CN111340906A - SPECT (single photon emission computed tomography) tomographic image reconstruction method, device and equipment combining ART (ART) and UNet algorithm - Google Patents
SPECT (single photon emission computed tomography) tomographic image reconstruction method, device and equipment combining ART (ART) and UNet algorithm Download PDFInfo
- Publication number
- CN111340906A CN111340906A CN202010115023.2A CN202010115023A CN111340906A CN 111340906 A CN111340906 A CN 111340906A CN 202010115023 A CN202010115023 A CN 202010115023A CN 111340906 A CN111340906 A CN 111340906A
- Authority
- CN
- China
- Prior art keywords
- unet
- art
- neural network
- spect
- network model
- 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
Images
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
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Nuclear Medicine (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The invention discloses a SPECT (single photon emission computed tomography) tomographic image reconstruction method, a device and equipment combining ART (ART) and UNet (UNet) algorithms, wherein the method comprises the following steps: acquiring projection data from a SPECT imaging system and attenuation coefficients from a CT system; calculating an initial solution of a reconstructed image according to the projection data and the attenuation coefficient by adopting an ART algorithm; and substituting the initial solution into a trained neural network model to output a reconstructed image, wherein the neural network model is an improved UNet neural network model, the output layer of the UNet neural network model is 1 layer, and the activation function is a sigmoid function. The method can make up the defect of the ART algorithm of the relaxation factors in the reconstruction of the SPECT tomography image to a certain extent, so that the method is suitable for the reconstruction of the SPECT image and improves the quality of the reconstructed image.
Description
Technical Field
The invention relates to the technical field of computers, in particular to a SPECT (single photon emission computed tomography) tomographic image reconstruction method, a SPECT tomographic image reconstruction device and SPECT tomographic image reconstruction equipment which combine an ART (ART) algorithm and a UNet algorithm.
Background
SPECT (Single-Photon Emission Computed Tomography) is a widely used medical detection technique, and SPECT image reconstruction aims to reconstruct an image of a target part from data collected by a medical instrument. The quality of the reconstructed image is crucial for medical diagnosis.
In the case that the values of the real images are all positive, the images reconstructed by the ART algorithm with the relaxation factor even show negative numbers, which makes the method less suitable for the reconstruction of the SPECT images.
Disclosure of Invention
The present invention is directed to solving at least one of the problems of the prior art. Therefore, the invention provides a SPECT (single photon emission computed tomography) tomographic image reconstruction method combining an ART algorithm and a UNet algorithm, which can make up the defect of the ART algorithm of a relaxation factor in the SPECT tomographic image reconstruction to a certain extent, so that the method is suitable for the SPECT image reconstruction and improves the quality of the reconstructed image.
The invention also provides a SPECT tomography image reconstruction device combining ART and UNet algorithms.
The invention also provides a SPECT tomography image reconstruction device combining ART and UNet algorithms.
The invention also provides a computer readable storage medium.
In a first aspect, an embodiment of the present invention provides a SPECT tomographic image reconstruction method combining ART and UNet algorithms: the method comprises the following steps:
acquiring projection data from a SPECT imaging system and attenuation coefficients from a CT system;
calculating an initial solution of a reconstructed image according to the projection data and the attenuation coefficient by adopting an ART algorithm;
and substituting the initial solution into a trained neural network model to output a reconstructed image, wherein the neural network model is a UNet neural network model, the output layer of the UNet neural network model is 1 layer, and the activation function is a sigmoid function.
Further, the input layer of the UNet neural network model is size 256 × 256 × 1.
Furthermore, the convolution layer of the UNet neural network model also comprises a Batch Norm layer.
Further, zero padding processing is carried out on the initial solution and the characteristic diagram.
Further, the UNet neural network model is obtained by the following training steps:
acquiring a data sample;
calculating an initial solution using an ART algorithm;
based on loss functions(theta is the parameter space of the neural network, G represents the output of the neural network, | | | | | sweet wind2A two-norm representative of a vector) and Adam algorithm to train the UNet neural network model.
