CN109272472A - Noise and artifact eliminating method towards medical power spectrum CT image - Google Patents
Noise and artifact eliminating method towards medical power spectrum CT image Download PDFInfo
- Publication number
- CN109272472A CN109272472A CN201811198845.0A CN201811198845A CN109272472A CN 109272472 A CN109272472 A CN 109272472A CN 201811198845 A CN201811198845 A CN 201811198845A CN 109272472 A CN109272472 A CN 109272472A
- Authority
- CN
- China
- Prior art keywords
- image
- power spectrum
- ray
- training
- noise
- 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
- 238000001228 spectrum Methods 0.000 title claims abstract description 30
- 238000000034 method Methods 0.000 title claims abstract description 28
- 238000012549 training Methods 0.000 claims abstract description 26
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 22
- 238000013170 computed tomography imaging Methods 0.000 claims abstract description 11
- 238000012360 testing method Methods 0.000 claims abstract description 8
- 230000000694 effects Effects 0.000 claims abstract description 7
- 238000002591 computed tomography Methods 0.000 claims abstract description 6
- 238000013528 artificial neural network Methods 0.000 claims description 13
- 230000008569 process Effects 0.000 claims description 13
- 239000008280 blood Substances 0.000 claims description 3
- 210000004369 blood Anatomy 0.000 claims description 3
- 210000000988 bone and bone Anatomy 0.000 claims description 3
- 239000000470 constituent Substances 0.000 claims description 3
- 238000001514 detection method Methods 0.000 claims description 3
- 210000004185 liver Anatomy 0.000 claims description 3
- 210000004072 lung Anatomy 0.000 claims description 3
- 210000005075 mammary gland Anatomy 0.000 claims description 3
- 238000013441 quality evaluation Methods 0.000 claims description 3
- 238000004088 simulation Methods 0.000 claims description 3
- 210000004872 soft tissue Anatomy 0.000 claims description 3
- 230000001537 neural effect Effects 0.000 claims description 2
- 208000027418 Wounds and injury Diseases 0.000 abstract description 3
- 230000006378 damage Effects 0.000 abstract description 3
- 208000014674 injury Diseases 0.000 abstract description 3
- 238000004519 manufacturing process Methods 0.000 abstract 1
- 230000005855 radiation Effects 0.000 description 5
- 238000003062 neural network model Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 238000002360 preparation method Methods 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 101100536251 Mus musculus Tmem120a gene Proteins 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000002059 diagnostic imaging Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 239000007788 liquid Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000010408 sweeping Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
- A61B6/52—Devices using data or image processing specially adapted for radiation diagnosis
- A61B6/5258—Devices using data or image processing specially adapted for radiation diagnosis involving detection or reduction of artifacts or noise
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- General Physics & Mathematics (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Medical Informatics (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- High Energy & Nuclear Physics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Optics & Photonics (AREA)
- Pathology (AREA)
- Radiology & Medical Imaging (AREA)
- Heart & Thoracic Surgery (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The present invention relates to computed tomography fields, influence brought by beam hardening, random noise to reduce, export more accurate medical image, reduce injury suffered by human body.For this purpose, of the invention, noise and artifact eliminating method towards medical power spectrum CT image, steps are as follows: Step1: building Dummy mould；Step2: power spectrum CT imaging is simulated by body mould with a large amount of x-ray photons；Step3: power spectrum CT imaging is simulated by body mould with relatively low-dose x-ray photon；Step4: reconstruction image is matched；Step5: training convolutional neural networks complete training；Step6: test network training effect.Present invention is mainly applied to the design and manufacture occasions of medical CT equipment.
Description
Technical field
The present invention relates to the fields computed tomography (Computed Tomography), for power spectrum CT different-energy
The lower problem of signal-to-noise ratio caused by the number of photons collected in section is less, using convolutional neural networks model to each energy section
Reconstruction image with noise and artifact is handled, and reduces dose of radiation while improving the precision of medical image.Specifically,
It is related to noise and artifact eliminating method towards medical power spectrum CT image.
