in the corresponding patient scan region with each bed position.It will obtain again
Each bed image according to according to obtain scan data sequence carry out image mosaic, obtain entire patient's image.Utilizing the
One equipment to patient carry out body scan when, due to scanning will continue longer time, probably need 10 to 15 minutes when
Between, and patient can generate the change of chest area position because of respiratory movement in this process.In this way, each bed data are
It is collected under the different respiratory movement stages.It is pseudo- that the image obtained after rebuilding to each bed data will will appear breathing
Shadow etc. influences the problem of picture quality.And since sweep time is long, patient may not be able to maintain the same position for a long time, or
The same posture of person, it is inadequate so as to cause sample data effective in the data of acquisition situations such as, to influence final image matter
Amount.Therefore, the first image is muting sensitivity image.
Step 202, the second image scanned based on the second equipment is obtained.
In step 202, wherein the second equipment is long wheelbase PET imaging device, the axis of the long wheelbase PET imaging device
To length between 50cm between 230cm.It, during the scanning process, can be by one since the axial length of PET imaging device is longer
A bed scanning obtains the scan data of entire patient's large area.Again the data that the scanning of single bed obtains are rebuild to obtain
Second image.Also, it is shorter being scanned process to patient using long wheelbase PET imaging device.
Further, the second image rebuild is the single bed image obtained based on the second equipment.It sets due to second
Standby is long wheelbase PET imaging device, when the length of wheelbase PET imaging device is 230cm on the spot, it is only necessary to scan single bed just
The scan data of available entire patient.Since scanning process is shorter, be then conducive to patient's holding position in the scanning process
It sets and posture is constant, effective sample data in such scan data is more sufficient, to guarantee picture quality.Also, it is acquiring
When scan data, patient is in same respiration phase, and the second obtained image will be less by the influence breathed, to protect
The second picture quality is demonstrate,proved.
It should be noted that, when needing to be scanned entire patient, also being needed when the length of the second equipment is 50 centimetres
Obtain the scan data of multiple beds.But when carrying out whole body to the same patient using the first equipment and the second equipment
When scanning, the time that the second equipment scanning process needs is more shorter than the time that the first equipment scanning process needs, so that the
Effective sample data in data that the scanning of two equipment obtains is more more sufficient than the scanning acquisition of the first equipment, and the first image
The problem of middle respiration artefacts, is even more serious than the second image.Therefore in the actual operation process, the image obtained based on the second equipment
The picture quality that the first equipment of mass ratio obtains is more high-quality.Therefore, second image is high temperature sensitivity.
Step 203, deep learning model is established, by the first image and the second image composition training set training depth
Learning model is spent, housebroken deep learning model is obtained.
It before step 203, further include that the first image and second image are subjected to image registration, described image
Registration includes Rigid Registration and non-rigid registration.
In the present embodiment, after being spliced due to image that the first image is multiple beds obtained entire patient or
Topography, and the second image is the single image of entire patient or part.It is formed using the first image and the second image
Before training set, need to indicate in the information and the second image that indicate patients body location in the first image by image registration
The information of patient body same position corresponds.That is, make the first equipment and the second equipment to patient's same position into
The information obtained when row scanning, is registrated in the first image and the second image.
Further, image registration includes Rigid Registration.The Rigid Registration refers to during being scanned, Huan Zhezhu
Dynamic movement or the variation of posture are so that difference occur in the first image and the second image.It, can be shorter by scanning process before registration
The second image as reference picture, using longer first image of scanning process as floating image.It is found by image registration
The coordinate transform optimal to one, so that the similarity measure between floating image and reference picture is maximum.
Further, image registration includes non-rigid registration.The non-rigid registration is patient during being scanned
Because of the difference that the first image and the second image occur in breathing and cardiac motion.When carrying out non-rigid registration, usually have
Two class method for registering.The first kind is the method for registering based on mutual information measure.Second class method passes through picture structure characterizing method
Multimode image registering is reduced to single mode image registration, such as passes through entropy diagram, weber local feature description son (Weber Local
Descriptor, WLD) and based on mode independence neighborhood description son (Modality Independent Neighborhood
Descriptor, MIND) etc. features picture structure is characterized, then using characterization result squared difference and (Sum of
Squared Difference, SSD) it is used as water rogulator to realize image registration.