Further, the data samples comprise artificially simulated data samples with gaussian noise.
In a second aspect, an embodiment of the present invention provides a SPECT tomographic image reconstruction apparatus combining ART and UNet algorithms, including:
the parameter acquisition unit is used for acquiring projection data from the SPECT imaging system and acquiring an attenuation coefficient from the CT system;
an initial solution calculation unit for calculating an initial solution of a reconstructed image according to the projection data and the attenuation coefficient by an ART algorithm;
and the reconstructed image output unit is used for substituting the initial solution into a trained neural network model to output a reconstructed image, the neural network model is a UNet neural network model, the output layer of the UNet neural network model is 1 layer, and the activation function is a sigmoid function.
In a third aspect, an embodiment of the present invention provides a SPECT tomographic image reconstruction apparatus combining ART and UNet algorithms, including:
at least one processor, and,
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the SPECT tomographic reconstruction method in combination with the ART and UNet algorithms.
In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having stored thereon computer-executable instructions for causing a computer to perform the SPECT tomographic reconstruction method in combination with the ART and UNet algorithms.
The SPECT tomography image reconstruction method, the device and the equipment which combine the ART algorithm and the UNet algorithm have at least the following beneficial effects: the initial solution of the reconstructed image is calculated through an ART algorithm, the initial solution is substituted into a trained and improved Unet neural network model to output the reconstructed image, the defects of the ART algorithm of a relaxation factor are overcome, the SPECT image reconstruction method is suitable for SPECT image reconstruction, the quality of the reconstructed image is improved, the interference of noise is weakened or even eliminated, the image with a clear interface is reconstructed, and meanwhile the image reconstruction time is shortened.
Drawings
FIG. 1 is a flowchart illustrating a SPECT tomographic image reconstruction method combining ART and UNet algorithms according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a UNet neural network model according to an embodiment of the SPECT tomographic image reconstruction method combining ART and UNet algorithms in accordance with the present invention;
FIG. 3 is a flowchart of UNet neural network model training in an embodiment of a SPECT tomographic reconstruction method incorporating ART and UNet algorithms in accordance with the present invention;
FIG. 4 is a lung CT image used in UNet neural network model training in an embodiment of the SPECT tomographic image reconstruction method incorporating ART and UNet algorithms in accordance with the present invention;
FIGS. 5 a-5 f are noisy projection views of a UNet neural network model training in an embodiment of a SPECT tomographic image reconstruction method incorporating ART and UNET algorithms in accordance with an embodiment of the present invention;
FIGS. 6a to 6f are initial solutions of reconstruction of training samples of a UNet neural network model according to an embodiment of a SPECT tomographic image reconstruction method combining ART and UNET algorithms according to an embodiment of the present invention;
FIGS. 7a to 7f are real solutions of reconstruction of training samples of a UNet neural network model according to an embodiment of the SPECT tomographic image reconstruction method combining ART and UNET algorithms according to the present invention;
FIG. 8a is a CT image of the lungs of an embodiment of a SPECT tomographic reconstruction method incorporating ART and UNet algorithms in accordance with an embodiment of the present invention;
FIG. 8b is a graph of the distribution of radioactive materials in the lungs in an embodiment of the SPECT tomographic reconstruction method incorporating ART and UNet algorithms in accordance with an embodiment of the present invention;
FIG. 8c is projection data of radioactive material in the lungs of an embodiment of a SPECT tomographic reconstruction method incorporating ART and UNet algorithms in accordance with an embodiment of the present invention;
FIGS. 9a to 9c are projection data obtained by adding 0%, 10% and 20% Gaussian noise to original projection data in an embodiment of a SPECT tomographic image reconstruction method combining ART and UNet algorithms according to an embodiment of the present invention;
fig. 10a to 10c are initial solutions of reconstruction images of SPECT tomographic images with the ART and UNet algorithms according to an embodiment of the present invention, in which 0%, 10%, and 20% of gaussian noise is added to the original projection data, respectively, in the SPECT tomographic image reconstruction method according to the embodiment of the present invention;
fig. 11a to 11c are reconstructed images of projection data obtained by adding 0%, 10%, and 20% gaussian noise to the original projection data in an embodiment of a SPECT tomographic image reconstruction method combining ART and UNet algorithms according to an embodiment of the present invention.