Background technique
The key of Medical CT imaging is broadly divided into two o'clock: first is that the precision of images is promoted, second is that reducing dose of radiation.With monochromatic light
Sub-count type detector and energy product parting detector are that the multi-power spectrum CT of representative is capable of providing more accurately image information, are improved
The precision of medical imaging, and the much narrower of energy interval division will can be imaged closer to single, can preferably eliminate and penetrate
Beam hardening artifact provides more fully human body information, increases advantage to improve the accuracy of diagnosis.But it is reducing
Under conditions of overall radiation dosage, the number of photons that each energy section is covered is reduced, and causes the noise of image relatively low, is generated
Low dosage noise and artifact, it is difficult to meet the accuracy requirement of medical image.
In general, X-ray interacted with matter energy during include complicated energetic interaction mechanism, and charge-trapping
Generate and collect the event that process is randomization.Single-photon counting module is to the more demanding of count rate, energy product parting
Detector layer-to-layer signal transfer is serious, they are difficult to pass through tradition under the conditions of number of photons is insufficient for image bring noise and artifact
Method is eliminated.Recently as the development of machine learning techniques, convolutional neural networks show big advantage, by power spectrum
Each energy section reconstruction image of CT is handled by convolutional neural networks, in the actual clinical needs for meeting high-precision power spectrum CT
While further decrease the dose of radiation of X-ray.
Summary of the invention
In order to overcome the deficiencies of the prior art, the problem of being directed to low dosage noise present in power spectrum CT imaging and artifact,
The present invention is directed to propose a kind of method of the medical X-ray power spectrum CT image procossing based on convolutional neural networks, by primary low
The X-ray scanning of dosage, by preparatory trained convolutional neural networks model to the figure of each energy section Noise and artifact
As being handled, reduction becomes with the reconstruction image with phase homogenous quantities under the conditions of high dose x-ray bombardment, can reduce beam
It is influenced brought by hardening, random noise, exports more accurate medical image, reduce injury suffered by human body.For this purpose, this
Invention adopts the technical scheme that noise and artifact eliminating method towards medical power spectrum CT image, steps are as follows:
Step1: building Dummy mould imitates human body constituent and builds body mould, simulates bone, blood, fat, soft tissue
Institutional framework and including liver, lung, mammary gland vitals；
Step2: simulating power spectrum CT imaging by body mould with a large amount of x-ray photons, selects an individual mould, uses clinical criteria
The X-ray of dosage is irradiated, using software emulation detector physical process and carry out image reconstruction, obtain different-energy area
Between medical image, replace body mould, repeat the above process, obtain label image of the data set as neural network；
Step3: power spectrum CT imaging is simulated by body mould with relatively low-dose x-ray photon, with the X-ray weight of relatively low-dose
Detection and reconstruction process in multiple Step2, obtain input picture of the data set as neural network；
Step4: reconstruction image is matched.It will be one group with image setting obtained by Integral mold different-energy section, with
Just convolutional neural networks can learn to the information that all can be composed, and by various dose X-ray by same Integral mold the institute in Step2
It obtains label image to be matched with input picture obtained in Step3, makes its one-to-one correspondence, repeat the above process, until completing
The matching of reconstruction image under the conditions of all body moulds；
Step5: training convolutional neural networks.After defining a suitable convolutional neural networks structure, Step4 is randomly choosed
The N group input picture of middle arrangement is trained neural network with corresponding N group label image, by back-propagation algorithm until
Loss function converges to minimum, completes training；
Step6: test network training effect is instructed using Step5 is removed in data set matched in Step4 for neural network
Input picture is inputted network as test data set by the remainder outside experienced image, is examined using image quality evaluation standard
The effect of network training is looked into, if qualified, network training is completed, if unqualified, repeatedly Step5 and Step6.
A large amount of x-ray photons are Standard Ratio dosage used in clinically normal CT scan, and relatively low-dose is in Step2
The a quarter of used number of photons.