Step 203 further include: using the first image of same sweep object and the second image as one group of image sequence;It obtains
The multiple groups described image sequence of multiple sweep objects;Using the first image in image sequence as input, in same image group
The second image as output, the training deep learning model.
Specifically, the first image and the second image are to be utilized respectively the first equipment to same sweep object and second set
For what is be scanned.And the first image and the second image at this time is by the image after registration.By same scanning
The first image and the second image that object scan obtains are as one group of image sequence.It is trained to deep learning model
When, using the first image obtained in image sequence by short wheelbase PET imaging device as the input of deep learning model, by image
Output of the second image obtained in sequence by long wheelbase PET imaging device as deep learning model.
It further, can be by being utilized respectively first to multiple sweep objects when being trained to deep learning model
The multiple series of images sequence that equipment and the second equipment obtain, repeatedly trains deep learning model.
Step 204, the target image scanned based on the first equipment is obtained.
In the present embodiment, the target image obtained is to be based on short wheelbase PET imaging device pair in actual clinical operation
What patient was scanned.During the scanning process, after being scanned by the first equipment to patient, sweeping for multiple beds is obtained
Data are retouched, and each scan data is rebuild respectively, obtain multiple bed images, then are obtained after each bed image is spliced
To target image.At this point, obtained target image is muting sensitivity image, and it is second-rate clinical image, it cannot be directly sharp
Patient is diagnosed with target image.
Step 205, the target image is inputted into housebroken deep learning model, obtains PET image.
In the present embodiment, target image obtained in step 204 is inputted into trained deep learning model.Due to
It is the first image by obtaining short axle PET imaging device as input, by long axis PET in training deep learning model
The second image that imaging device obtains has deep learning model by the energy of first the second image of image optimization as output
Power.And when being trained to deep learning model, multiple series of images sequence is used to be trained, so that by deep learning
The first picture quality and the second image after model optimization is more close.Depth in this way by target image input after trained
Learning model obtains highly sensitive PET image.The PET image is used directly for diagnosing patient.
It should be noted that step 201 instructs deep learning model with the second image using the first image to 203
Experienced process, and step 204 and step 205 are to be used in clinic the deep learning model of trained mistake
Process.Above-mentioned two process can not be completed in a period, actually faced the process that deep learning model is trained
Bed operation before has completed, deep learning model can also be trained repeatedly, make optimized first image without
It limits close to the second image.
In above-mentioned PET image acquisition methods, the first image scanned based on the first equipment is obtained, obtains and is based on second
The first image and the second image are formed training set, training deep learning model by the second image that equipment scans;It obtains
The target image scanned based on the first equipment is taken, the target image is inputted into housebroken deep learning model, is obtained
PET image.In the method, the first image is what short wheelbase PET imaging device obtained, therefore the first image is muting sensitivity figure
Picture cannot be used directly to diagnose patient, and what the second image obtained for long wheelbase PET imaging device, therefore the second figure
As being high temperature sensitivity, can be used directly to diagnose patient.Deep learning model is trained again, utilizes first
Input of the image as deep learning model, output of second image as deep learning model, make it is trained after depth
Habit model has the ability by the first image optimization for the second image.In clinical manipulation, obtained based on short wheelbase PET imaging device
The target image taken is muting sensitivity image, can be direct by the way that the target image is inputted that housebroken deep learning model obtains
For the PET image diagnosed to patient.Such method can lead to the client for using short wheelbase PET imaging device at present
It crosses depth learning model to optimize muting sensitivity image, to obtain the height scanned with long wheelbase PET imaging device
The PET image of sensitivity image uniform quality.
It should be understood that although each step in the flow chart of Fig. 1 is successively shown according to the instruction of arrow, this
A little steps are not that the inevitable sequence according to arrow instruction successively executes.Unless expressly state otherwise herein, these steps
It executes there is no the limitation of stringent sequence, these steps can execute in other order.Moreover, at least part in Fig. 1
Step may include that perhaps these sub-steps of multiple stages or stage are executed in synchronization to multiple sub-steps
It completes, but can execute at different times, the execution sequence in these sub-steps or stage, which is also not necessarily, successively to be carried out,
But it can be executed in turn or alternately at least part of the sub-step or stage of other steps or other steps.