Detailed Description
The concept and technical effects of the present invention will be clearly and completely described below in conjunction with the embodiments to fully understand the objects, features and effects of the present invention. It is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all embodiments, and those skilled in the art can obtain other embodiments without inventive effort based on the embodiments of the present invention, and all embodiments are within the protection scope of the present invention.
In the description of the present invention, if an orientation description is referred to, for example, the orientations or positional relationships indicated by "upper", "lower", "front", "rear", "left", "right", etc. are based on the orientations or positional relationships shown in the drawings, only for convenience of describing the present invention and simplifying the description, but not for indicating or implying that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as limiting the present invention. If a feature is referred to as being "disposed," "secured," "connected," or "mounted" to another feature, it can be directly disposed, secured, or connected to the other feature or indirectly disposed, secured, connected, or mounted to the other feature.
In the description of the embodiments of the present invention, if "a number" is referred to, it means one or more, if "a plurality" is referred to, it means two or more, if "greater than", "less than" or "more than" is referred to, it is understood that the number is not included, and if "greater than", "lower" or "inner" is referred to, it is understood that the number is included. If reference is made to "first" or "second", this should be understood to distinguish between features and not to indicate or imply relative importance or to implicitly indicate the number of indicated features or to implicitly indicate the precedence of the indicated features.
Referring to fig. 1, a flow chart of a SPECT tomographic image reconstruction method combining ART and UNet algorithms in an embodiment of the present invention is shown. The method specifically comprises the following steps:
s1, acquiring projection data from a SPECT imaging system and acquiring an attenuation coefficient from a CT system;
s2, calculating an initial solution of a reconstructed image according to the projection data and the attenuation coefficient by adopting an ART algorithm;
and (3) obtaining an initial solution of a reconstructed image by inverting the projection data by adopting a classical ART algorithm, wherein an iterative expression of the ART algorithm is as follows:
wherein x is [ x ]i]N×1Is an image vector; p is a radical ofiIs the ith receiver sideThe obtained numerical value; a isiAn ith row element representing a projection matrix; "·" denotes the inner product; r represents a relaxation factor; k is the current iteration step number, and when k is 0, x(k)=0。
And S3, substituting the initial solution into a trained neural network model to output a reconstructed image, wherein the neural network model is a UNet neural network model, the output layer of the UNet neural network model is 1 layer, and the activation function is a sigmoid function.
According to the embodiment of the invention, the initial solution of the reconstructed image is calculated through the ART algorithm, the image reconstruction is carried out through the improved UNet neural network model, and the defects of the ART algorithm of the relaxation factors are made up, so that the method is suitable for the reconstruction of the SPECT image and the quality of the reconstructed image is improved.
In another embodiment, referring to fig. 2, the input layer size of the UNet neural network model is 256 × 256 × 1, and before the convolution operation, the embodiment of the present invention performs zero padding on the image to ensure that the feature map size is unchanged before and after the convolution operation, thereby avoiding the cropping operation in the original UNet.
In order to accelerate learning efficiency, inhibit overfitting and reduce sensitivity of a neural network to a weight initial value, a Batch Norm layer is added between convolution and nonlinear activation operation, namely, data is normalized after the convolution operation, then the normalized data is subjected to scaling and translation transformation, and finally, nonlinear function activation is used, so that the embodiment has high operation speed and high anti-noise interference performance.
