CN103955899A - Dynamic PET image denoising method based on combined image guiding - Google Patents
Dynamic PET image denoising method based on combined image guiding Download PDFInfo
- Publication number
- CN103955899A CN103955899A CN201410180289.XA CN201410180289A CN103955899A CN 103955899 A CN103955899 A CN 103955899A CN 201410180289 A CN201410180289 A CN 201410180289A CN 103955899 A CN103955899 A CN 103955899A
- Authority
- CN
- China
- Prior art keywords
- image
- dynamic pet
- dynamic
- pet image
- guiding
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Abstract
The invention discloses a dynamic PET image denoising method based on combined image guiding. The method comprises the first step of utilizing a PET imaging device to carry out dynamic scanning and reestablishing a dynamic PET image, the second step of calculating and obtaining a combined image according to the reestablished PET image obtained in the first step, the third step of defining a guiding filter model and a kernel function of the guiding filter model, the fourth step of regarding the combined image obtained in the second step as a guiding image, and converting the model obtained in the third step to obtain an equation with a constraint target function for the single-frame dynamic PET image, and the fifth step of carrying out calculation to obtain the denoised dynamic PET image based on the linear regression method and overall parameter selection on the equation according to the result obtained in the fourth step. According to the method, the combined image serves as the guiding image, the noise of the single-frame dynamic PET image can be effectively reduced, and the quality of the dynamic PET image can be greatly improved.
Description
Technical field
The present invention relates to a kind of image de-noising method of medical image, relate in particular to a kind of dynamic PET image de-noising method based on combination image guiding.
Background technology
Positron emission computerized tomography (Positron Emission Tomography, PET) is one of state-of-the-art clinical examination technology of Medical Imaging, can to disease, diagnose at molecular level.Dynamic pet imaging, can not only show and the space distribution of radioactive tracer agent concentration also disclose the dynamic process of tracer agent metabolism in time.By applied dynamics model, can obtain and there is the kinetic parameter that enriches meaning, thereby provide guidance for medical diagnosis on disease, treatment and drug development.Yet, in dynamic pet imaging, in order to follow the trail of in time the tracer agent activity of histoorgan, change, the sweep time of the dynamic PET image of each frame is extremely short, the event that meets of corresponding each frame reduces, thereby make the single frames PET image that reconstructs affected by noise very large, further affect the estimation of kinetic parameter.So, in dynamic pet imaging urgently high-quality, efficiently wave filter is to dynamic PET image denoising.
The filtering method extensively adopting is at present gaussian filtering (Gaussian Filter, GF), the method is weighted on average entire image, and the value of each pixel is all obtained after the Euclidean distance weighted mean of location of pixels by itself and other pixel values in neighborhood.Easily see, gaussian filtering only considers that local neighborhood pixel Euclidean distance in position carries out filtering to pixel, and its filtering core is spatially constant, often causes the blurring effect of image.Huang in 2006 etc. propose HYPR method (HighlY constrained backPRojection, HYPR), utilize the high s/n ratio (Signal to Noise Ratio, SNR) of combination image to improve the signal to noise ratio (S/N ratio) of time-series image.2008, Vandehey etc. were applied to HYPR in dynamic PET image denoising, and emulation and real human body experimental data all show to use the SNR of the dynamic PET image after the denoising of HYPR method well to be improved.Yet HYPR denoising effect greatly depends on the single frames PET picture quality for the treatment of denoising.
In recent years, the employing Local Linear Models such as He have proposed a kind of navigational figure filtering method and have been widely applied in image denoising, and in the method, filtering core is the edge preserving smoothing operator being determined by navigational figure information.The relevant practice of the prior art still can not obtain good PET picture quality.
Not enough for prior art, provide a kind of dynamic PET image de-noising method based on combination image guiding very necessary to overcome prior art deficiency.
Summary of the invention
The invention discloses a kind of dynamic PET image de-noising method based on combination image guiding.The method is usingd combination image as navigational figure, can significantly improve dynamic PET picture quality.