In one embodiment, as shown in Fig. 2, providing a kind of PET image acquisition device, comprising: first obtains module
301, second module 302, training module 303, third acquisition module 304 is obtained and obtains module 305, in which:
First obtains module 301, for obtaining the first image scanned based on the first equipment.
Second obtains module 302, for obtaining the second image scanned based on the second equipment.
Training module 303, for establishing deep learning model, by the first image and the composition training training of the second image
Practice the deep learning model, obtains housebroken deep learning model.
Third obtains module 304, for obtaining the target image scanned based on the first equipment.
Module 305 is obtained, for the target image to be inputted the housebroken deep learning model, obtains PET figure
Picture.
In one embodiment, training pattern 303 is also used to: training set is formed in the first image and the second image,
Before training deep learning model, the first image and second image are subjected to image registration, described image registration packet
Include Rigid Registration and non-rigid registration.
In one embodiment, as shown in figure 3, providing a kind of deep learning model training apparatus, comprising: image sequence
Comprising modules 401, multiple series of images retrieval module 402 and deep learning model training module 403, wherein
Image sequence comprising modules 401, for using the first image of same sweep object and the second image as a group picture
As sequence.
Multiple series of images retrieval module 402, for obtaining the multiple groups described image sequence of multiple sweep objects.
Deep learning model training module 403, for using the first image in image sequence as input, with same image
The second image in group is as output, the training deep learning model.
Specific about PET image acquisition device limits the restriction that may refer to above for PET image acquisition methods,
Details are not described herein.Modules in above-mentioned PET image acquisition device can be fully or partially through software, hardware and combinations thereof
To realize.Above-mentioned each module can be embedded in the form of hardware or independently of in the processor in computer equipment, can also be with soft
Part form is stored in the memory in computer equipment, executes the corresponding behaviour of the above modules in order to which processor calls
Make.
In one embodiment, a kind of computer equipment is provided, which can be terminal, internal structure
Figure can be as shown in Figure 3.The computer equipment includes processor, the memory, network interface, display connected by system bus
Screen and input unit.Wherein, the processor of the computer equipment is for providing calculating and control ability.The computer equipment is deposited
Reservoir includes non-volatile memory medium, built-in storage.The non-volatile memory medium is stored with operating system and computer journey
Sequence.The built-in storage provides environment for the operation of operating system and computer program in non-volatile memory medium.The calculating
The network interface of machine equipment is used to communicate with external terminal by network connection.When the computer program is executed by processor with
Realize a kind of PET image acquisition methods.The display screen of the computer equipment can be liquid crystal display or electric ink is shown
Screen, the input unit of the computer equipment can be the touch layer covered on display screen, be also possible on computer equipment shell
Key, trace ball or the Trackpad of setting can also be external keyboard, Trackpad or mouse etc..
It will be understood by those skilled in the art that structure shown in Fig. 3, only part relevant to application scheme is tied
The block diagram of structure does not constitute the restriction for the computer equipment being applied thereon to application scheme, specific computer equipment
It may include perhaps combining certain components or with different component layouts than more or fewer components as shown in the figure.
In one embodiment, a kind of computer equipment, including memory and processor are provided, is stored in memory
Computer program, the processor perform the steps of when executing computer program
Obtain the first image scanned based on the first equipment;
Obtain the second image scanned based on the second equipment;
Deep learning model is established, by the first image and the second image composition training set training deep learning mould
Type obtains housebroken deep learning model;
Obtain the target image scanned based on the first equipment;
The target image is inputted into the housebroken deep learning model, obtains PET image.
In one embodiment, it is also performed the steps of when processor executes computer program
Training set is formed in the first image and the second image, before training deep learning model, by first figure
As carrying out image registration with second image, described image registration includes Rigid Registration and non-rigid registration.
In one embodiment, it is also performed the steps of when processor executes computer program
First equipment is short wheelbase PET imaging device, the axial length of the short wheelbase PET imaging device between
10cm is between 50cm.
In one embodiment, it is also performed the steps of when processor executes computer program
Second equipment is long wheelbase PET imaging device, the axial length of the long wheelbase PET imaging device between
50cm is between 230cm.