Referring to fig. 3, the embodiment of the present invention further includes a UNet neural network model training step:
s01, acquiring a data sample;
sufficient data samples are needed for training the parameters of the improved UNet model, and in order to construct an image database, a lung CT image is firstly adopted as an attenuation image, as shown in FIG. 4. Artificially simulating a plurality of groups of data samples g with 5% -25% Gaussian noise(n)N is 1, …, N, i.e., N noisy projection views (some samples are shown in fig. 5a to 5 f), and N should be large enough to satisfy training requirements. Then, the ART algorithm is used for respectively inverting the N groups of data to obtainInitial solution to N reconstructed images(some samples are shown in FIGS. 6 a-6 f). The initial solution and the real solution of the N reconstructed images are obtained(partial samples are shown in figures 7 a-7 f) as input and label, respectively, for the improved UNet model. In actual practice, the value of N is suggested to be around 1000.
S02, calculating an initial solution by using an ART algorithm;
and S03, training the UNet neural network model based on a preset loss function and an Adam algorithm.
Wherein the loss function is(theta is the parameter space of the neural network, G represents the output of the neural network, | | | | | sweet wind2Representing the two-norm of the vector).
Inputting training samples in batches, optimizing a loss function by using an Adam algorithm, and updating model parameters. And setting a proper iteration number by taking the condition that over-fitting or under-fitting does not occur as a standard. And after the training is finished, saving the parameters of the neural network for the use of a subsequent image reconstruction task.
In a specific embodiment, the attenuation map is a pulmonary CT map; the distribution diagram of the radioactive substance is simulation data, the numerical value of the area with the distribution of the radioactive substance is 1, and the numerical values of the rest areas are 0; fig. 8a is an attenuation map, fig. 8b is a distribution map of a radioactive material, and fig. 8c is projection data.
0%, 10% and 20% of Gaussian noise is added to the original data, and the projection data are shown in FIGS. 9a to 9 c; reconstructing the projection data with 0%, 10% and 20% of Gaussian noise by ART algorithm to obtain the initial solution of the reconstructed image, wherein the reconstruction effect is shown in FIGS. 10 a-10 c; the initial solution is input into the trained neural network, the final solution of the image is output, and the reconstruction effect is shown in fig. 11a to 11 c.
It can be seen from the above embodiments that, in the embodiments of the present invention, under the influence of gaussian noise of different degrees on data, an image without noise and artifacts and with a clear interface can still be reconstructed by the method, and the time consumed for reconstructing the image is far shorter than that of the conventional iterative reconstruction algorithm.
The specific application of the embodiment of the invention is lung perfusion imaging, and the embodiment of the invention can also be transferred to other applications, such as myocardial perfusion imaging and the like, and has good transfer effect.
An embodiment of the invention also provides a SPECT tomography image reconstruction device combining ART and UNet algorithms, which comprises:
the parameter acquisition unit is used for acquiring projection data from the SPECT imaging system and acquiring an attenuation coefficient from the CT system;
an initial solution calculation unit for calculating an initial solution of a reconstructed image from the projection data by an ART algorithm;
and the reconstructed image output unit is used for substituting the initial solution into a trained neural network model to output a reconstructed image, the neural network model is a UNet neural network model, the output layer of the UNet neural network model is 1 layer, and the activation function is a sigmoid function.
An embodiment of the present invention also provides a SPECT tomographic image reconstruction apparatus combining ART and UNet algorithms, including:
at least one processor, and,
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the SPECT image reconstruction method.
An embodiment of the present invention also provides a computer-readable storage medium storing computer-executable instructions for causing a computer to perform the SPECT image reconstruction method.
The embodiments of the present invention have been described in detail with reference to the accompanying drawings, but the present invention is not limited to the above embodiments, and various changes can be made within the knowledge of those skilled in the art without departing from the gist of the present invention. Furthermore, the embodiments of the present invention and the features of the embodiments may be combined with each other without conflict.