Object of the present invention can realize by following technical measures, comprises that step is as follows:
A dynamic PET image de-noising method based on combination image guiding, comprises the following steps:
(1) utilize PET imaging device to carry out dynamic scan and rebuild dynamic PET image, described dynamic PET image each two field picture in sequence forms;
(2) the dynamic PET image obtaining according to step (1), calculates combination image;
(3) build guiding filter model;
(4) take the combination image that step (2) obtains is navigational figure, and the model of step (3) is transformed, and obtains the equation for the belt restraining objective function of the dynamic PET image of single frames;
(5) equation of the belt restraining objective function being obtained by step (4), on the basis based on global parameter in equation is selected, adopts linear regression method to calculate, and obtains the dynamic PET image after denoising.
Above-mentioned steps (2) the specifically summation that adds up of the pixel value by each frame PET image respective pixel in dynamic sequence obtains combination image.
In step described above (3), the guiding filter model of member is linear model, specifically supposes wave filter output
it is navigational figure
be centered close to pixel
window
linear transformation, guiding filter model be expressed as:
;
Wherein
that supposition is at window
in the linear coefficient that remains unchanged,
it is window
in any one pixel.
The equation of the belt restraining objective function for the dynamic PET image of single frames that above-mentioned steps (4) obtains is specially:
;
Wherein
that supposition is at window
in the linear coefficient that remains unchanged,
it is window
in any one pixel,
for navigational figure,
be wave filter input picture, be original noise image,
be an overall regularization parameter, overall regularization parameter is controlled
span.
Above-mentioned steps (5) specifically adopts linear regression method to minimize objective function, tries to achieve
with
:
;
;
Wherein
with
it is navigational figure
at window
average and variance,
it is window
the pixel number comprising,
it is noise image
be included in window
in the average of pixel;
The denoising image obtaining is:
;
Wherein
with
.
Dynamic PET image de-noising method based on combination image guiding of the present invention be take combination image as navigational figure, take full advantage of the feature that combination image in dynamic pet imaging has high s/n ratio dynamic PET image is carried out to filtering processing, can effectively reduce the dynamic PET picture noise of single frames, increase substantially the quality of dynamic PET image, thus auxiliary clinical diagnosis better.
Accompanying drawing explanation
The present invention is further illustrated to utilize accompanying drawing, but content in accompanying drawing does not form any limitation of the invention.
Fig. 1 is the process flow diagram that the present invention is based on the dynamic PET image de-noising method of combination image guiding.
Fig. 2 is the reconstruction noise image of Derenzo body mould and the denoising image of employing distinct methods processing of scanning.
Fig. 3 has described to use the sequence that in Fig. 2, distinct methods is processed to add with the average image at the outline of straight line figure that is 1.5mm and 2.5mm circular hole through diameter.
Fig. 3 illustrates to use the sequence of distinct methods processing in Fig. 2 to add with the average image and sequence and adds with the average image at the outline of straight line figure that is 1.5mm and 2.5mm circular hole through diameter.
Fig. 4 is the SD comparative result schematic diagram of HYPR and GIF method.
Fig. 5 is the NEMA NU4-2008 IQ physical bodies mould reconstruction noise image of scanning and the denoising image that adopts distinct methods to process.
Fig. 6 is for carrying out the result schematic diagram of qualitative assessment with NEMA NU4-2008 IQ physical bodies mould, the area-of-interest of encircled in Fig. 6 (a) for selecting in the circular hole that is 5mm at diameter to a two field picture, Fig. 6 (b) has provided the SD distribution plan of 20 two field pictures after distinct methods processing.
Embodiment
The invention will be further described with the following Examples.
embodiment 1.
A dynamic PET image de-noising method based on combination image guiding, as shown in Figure 1, comprises the following steps.
(1) utilize PET imaging device to carry out dynamic scan and rebuild dynamic PET image, dynamically each two field picture (1st frame to the of PET image in sequence
kframe) form.Experimental data acquisition mode of the present invention is full three-dimensional acquisition.
(2) the dynamic PET image obtaining according to step (1), calculates combination image.
Step (2) the specifically summation that adds up of the pixel value by each frame PET image respective pixel in dynamic sequence obtains combination image.
(3) build guiding filter model.
In step (3), the guiding filter model of member is linear model, specifically supposes wave filter output
it is navigational figure
be centered close to pixel
window
linear transformation, guiding filter model be expressed as:
;
Wherein
that supposition is at window
in the linear coefficient that remains unchanged,
it is window
in any one pixel.