The features of the present invention and beneficial effect are:
The invention proposes a kind of noises and artifact eliminating method towards medical power spectrum CT image.Pass through convolutional Neural net
Network model handles the low signal-to-noise ratio (SNR) images in power spectrum CT different-energy section, improves while reducing X-ray radiation dosage
Picture quality provides advantage for accurate medical treatment.Trained neural network has faster processing speed simultaneously, compared to
The saving such as iterative reconstruction algorithm many times, improve the efficiency of image reconstruction.
Detailed description of the invention:
A kind of removing method frame diagram of noise and artifact towards medical X-ray energy spectrum CT image of Fig. 1.
The preparation method flow chart of Fig. 2 neural network model.
Specific embodiment
The problem of being directed to low dosage noise present in above-mentioned power spectrum CT imaging and artifact, the invention proposes a kind of bases
In the method for the medical X-ray power spectrum CT image procossing of convolutional neural networks, as shown in Figure 1.By the X-ray of a low dosage
Scanning is handled the image of each energy section Noise and artifact by preparatory trained convolutional neural networks model,
Reduction becomes with the reconstruction image with high dose x-ray bombardment phase homogenous quantities, can reduce beam hardening, random noise institute band
The influence come exports more accurate medical image, reduces injury suffered by human body.
The invention proposes a kind of noises and artifact eliminating method towards medical power spectrum CT image, pass through power spectrum CT first
The image in resulting different-energy section is rebuild under the conditions of system acquisition low dose X-ray, it is then logical as input data
Preparatory trained convolutional neural networks model is crossed, obtains having identical as resulting image is rebuild under high dose X-ray
The image for being suitable for the precisely different-energy section of medical treatment of quality.Convolutional neural networks model is by simulating low dosage X respectively
Obtained by the image training in the different-energy section reconstructed under ray and high dose X-ray as same Integral mold, this mistake
Journey frame diagram as shown in Figure 1, the specific preparation method flow chart of convolutional neural networks model as shown in Fig. 2, specific embodiment
It is as follows:
Step1: building Dummy mould.It imitates human body constituent and builds body mould, it should simulation bone, blood as far as possible
The vitals such as the institutional frameworks such as liquid, fat, soft tissue and liver, lung, mammary gland.
Step2: power spectrum CT imaging is simulated by body mould with a large amount of x-ray photons.The individual mould of selection one, uses clinical criteria
The X-ray of dosage is irradiated, using software emulation detector physical process and carry out image reconstruction, obtain different-energy area
Between medical image.Body mould is replaced, repeats the above process, obtains label image of the data set as neural network.
Step3: power spectrum CT imaging is simulated by body mould with a small amount of x-ray photon.(about with the X-ray of relatively low-dose
The a quarter of number of photons used in Step2) repeat Step2 in detection and reconstruction process, obtain data set as mind
Input picture through network.
Step4: reconstruction image is matched.It will be one group with image setting obtained by Integral mold different-energy section, with
Just convolutional neural networks can learn to the information that all can be composed, and by various dose X-ray by same Integral mold the institute in Step2
It obtains label image to be matched with input picture obtained in Step3, makes its one-to-one correspondence.It repeats the above process, until completing
The matching of reconstruction image under the conditions of all body moulds.
Step5: training convolutional neural networks.After defining a suitable convolutional neural networks structure, Step4 is randomly choosed
The N group input picture of middle arrangement is trained neural network with corresponding N group label image, by back-propagation algorithm until
Loss function converges to minimum, completes training.
Step6: test network training effect.It is instructed using Step5 is removed in data set matched in Step4 for neural network
Input picture is inputted network as test data set by the remainder outside experienced image, is examined using image quality evaluation standard
The effect of network training is looked into, if qualified, network training is completed, if unqualified, repeatedly Step5 and Step6.
High dose X-ray mentioned by the present invention is clinically Standard Ratio dosage used in normal CT scan, with sweeping
It retouches the difference of body positions and changes, low dose X-ray dosage is usually a quarter of standard dose.
The volume that can restore the image with noise and artifact that low dose X-ray reconstructs is generated in above-mentioned steps
Under conditions of product neural network model, so that it may for being clinically directed to the processing of low dosage power spectrum CT image.By power spectrum CT at
Image as system acquisition low dosage different-energy section with noise and artifact inputs in convolutional neural networks model, defeated
Out with the medical image of the reconstruction image equal quality in high dose different-energy section.