Claims (9)
1. A SPECT tomography image reconstruction method combining ART and UNet algorithm is characterized by comprising the following steps:
acquiring projection data from a SPECT imaging system and attenuation coefficients from a CT system;
calculating an initial solution of a reconstructed image according to the projection data and the attenuation coefficient by adopting an ART algorithm;
and substituting the initial solution into a trained neural network model to output a reconstructed image, wherein the neural network model is a UNet neural network model, the output layer of the UNet neural network model is 1 layer, and the activation function is a sigmoid function.
2. The SPECT tomographic reconstruction method in combination with ART and UNet algorithm of claim 1, wherein the input slice size of the UNet neural network model is 256 × 256 × 1.
3. The SPECT tomographic image reconstruction method combining ART and UNet algorithm as claimed in claim 1, wherein the convolution layer of the UNet neural network model further includes a Batch Norm layer.
4. The SPECT tomographic reconstruction method in combination with ART and UNet algorithms of claim 1, further comprising zero-padding the initial solution and the feature map.
5. The SPECT tomographic image reconstruction method combining ART and UNet algorithm as claimed in claim 1, wherein the UNet neural network model is obtained by the following training steps:
acquiring a data sample;
calculating an initial solution using an ART algorithm;
6. The SPECT tomographic reconstruction method in combination with ART and UNet algorithms of claim 5, wherein the data samples comprise artificially simulated data samples with Gaussian noise.
7. A SPECT tomographic image reconstruction apparatus combining ART and UNet algorithms, comprising:
the parameter acquisition unit is used for acquiring projection data from the SPECT imaging system and acquiring an attenuation coefficient from the CT system;
an initial solution calculation unit for calculating an initial solution of a reconstructed image according to the projection data and the attenuation coefficient by an ART algorithm;
and the reconstructed image output unit is used for substituting the initial solution into a trained neural network model to output a reconstructed image, the neural network model is a UNet neural network model, the output layer of the UNet neural network model is 1 layer, and the activation function is a sigmoid function.
8. A SPECT tomographic reconstruction apparatus combining ART and UNet algorithms, comprising:
at least one processor, and,
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform a SPECT tomographic reconstruction method in combination with an ART and UNet algorithm as claimed in any one of claims 1 to 6.
9. A computer readable storage medium having stored thereon computer executable instructions for causing a computer to perform a SPECT tomographic reconstruction method in combination with an ART and UNet algorithm as claimed in any one of claims 1 to 6.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010115023.2A CN111340906A (en) | 2020-02-25 | 2020-02-25 | SPECT (single photon emission computed tomography) tomographic image reconstruction method, device and equipment combining ART (ART) and UNet algorithm |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010115023.2A CN111340906A (en) | 2020-02-25 | 2020-02-25 | SPECT (single photon emission computed tomography) tomographic image reconstruction method, device and equipment combining ART (ART) and UNet algorithm |
Publications (1)
Publication Number | Publication Date |
---|---|
CN111340906A true CN111340906A (en) | 2020-06-26 |
Family
ID=71183718
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010115023.2A Pending CN111340906A (en) | 2020-02-25 | 2020-02-25 | SPECT (single photon emission computed tomography) tomographic image reconstruction method, device and equipment combining ART (ART) and UNet algorithm |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111340906A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114723845A (en) * | 2022-06-10 | 2022-07-08 | 中国医学科学院北京协和医院 | Saline contrast lung perfusion image reconstruction method, system and equipment with SPECT as standard |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20060204070A1 (en) * | 2005-03-05 | 2006-09-14 | Hinshaw Waldo S | Three-dimensional x-ray imaging with Fourier reconstruction |