(4) take the combination image that step (2) obtains is navigational figure, and the model of step (3) is transformed, and obtains the equation for the belt restraining objective function of the dynamic PET image of single frames.
The equation of the belt restraining objective function for the dynamic PET image of single frames that step (4) obtains is specially:
;
Wherein
that supposition is at window
in the linear coefficient that remains unchanged,
it is window
in any one pixel,
for navigational figure,
be wave filter input picture, be original noise image,
be an overall regularization parameter, overall regularization parameter is controlled
span.
(5) equation of the belt restraining objective function being obtained by step (4), on the basis based on global parameter in equation is selected, adopts linear regression method to calculate, and obtains the dynamic PET image after denoising.
Wherein, step (5) specifically adopts linear regression method to minimize objective function, tries to achieve
with
:
;
;
Wherein
with
it is navigational figure
at window
average and variance,
it is window
the pixel number comprising,
it is noise image
be included in window
in the average of pixel;
The denoising image obtaining is:
;
Wherein
with
.
Dynamic PET image de-noising method based on combination image guiding of the present invention be take combination image as navigational figure, take full advantage of the feature that combination image in dynamic pet imaging has high s/n ratio dynamic PET image is carried out to filtering processing, can effectively reduce the dynamic PET picture noise of single frames, increase substantially the quality of dynamic PET image, thus auxiliary clinical diagnosis better.
embodiment 2.
Adopt method of the present invention and method of the prior art to test, obtain relevant comparative result.
Fig. 2 has shown Derenzo body mould reconstruction noise image and the denoising image of scanning.Derenzo body mould adopts acrylic acid processing, and its body diameter is 40mm, and length is 13mm.Body mould exist a plurality of different-diameter sizes circular hole (diameter is followed successively by 0.8,, 1.0,1.25,1.5,2.0,2.5mm), and these circular holes be according to same diameter, assemble arranged together.In Derenzo body mould, injecting total activity is 18.54MBq's
18f-FDG solution, the enterprising line scanning of the PET of Bing Nanfang Hospital center toy InvoenmicroPET.PET system is set to default setting, and body mould is carried out to 20min dynamic scan, every frame period one minute, altogether 20 frames.Method for reconstructing is filtered back projection's method (filter backprojection, FBP), and the view data finally obtaining is
, pixel size is
.In figure, filtered back-projection method (filter backprojection is used in ' FBP ' representative, FBP) noise image of rebuilding, gaussian filtering (gaussian filter is used in ' GF ' representative, GF) method is to the filtered result images of FBP noise image, highly filtered back-projection method (the HighlY constrained backPRojection of convergence is used in ' HYPR ' representative, HYPR) to the filtered result images of FBP noise image, combination image guiding filtering (guilded image filter is used in ' GIF ' representative, GIF) method is to the filtered result images of FBP noise image.The 1st frame (Frame 1), the 10th frame (Frame 10), the 20th frame (Frame 20) image chosen wherein carry out result displaying, and image result shows, GIF method of the present invention has good denoising effect.
Fig. 3 illustrates to use the sequence of distinct methods processing in Fig. 2 to add with the average image and sequence and adds with the average image at the outline of straight line figure that is 1.5mm and 2.5mm circular hole through diameter.That profile diagram characterizes is error bar figure, be that each pixel of calculated line process is (because be that sequence adds the pixel with the average image, so each pixel is asked by 20 points corresponding to 20 two field pictures average) average and standard variance (standard deviation, SD)." work " font of curve represents the bound of SD, is average in the middle of I shape, when the I shape of curve is longer, shows that the noise of image is larger, and SD is larger.As seen from Figure 2, GF, HYPR have approximate consistent average (or gray-scale value) with the filtered image average of GIF and FBP noise image, but SD obviously reduces compared with FBP noise image, and wherein the filter effect of HYPR and GIF is good compared with GF.
Fig. 4 is the SD comparative result schematic diagram of HYPR and GIF method.In order further to compare the filter effect of HYPR and GIF, the comparison of in figure, the SD of HYPR and GIF being put together.The noise SD that obviously can find out GIF method of the present invention is less, illustrates that method of the present invention has better denoising effect.
embodiment 3.
Adopt method of the present invention and method of the prior art to carry out another experiment, obtain relevant comparative result.