The present invention is further illustrated below by example, but does not therefore limit the invention to the example ranges
Interior, those of ordinary skill in the art conceive according to the present invention, and the simple change made, should be claimed in the present invention
In range.Below in conjunction with attached drawing, it is described as follows:
Used body mould should meet human body reality when present invention training neural network model, with human body vitals height
It is similar.The X-ray source of proposed adoption is generated by simulation GE Maxiray125, and wherein peak tube voltage and electric current are respectively set to
EMax=120keV, 0.5mAs and EMax=120keV, 2mAs simulates low dose X-ray and high dose X-ray, Qi Tacan respectively
Number is default setting.By layer-stepping energy product parting detector, it is grouped the charge number in each energy section of adding up, carries out energy
Reconstruction image after parsing forms the data set of neural network.
N group image in data set is randomly selected as training set, using low dosage different-energy section reconstruction image as
Input picture, high dose different-energy section reconstruction image are inputted in convolutional neural networks and are trained, directly as label image
Minimum is converged to loss function, completes the training of network.Non-training collection part is used as test set in data set, by input picture
Network is inputted, output image is evaluated with indexs such as PSNR (Y-PSNR), CNR (comparison noise ratio), if reaching institute
Required standard then completes the training of convolutional neural networks, if not up to standard, chooses N group image again and is trained, adjust
Whole network structure and its parameter, until output image meets the requirements.After the training for completing neural network model, it can be used for low dosage
The processing for the medical image with noise and artifact rebuild under X-ray.
Claims (2)
1. a kind of noise and artifact eliminating method towards medical power spectrum CT image, characterized in that steps are as follows:
Step1: building Dummy mould imitates human body constituent and builds body mould, the group of simulation bone, blood, fat, soft tissue
Knit structure and including liver, lung, mammary gland vitals；
Step2: simulating power spectrum CT imaging by body mould with a large amount of x-ray photons, an individual mould is selected, with clinical criteria dosage
X-ray be irradiated, using software emulation detector physical process and carry out image reconstruction, obtain different-energy section
Medical image replaces body mould, repeats the above process, obtain label image of the data set as neural network；
Step3: power spectrum CT imaging is simulated by body mould with relatively low-dose x-ray photon, is repeated with the X-ray of relatively low-dose
Detection and reconstruction process in Step2, obtain input picture of the data set as neural network；
Step4: reconstruction image is matched.It will be one group with image setting obtained by Integral mold different-energy section, to roll up
Product neural network can learn to the information that all can be composed, and by various dose X-ray as same Integral mold in Step2 gained mark
Label image is matched with input picture obtained in Step3, is made its one-to-one correspondence, is repeated the above process, all until completing
The matching of reconstruction image under the conditions of body mould；
Step5: training convolutional neural networks.After defining a suitable convolutional neural networks structure, randomly choose whole in Step4
The N group input picture of reason is trained neural network with corresponding N group label image, by back-propagation algorithm until loss
Function convergence completes training to minimum；
Step6: test network training effect is used for neural metwork training using Step5 is removed in data set matched in Step4
Input picture is inputted network as test data set by the remainder outside image, checks net using image quality evaluation standard
The effect of network training, if qualified, network training is completed, if unqualified, repeatedly Step5 and Step6.