In one embodiment, it is also performed the steps of when processor executes computer program
Using the first image of same sweep object and the second image as one group of image sequence;
Obtain the multiple groups described image sequence of multiple sweep objects;
Using the first image in image sequence as input, using the second image in same image group as output, training
The deep learning model.
In one embodiment, it is also performed the steps of when processor executes computer program
The first image is the more bed stitching images obtained based on the first equipment.
In one embodiment, it is also performed the steps of when processor executes computer program
Second image is the single bed image obtained based on the second equipment.
In one embodiment, a kind of computer readable storage medium is provided, computer program is stored thereon with, is calculated
Machine program performs the steps of when being executed by processor
Obtain the first image scanned based on the first equipment;
Obtain the second image scanned based on the second equipment;
Deep learning model is established, by the first image and the second image composition training set training deep learning mould
Type obtains housebroken deep learning model;
Obtain the target image scanned based on the first equipment;
The target image is inputted into the housebroken deep learning model, obtains PET image.
In one embodiment, it is also performed the steps of when computer program is executed by processor
Training set is formed in the first image and the second image, before training deep learning model, by first figure
As carrying out image registration with second image, described image registration includes Rigid Registration and non-rigid registration.
In one embodiment, it is also performed the steps of when computer program is executed by processor
First equipment is short wheelbase PET imaging device, the axial length of the short wheelbase PET imaging device between
10cm is between 50cm.
In one embodiment, it is also performed the steps of when computer program is executed by processor
Second equipment is long wheelbase PET imaging device, the axial length of the long wheelbase PET imaging device between
50cm is between 230cm.
In one embodiment, it is also performed the steps of when computer program is executed by processor
Using the first image of same sweep object and the second image as one group of image sequence;
Obtain the multiple groups described image sequence of multiple sweep objects;
Using the first image in image sequence as input, using the second image in same image group as output, training
The deep learning model.
In one embodiment, it is also performed the steps of when computer program is executed by processor
The first image is the more bed stitching images obtained based on the first equipment.
In one embodiment, it is also performed the steps of when computer program is executed by processor
Second image is the single bed image obtained based on the second equipment.
Those of ordinary skill in the art will appreciate that realizing all or part of the process in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, the computer program can be stored in a non-volatile computer
In read/write memory medium, the computer program is when being executed, it may include such as the process of the embodiment of above-mentioned each method.Wherein,
To any reference of memory, storage, database or other media used in each embodiment provided herein,
Including non-volatile and/or volatile memory.Nonvolatile memory may include read-only memory (ROM), programming ROM
(PROM), electrically programmable ROM (EPROM), electrically erasable ROM (EEPROM) or flash memory.Volatile memory may include
Random access memory (RAM) or external cache.By way of illustration and not limitation, RAM is available in many forms,
Such as static state RAM (SRAM), dynamic ram (DRAM), synchronous dram (SDRAM), double data rate sdram (DDRSDRAM), enhancing
Type SDRAM (ESDRAM), synchronization link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM
(RDRAM), direct memory bus dynamic ram (DRDRAM) and memory bus dynamic ram (RDRAM) etc..
Each technical characteristic of above embodiments can be combined arbitrarily, for simplicity of description, not to above-described embodiment
In each technical characteristic it is all possible combination be all described, as long as however, the combination of these technical characteristics be not present lance
Shield all should be considered as described in this specification.
The several embodiments of the application above described embodiment only expresses, the description thereof is more specific and detailed, but simultaneously
It cannot therefore be construed as limiting the scope of the patent.It should be pointed out that coming for those of ordinary skill in the art
It says, without departing from the concept of this application, various modifications and improvements can be made, these belong to the protection of the application
Range.Therefore, the scope of protection shall be subject to the appended claims for the application patent.
Claims (10)
1. a kind of PET image acquisition methods, which is characterized in that the described method includes:
Obtain the first image scanned based on the first equipment;
Obtain the second image scanned based on the second equipment;
Deep learning model is established, the first image and the second image composition training set are trained into the deep learning model,
Obtain housebroken deep learning model;
Obtain the target image scanned based on the first equipment;
The target image is inputted into the housebroken deep learning model, obtains PET image.
2. the method according to claim 1, wherein the method also includes:
The first image and the second image are being formed into training set, before training the deep learning model, by described first
Image and second image carry out image registration, and described image registration includes Rigid Registration and non-rigid registration.