Fig. 5 is the NEMA NU4-2008 IQ physical bodies mould reconstruction noise image of scanning and the denoising image that adopts distinct methods to process.This physical bodies mould is mainly for distinct methods is carried out to qualitative assessment, adopts organic glass processing, and its body diameter is 30mm, and length is 50mm, and comprising diameter is that 30mm, height are the cylindrical cavity of 30mm, also has the solid portion of high 20mm.Wherein solid portion have that 5 diameters are followed successively by 1,2,3,4,5mm can cavity filling rod, cavity Bang Yu cylindrical cavity UNICOM.In body mould, injecting 21ml activity concentration is 174.8kBq/ml's
18f-FDG solution, the total activity obtaining is 3.67MBq.In the enterprising line scanning of the PET of Nanfang Hospital center toy InvoenmicroPET.PET system is set to default setting, and body mould is carried out to 20min dynamic scan, every frame period one minute, altogether 20 frames.Method for reconstructing is filtered back projection's method (filter backprojection, FBP), and the view data finally obtaining is
, pixel size is
.The introduction of distinct methods is identical with explanation in embodiment 2.Choose the 1st frame (Frame 1), the 10th frame (Frame 10), the 20th frame (Frame 20) image and carry out result displaying, image result shows that GIF method of the present invention has good denoising effect.
Fig. 6 is for to carry out qualitative assessment result with NEMA NU4-2008 IQ physical bodies mould.The area-of-interest of encircled for selecting in the circular hole that is 5mm at diameter to a two field picture in Fig. 6 (a), its diameter is 9 pixels (3.6mm), the pixel in area-of-interest is used for calculating standard variance (standard deviation, SD).To be similar to Fig. 6 (a), to every two field picture (20 frames altogether), select area-of-interest to ask SD, Fig. 6 (b) has provided the SD distribution plan of 20 two field pictures after distinct methods processing, as can be seen from the figure, the SD of GF, HYPR, GIF is less than FBP noise image, wherein the SD of GIF is always minimum, and the GIF method that the present invention's proposition be described compares GF and HYPR method has better denoising effect.
The results show, dynamic PET image de-noising method based on combination image guiding of the present invention be take combination image as navigational figure, take full advantage of the feature that combination image in dynamic pet imaging has high s/n ratio dynamic PET image is carried out to filtering processing, can effectively reduce the dynamic PET picture noise of single frames, increase substantially the quality of dynamic PET image, thus auxiliary clinical diagnosis better.
Finally should be noted that; above embodiment is only in order to illustrate technical scheme of the present invention but not limiting the scope of the invention; although the present invention is explained in detail with reference to preferred embodiment; those of ordinary skill in the art is to be understood that; can modify or be equal to replacement technical scheme of the present invention, and not depart from essence and the scope of technical solution of the present invention.
Claims (5)
1. the dynamic PET image de-noising method based on combination image guiding, is characterized in that, comprises the following steps:
(1) utilize PET imaging device to carry out dynamic scan and rebuild dynamic PET image, described dynamic PET image each two field picture in sequence forms;
(2) the dynamic PET image obtaining according to step (1), calculates combination image;
(3) build guiding filter model;
(4) take the combination image that step (2) obtains is navigational figure, and the model of step (3) is transformed, and obtains the equation for the belt restraining objective function of the dynamic PET image of single frames;
(5) equation of the belt restraining objective function being obtained by step (4), on the basis based on global parameter in equation is selected, adopts linear regression method to calculate, and obtains the dynamic PET image after denoising.
2. the dynamic PET image de-noising method based on combination image guiding according to claim 1, is characterized in that: described step (2) the specifically pixel value by each frame PET image respective pixel in dynamic sequence add up to sue for peace and obtains combination image.
3. the dynamic PET image de-noising method based on combination image guiding according to claim 2, is characterized in that: in described step (3), the guiding filter model of member is linear model, specifically supposes wave filter output
it is navigational figure
be centered close to pixel
window
linear transformation, guiding filter model be expressed as:
;
Wherein
that supposition is at window
in the linear coefficient that remains unchanged,
it is window
in any one pixel.
4. the dynamic PET image de-noising method based on combination image guiding according to claim 3, is characterized in that: the equation of the belt restraining objective function for the dynamic PET image of single frames that described step (4) obtains is specially:
;
Wherein
that supposition is at window
in the linear coefficient that remains unchanged,
it is window
in any one pixel,
for navigational figure,
be wave filter input picture, be original noise image,
be an overall regularization parameter, overall regularization parameter is controlled
span.