2. noise and artifact eliminating method as described in claim 1 towards medical power spectrum CT image, characterized in that a large amount of X
Ray photons are Standard Ratio dosage used in clinically normal CT scan, and relatively low-dose is number of photons used in Step2
A quarter.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811198845.0A CN109272472B (en) | 2018-10-15 | 2018-10-15 | Noise and artifact eliminating method for medical energy spectrum CT image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811198845.0A CN109272472B (en) | 2018-10-15 | 2018-10-15 | Noise and artifact eliminating method for medical energy spectrum CT image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109272472A true CN109272472A (en) | 2019-01-25 |
| CN109272472B CN109272472B (en) | 2022-07-15 |
Family
ID=65196869
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811198845.0A Active CN109272472B (en) | 2018-10-15 | 2018-10-15 | Noise and artifact eliminating method for medical energy spectrum CT image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109272472B (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109978769A (en) * | 2019-04-04 | 2019-07-05 | 深圳安科高技术股份有限公司 | A kind of CT scan image data interpolation method and its system |
| CN110246199A (en) * | 2019-05-26 | 2019-09-17 | 天津大学 | A kind of projection domain data noise minimizing technology towards power spectrum CT |
| CN110517194A (en) * | 2019-07-19 | 2019-11-29 | 深圳安科高技术股份有限公司 | A kind of training method of substance decomposition model, substance decomposition method and relevant device |
| CN110728727A (en) * | 2019-09-03 | 2020-01-24 | 天津大学 | Low-dose energy spectrum CT projection data recovery method |
| CN110742635A (en) * | 2019-10-08 | 2020-02-04 | 南京安科医疗科技有限公司 | Composite energy spectrum CT imaging method |
| CN112581554A (en) * | 2019-09-30 | 2021-03-30 | 中国科学院深圳先进技术研究院 | CT imaging method, device, storage equipment and medical imaging system |
| CN113454489A (en) * | 2019-02-07 | 2021-09-28 | 浜松光子学株式会社 | Image processing apparatus, image processing method, and program |
| CN113689359A (en) * | 2021-09-23 | 2021-11-23 | 上海联影医疗科技股份有限公司 | Image artifact removing model and training method and system thereof |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN202049120U (en) * | 2011-03-04 | 2011-11-23 | 首都师范大学 | System for eliminating geometric artifacts in CT (computed tomography) image |
| CN102652674A (en) * | 2011-03-04 | 2012-09-05 | 首都师范大学 | Method and system for eliminating geometrical artifacts in CT (Computerized Tomography) image |
| US20140105477A1 (en) * | 2012-10-12 | 2014-04-17 | Mayo Foundation For Medical Education And Research | Low dose cardiac ct imaging with time-adaptive filtration |
| CN108564553A (en) * | 2018-05-07 | 2018-09-21 | 南方医科大学 | Low-dose CT image noise suppression method based on convolutional neural networks |
-
2018
- 2018-10-15 CN CN201811198845.0A patent/CN109272472B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN202049120U (en) * | 2011-03-04 | 2011-11-23 | 首都师范大学 | System for eliminating geometric artifacts in CT (computed tomography) image |
| CN102652674A (en) * | 2011-03-04 | 2012-09-05 | 首都师范大学 | Method and system for eliminating geometrical artifacts in CT (Computerized Tomography) image |
| US20140105477A1 (en) * | 2012-10-12 | 2014-04-17 | Mayo Foundation For Medical Education And Research | Low dose cardiac ct imaging with time-adaptive filtration |
| CN108564553A (en) * | 2018-05-07 | 2018-09-21 | 南方医科大学 | Low-dose CT image noise suppression method based on convolutional neural networks |
Non-Patent Citations (2)
| Title |
|---|
| 刘明娜: "基于视觉系统和特征提取的图像质量客观评价方法及应用研究", 《中国博士学位论文全文数据库.医药卫生科技》 * |
| 周丽平: "螺旋束CT系统重建及射束硬化伪影抑制方法", 《中国优秀硕士学位论文全文数据库.信息科技辑》 * |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113454489A (en) * | 2019-02-07 | 2021-09-28 | 浜松光子学株式会社 | Image processing apparatus, image processing method, and program |
| US11893660B2 (en) | 2019-02-07 | 2024-02-06 | Hamamatsu Photonics K.K. | Image processing device and image processing method |
| CN109978769B (en) * | 2019-04-04 | 2023-06-20 | 深圳安科高技术股份有限公司 | CT scanning image data interpolation method and system thereof |