3. the method according to claim 1, wherein first equipment be short wheelbase PET imaging device, it is described
The axial length of short wheelbase PET imaging device is between 10cm between 50cm.
4. the method according to claim 1, wherein second equipment be long wheelbase PET imaging device, it is described
The axial length of long wheelbase PET imaging device is between 50cm between 230cm.
5. the method according to claim 1, wherein the first image and the second image form training set, instruction
Practicing deep learning model includes:
Using the first image of same sweep object and the second image as one group of image sequence;
Obtain the multiple groups described image sequence of multiple sweep objects;
Using the first image in image sequence as input, using the second image in same image group as output, described in training
Deep learning model.
6. the method according to claim 1, wherein the first image is more obtained based on the first equipment
Position stitching image.
7. the method according to claim 1, wherein second image is single obtained based on the second equipment
Bit image.
8. a kind of PET image acquisition device, which is characterized in that described device includes:
First obtains module, for obtaining the first image scanned based on the first equipment;
Second obtains module, for obtaining the second image scanned based on the second equipment;
Training module will be described in the first image and the composition training set training of the second image for establishing deep learning model
Deep learning model obtains housebroken deep learning model;
Third obtains module, for obtaining the target image scanned based on the first equipment;
Module is obtained, for the target image to be inputted the housebroken deep learning model, obtains PET image.
9. a kind of computer equipment, including memory and processor, the memory are stored with computer program, feature exists
In the step of processor realizes any one of claims 1 to 7 the method when executing the computer program.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the computer program
The step of method described in any one of claims 1 to 7 is realized when being executed by processor.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910713405.2A CN110443868B (en) | 2019-08-02 | 2019-08-02 | PET image acquisition method, PET image acquisition device, computer equipment and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910713405.2A CN110443868B (en) | 2019-08-02 | 2019-08-02 | PET image acquisition method, PET image acquisition device, computer equipment and storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110443868A true CN110443868A (en) | 2019-11-12 |
CN110443868B CN110443868B (en) | 2023-09-26 |
Family
ID=68433101
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910713405.2A Active CN110443868B (en) | 2019-08-02 | 2019-08-02 | PET image acquisition method, PET image acquisition device, computer equipment and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110443868B (en) |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
SU1518744A2 (en) * | 1987-11-19 | 1989-10-30 | Научно-Исследовательский Институт Интроскопии | Computing tomograph |
JPH07229969A (en) * | 1994-02-17 | 1995-08-29 | Hitachi Medical Corp | Positron ct apparatus |
CN1242976A (en) * | 1998-07-24 | 2000-02-02 | 深圳安科高技术有限公司 | Computer tomoscanning device |
CN104287760A (en) * | 2013-07-16 | 2015-01-21 | 上海联影医疗科技有限公司 | Computed tomography control device and method |
CN105389788A (en) * | 2015-10-13 | 2016-03-09 | 沈阳东软医疗系统有限公司 | Positron emission tomography (PET) multi-bed image reconstruction method, PET multi-bed image reconstruction device, PET multi-bed image combining method and PET multi-bed image combining device |
CN107115119A (en) * | 2017-04-25 | 2017-09-01 | 上海联影医疗科技有限公司 | The acquisition methods of PET image attenuation coefficient, the method and system of correction for attenuation |
CN108416755A (en) * | 2018-03-20 | 2018-08-17 | 南昌航空大学 | A kind of image de-noising method and system based on deep learning |
CN109360152A (en) * | 2018-10-15 | 2019-02-19 | 天津大学 | 3 d medical images super resolution ratio reconstruction method based on dense convolutional neural networks |
-
2019
- 2019-08-02 CN CN201910713405.2A patent/CN110443868B/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
SU1518744A2 (en) * | 1987-11-19 | 1989-10-30 | Научно-Исследовательский Институт Интроскопии | Computing tomograph |
JPH07229969A (en) * | 1994-02-17 | 1995-08-29 | Hitachi Medical Corp | Positron ct apparatus |