5. the dynamic PET image de-noising method based on combination image guiding according to claim 1, is characterized in that:
Described step (5) specifically adopts linear regression method to minimize objective function, tries to achieve
with
:
;
;
Wherein
with
it is navigational figure
at window
average and variance,
it is window
the pixel number comprising,
it is noise image
be included in window
in the average of pixel;
The denoising image obtaining is:
;
Wherein
with
.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410180289.XA CN103955899A (en) | 2014-05-02 | 2014-05-02 | Dynamic PET image denoising method based on combined image guiding |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410180289.XA CN103955899A (en) | 2014-05-02 | 2014-05-02 | Dynamic PET image denoising method based on combined image guiding |
Publications (1)
Publication Number | Publication Date |
---|---|
CN103955899A true CN103955899A (en) | 2014-07-30 |
Family
ID=51333168
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410180289.XA Pending CN103955899A (en) | 2014-05-02 | 2014-05-02 | Dynamic PET image denoising method based on combined image guiding |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103955899A (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104318525A (en) * | 2014-10-17 | 2015-01-28 | 合肥工业大学 | Space guiding filtering based image detail enhancement method |
US10043295B2 (en) | 2015-10-13 | 2018-08-07 | Shenyang Neusoft Medical Systems Co., Ltd. | Reconstruction and combination of pet multi-bed image |
CN109035160A (en) * | 2018-06-29 | 2018-12-18 | 哈尔滨商业大学 | The fusion method of medical image and the image detecting method learnt based on fusion medical image |
JP2019113475A (en) * | 2017-12-26 | 2019-07-11 | 浜松ホトニクス株式会社 | Image processing device and image processing method |
CN110858391A (en) * | 2018-08-23 | 2020-03-03 | 通用电气公司 | Patient-specific deep learning image denoising method and system |
WO2020162296A1 (en) | 2019-02-07 | 2020-08-13 | 浜松ホトニクス株式会社 | Image processing device and image processing method |
CN113052933A (en) * | 2021-03-15 | 2021-06-29 | 深圳高性能医疗器械国家研究院有限公司 | Parameter imaging method and system |
WO2021153604A1 (en) | 2020-01-29 | 2021-08-05 | 浜松ホトニクス株式会社 | Image processing device and image processing method |
WO2022133639A1 (en) * | 2020-12-21 | 2022-06-30 | 深圳先进技术研究院 | Medical image processing method and apparatus, and device and storage medium |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101251583A (en) * | 2007-02-19 | 2008-08-27 | 威斯康星校友研究基金会 | Localized and highly constrained image reconstruction method |
CN103136731A (en) * | 2013-02-05 | 2013-06-05 | 南方医科大学 | Parameter imaging method of dynamic Positron Emission Tomography (PET) images |
-
2014
- 2014-05-02 CN CN201410180289.XA patent/CN103955899A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101251583A (en) * | 2007-02-19 | 2008-08-27 | 威斯康星校友研究基金会 | Localized and highly constrained image reconstruction method |
CN103136731A (en) * | 2013-02-05 | 2013-06-05 | 南方医科大学 | Parameter imaging method of dynamic Positron Emission Tomography (PET) images |
Non-Patent Citations (5)
Title |
---|
BRADLEY T. CHRISTIAN ET AL.: "Dynamic PET Denoising with HYPR Processing", 《JOURNAL OF NUCLEAR MEDICINE》 * |
JOYITA DUTTA ET AL.: "Non-Local Means Denoising of Dynamic PET Images", 《PLOS ONE》 * |
KAIMING HE ET AL.: "Guided image filtering", 《COMPUTER VISION–ECCV 2010》 * |
KAIMING HE ET AL.: "Guided Image Filtering", 《EEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE》 * |
边兆英: "基于区域时空先验的动态PET重建及PET图像恢复算法研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104318525B (en) * | 2014-10-17 | 2017-02-15 | 合肥工业大学 | Space guiding filtering based image detail enhancement method |
CN104318525A (en) * | 2014-10-17 | 2015-01-28 | 合肥工业大学 | Space guiding filtering based image detail enhancement method |
US10043295B2 (en) | 2015-10-13 | 2018-08-07 | Shenyang Neusoft Medical Systems Co., Ltd. | Reconstruction and combination of pet multi-bed image |
JP7018306B2 (en) | 2017-12-26 | 2022-02-10 | 浜松ホトニクス株式会社 | Image processing device and image processing method |