| CN109978769A (en) * | 2019-04-04 | 2019-07-05 | 深圳安科高技术股份有限公司 | A kind of CT scan image data interpolation method and its system |
| CN110246199A (en) * | 2019-05-26 | 2019-09-17 | 天津大学 | A kind of projection domain data noise minimizing technology towards power spectrum CT |
| CN110246199B (en) * | 2019-05-26 | 2023-05-09 | 天津大学 | Projection domain data noise removing method for energy spectrum CT |
| CN110517194A (en) * | 2019-07-19 | 2019-11-29 | 深圳安科高技术股份有限公司 | A kind of training method of substance decomposition model, substance decomposition method and relevant device |
| CN110728727B (en) * | 2019-09-03 | 2023-04-18 | 天津大学 | Low-dose energy spectrum CT projection data recovery method |
| CN110728727A (en) * | 2019-09-03 | 2020-01-24 | 天津大学 | Low-dose energy spectrum CT projection data recovery method |
| WO2021062885A1 (en) * | 2019-09-30 | 2021-04-08 | 中国科学院深圳先进技术研究院 | Ct imaging method and apparatus, storage medium, and medical imaging system |
| CN112581554A (en) * | 2019-09-30 | 2021-03-30 | 中国科学院深圳先进技术研究院 | CT imaging method, device, storage equipment and medical imaging system |
| CN112581554B (en) * | 2019-09-30 | 2024-02-27 | 中国科学院深圳先进技术研究院 | CT imaging method, device, storage equipment and medical imaging system |
| CN110742635A (en) * | 2019-10-08 | 2020-02-04 | 南京安科医疗科技有限公司 | Composite energy spectrum CT imaging method |
| CN113689359A (en) * | 2021-09-23 | 2021-11-23 | 上海联影医疗科技股份有限公司 | Image artifact removing model and training method and system thereof |
| CN113689359B (en) * | 2021-09-23 | 2024-05-14 | 上海联影医疗科技股份有限公司 | Image artifact removal model and training method and system thereof |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109272472B (en) | 2022-07-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109272472A (en) | Noise and artifact eliminating method towards medical power spectrum CT image | |
| Spadea et al. | Deep learning based synthetic‐CT generation in radiotherapy and PET: a review | |
| CN105608721B (en) | Computer tomography artifact correction method and device | |
| CN107330949A (en) | A kind of artifact correction method and system | |
| CN110464353A (en) | A kind of pseudo- CT synthetic method and application based on depth convolutional neural networks | |
| CN103961125B (en) | A kind of CT value correcting method for Cone-Beam CT | |
| CN110675461A (en) | CT image recovery method based on unsupervised learning | |
| CN111429379B (en) | Low-dose CT image denoising method and system based on self-supervision learning | |
| CN106157320A (en) | A kind of image blood vessel segmentation method and device | |
| CN113470139B (en) | CT image reconstruction method based on MRI | |
| US20230121358A1 (en) | Ct image generation method for attenuation correction of pet images | |
| CN104463840A (en) | Fever to-be-checked computer aided diagnosis method based on PET/CT images | |
| CN105678821B (en) | A kind of dynamic PET images method for reconstructing based on self-encoding encoder image co-registration | |
| CN109191462A (en) | A kind of CT anthropomorphic phantom generation method | |
| Zhang et al. | PET image reconstruction using a cascading back-projection neural network | |
| CN105796121A (en) | CT and X-ray luminescence computed dual-mode synchronous tomography method | |
| CN109255354A (en) | medical CT-oriented computer image processing method and device | |
| CN110533734A (en) | Multi-power spectrum based on traditional single energy CT is segmented sparsely scanning iterative reconstruction approach | |
| CN116630738A (en) | Energy spectrum CT imaging method based on depth convolution sparse representation reconstruction network | |
| CN110270015B (en) | sCT generation method based on multi-sequence MRI | |
| Costa | Computer tomography phantom applications | |
| CN110246199A (en) | A kind of projection domain data noise minimizing technology towards power spectrum CT | |
| CN106062818A (en) | Method and apparatus for gate specific mr-based attenuation correction of timed-gated pet studies | |
| CN104000618A (en) | Breathing movement gating correction technology implemented with ring true photon number gating method | |
| Segars et al. | Investigation of respiratory gating in quantitative myocardial SPECT |
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 |