CN1242976A (en) * | 1998-07-24 | 2000-02-02 | 深圳安科高技术有限公司 | Computer tomoscanning device |
CN104287760A (en) * | 2013-07-16 | 2015-01-21 | 上海联影医疗科技有限公司 | Computed tomography control device and method |
CN105389788A (en) * | 2015-10-13 | 2016-03-09 | 沈阳东软医疗系统有限公司 | Positron emission tomography (PET) multi-bed image reconstruction method, PET multi-bed image reconstruction device, PET multi-bed image combining method and PET multi-bed image combining device |
CN107115119A (en) * | 2017-04-25 | 2017-09-01 | 上海联影医疗科技有限公司 | The acquisition methods of PET image attenuation coefficient, the method and system of correction for attenuation |
CN108416755A (en) * | 2018-03-20 | 2018-08-17 | 南昌航空大学 | A kind of image de-noising method and system based on deep learning |
CN109360152A (en) * | 2018-10-15 | 2019-02-19 | 天津大学 | 3 d medical images super resolution ratio reconstruction method based on dense convolutional neural networks |
Non-Patent Citations (2)
Title |
---|
杨义: "正电子发射断层扫描仪(PET)中数据通信的实现", 《中国优秀硕士学位论文全文数据库(电子期刊)工程科技而Ⅱ辑》, pages 030 - 58 * |
袁圆: "基于医用CT图像的配准方法及精度研究", 《中国优秀硕士学位论文全文数据库(电子期刊)信息科技辑》, pages 138 - 1566 * |
Also Published As
Publication number | Publication date |
---|---|
CN110443868B (en) | 2023-09-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11020077B2 (en) | Simultaneous CT-MRI image reconstruction | |
CN105078495B (en) | PET-CT scanning imaging method and related imaging method | |
CN109308728B (en) | Positron emission computed tomography image processing method and device | |
CN110960241A (en) | Method and device for determining scanning parameters of medical image scanning and computer equipment | |
CN104220898B (en) | Method for generating PET absorption-corrected image from MR image | |
CN110996800B (en) | System, method, and non-transitory computer readable medium for determining PET imaging kinetic parameters | |
CN109961419B (en) | Correction information acquisition method for attenuation correction of PET activity distribution image | |
CN110742632B (en) | PET image attenuation correction method and PET-CT equipment | |
CN1895185B (en) | Method for displaying information of checking region of checked object and influence of drug in vivo. | |
CN110136076A (en) | Medical scanning imaging method, device, storage medium and computer equipment | |
CN109199422A (en) | CT preview image rebuilds optimization method, device, computer equipment and storage medium | |
CN110223247B (en) | Image attenuation correction method, device, computer equipment and storage medium | |
CN110507349A (en) | Automatic scanning method, device, computer equipment and computer readable storage medium | |
CN110215227A (en) | Time window setting method, device, computer equipment and storage medium | |
CN110269590A (en) | Pharmacokinetic parameter determines method, apparatus, computer equipment and storage medium | |
CN111402356B (en) | Parameter imaging input function extraction method and device and computer equipment | |
CN110299199A (en) | Medical image scan method, device, computer equipment and storage medium | |
CN110215226B (en) | Image attenuation correction method, image attenuation correction device, computer equipment and storage medium | |
CN110742639B (en) | Scanning system configuration method, scanning system configuration device, computer equipment and readable storage medium | |
CN110443868A (en) | PET image acquisition methods, device, computer equipment and storage medium | |
CN110215203A (en) | Electrocardiosignal acquisition methods, device, computer equipment and storage medium | |
CN107569251A (en) | Medical imaging procedure and system and non-transient computer readable storage medium storing program for executing | |
CN110811665A (en) | PET image attenuation correction method, apparatus, computer device and storage medium | |
CN111700633A (en) | PET image attenuation correction method, PET image attenuation correction device, computer equipment and readable storage medium | |
CN110507343B (en) | Playground acquisition method, PET image correction device and computer equipment |
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 | ||
CB02 | Change of applicant information | ||
CB02 | Change of applicant information |
Address after: 201807 Shanghai City, north of the city of Jiading District Road No. 2258 Applicant after: Shanghai Lianying Medical Technology Co.,Ltd. Address before: 201807 Shanghai City, north of the city of Jiading District Road No. 2258 Applicant before: SHANGHAI UNITED IMAGING HEALTHCARE Co.,Ltd. |
|
GR01 | Patent grant | ||
GR01 | Patent grant |