CN109493951A (en) * | 2017-11-08 | 2019-03-19 | 上海联影医疗科技有限公司 | For reducing the system and method for dose of radiation |
CN109727227A (en) * | 2018-11-23 | 2019-05-07 | 哈尔滨工业大学(威海) | A kind of diagnosis of thyroid illness method based on SPECT image |
-
2020
- 2020-02-25 CN CN202010115023.2A patent/CN111340906A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20060204070A1 (en) * | 2005-03-05 | 2006-09-14 | Hinshaw Waldo S | Three-dimensional x-ray imaging with Fourier reconstruction |
CN109493951A (en) * | 2017-11-08 | 2019-03-19 | 上海联影医疗科技有限公司 | For reducing the system and method for dose of radiation |
CN109727227A (en) * | 2018-11-23 | 2019-05-07 | 哈尔滨工业大学(威海) | A kind of diagnosis of thyroid illness method based on SPECT image |
Non-Patent Citations (3)
Title |
---|
TOMOHIRO SUZUKI 等: "IMAGE CORRECTION IN EMISSION TOMOGRAPHY USING DEEP CONVOLUTION NEURAL NETWORK", 《ICASSP》 * |
潘沛克 等: "基于U-net 模型的全自动鼻咽肿瘤MR图像分割", 《计算机应用》 * |
石明国: "《现代医学影像技术学》", 30 April 2007 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114723845A (en) * | 2022-06-10 | 2022-07-08 | 中国医学科学院北京协和医院 | Saline contrast lung perfusion image reconstruction method, system and equipment with SPECT as standard |
CN114723845B (en) * | 2022-06-10 | 2022-09-02 | 中国医学科学院北京协和医院 | Saline contrast lung perfusion image reconstruction method, system and equipment with SPECT as standard |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109978778B (en) | Convolutional neural network medical CT image denoising method based on residual learning | |
US10387765B2 (en) | Image correction using a deep generative machine-learning model | |
Dong et al. | A deep learning reconstruction framework for X-ray computed tomography with incomplete data | |
CN110660123B (en) | Three-dimensional CT image reconstruction method and device based on neural network and storage medium | |
CN110047113A (en) | Neural network training method and equipment, image processing method and equipment and storage medium | |
CN107610195A (en) | The system and method for image conversion | |
CN110930318B (en) | Low-dose CT image repairing and denoising method | |
CN109859285B (en) | Electrical impedance image reconstruction method based on cavity convolution network | |
Xu et al. | Accelerated fast iterative shrinkage thresholding algorithms for sparsity‐regularized cone‐beam CT image reconstruction | |
Ma et al. | Sinogram denoising via attention residual dense convolutional neural network for low-dose computed tomography | |
Xu et al. | Efficient low‐dose CT artifact mitigation using an artifact‐matched prior scan | |
CN111275786A (en) | SPECT (single photon emission computed tomography) tomographic image reconstruction method, device and equipment combining ML-EM (maximum likelihood-effective volume) and Unet algorithm | |
CN110599530B (en) | MVCT image texture enhancement method based on double regular constraints | |
Savanier et al. | Deep unfolding of the DBFB algorithm with application to ROI CT imaging with limited angular density | |
Chan et al. | An attention-based deep convolutional neural network for ultra-sparse-view CT reconstruction | |
CN111340906A (en) | SPECT (single photon emission computed tomography) tomographic image reconstruction method, device and equipment combining ART (ART) and UNet algorithm | |
CN116503506B (en) | Image reconstruction method, system, device and storage medium | |
Ye et al. | Momentum-net for low-dose CT image reconstruction | |
CN113706643B (en) | Head CT metal artifact correction method based on homomorphic adaptation learning | |
CN111080736B (en) | Low-dose CT image reconstruction method based on sparse transformation | |
CN113344876A (en) | Deformable registration method between CT and CBCT | |
Langet et al. | Compressed‐sensing‐based content‐driven hierarchical reconstruction: Theory and application to C‐arm cone‐beam tomography | |
CN116152373A (en) | Low-dose CT image reconstruction method combining neural network and convolutional dictionary learning | |
CN112132924B (en) | CT reconstruction method based on deep neural network | |
CN114926559A (en) | PET reconstruction method based on dictionary learning thought attenuation-free correction |
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 |