JP2019113475A (en) * | 2017-12-26 | 2019-07-11 | 浜松ホトニクス株式会社 | Image processing device and image processing method |
CN109035160A (en) * | 2018-06-29 | 2018-12-18 | 哈尔滨商业大学 | The fusion method of medical image and the image detecting method learnt based on fusion medical image |
CN109035160B (en) * | 2018-06-29 | 2022-06-21 | 哈尔滨商业大学 | Medical image fusion method and image detection method based on fusion medical image learning |
CN110858391A (en) * | 2018-08-23 | 2020-03-03 | 通用电气公司 | Patient-specific deep learning image denoising method and system |
CN110858391B (en) * | 2018-08-23 | 2023-10-10 | 通用电气公司 | Patient-specific deep learning image denoising method and system |
WO2020162296A1 (en) | 2019-02-07 | 2020-08-13 | 浜松ホトニクス株式会社 | Image processing device and image processing method |
US11893660B2 (en) | 2019-02-07 | 2024-02-06 | Hamamatsu Photonics K.K. | Image processing device and image processing method |
WO2021153604A1 (en) | 2020-01-29 | 2021-08-05 | 浜松ホトニクス株式会社 | Image processing device and image processing method |
WO2022133639A1 (en) * | 2020-12-21 | 2022-06-30 | 深圳先进技术研究院 | Medical image processing method and apparatus, and device and storage medium |
CN113052933A (en) * | 2021-03-15 | 2021-06-29 | 深圳高性能医疗器械国家研究院有限公司 | Parameter imaging method and system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103955899A (en) | Dynamic PET image denoising method based on combined image guiding | |
CN103020928B (en) | The metal artifacts reduction method of cone-beam CT system | |
CN109389655B (en) | Reconstruction of time-varying data | |
US20060061570A1 (en) | Method and system for progressive multi-resolution three-dimensional image reconstruction using region of interest information | |
EP3673457B1 (en) | A method of generating an enhanced tomographic image of an object | |
Wagner et al. | 4D interventional device reconstruction from biplane fluoroscopy | |
CN103136731B (en) | A kind of parameter imaging method of dynamic PET images | |
US10867375B2 (en) | Forecasting images for image processing | |
US20120275673A1 (en) | Truncation correction imaging enhancement method and system | |
US20230059132A1 (en) | System and method for deep learning for inverse problems without training data | |
US11481937B2 (en) | Positron emission tomography image reconstruction method | |
DE102014201134A1 (en) | Method and device for generating a 2-D projection image of a vascular system | |
DE102019201079A1 (en) | Method and imaging device for generating a motion-compensated image, computer program and storage medium | |
CN105931280A (en) | GPU-based fast three-dimensional CT (computed tomography) iterative reconstruction method | |
US10013778B2 (en) | Tomography apparatus and method of reconstructing tomography image by using the tomography apparatus | |
CN114494479A (en) | System and method for simultaneous attenuation correction, scatter correction, and denoising of low dose PET images using neural networks | |
Silva et al. | Image denoising methods for tumor discrimination in high-resolution computed tomography | |
CN103942763A (en) | Voxel level PET (positron emission tomography) image partial volume correction method based on MR (magnetic resonance) information guide | |
US20120275656A1 (en) | Method for supplying a 3d x-ray image data record for a moving object with highly absorbent material | |
CN111080736B (en) | Low-dose CT image reconstruction method based on sparse transformation | |
US20090304256A1 (en) | Method of extrapolating a grey-level bidimensional image in a radiography device and a radiography device implementing such a method | |
CN111179288A (en) | Interactive contrast blood vessel segmentation method and system | |
CN108765303B (en) | Integral enhancement method of digital subtraction blood vessel imaging image | |
KR102538329B1 (en) | A method for correcting a motion in a coronary image using a convolutional neural network | |
CN114359431A (en) | Method and system for directly reconstructing parameter image |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20140730 |
|
RJ01 | Rejection of invention patent application after publication |