CN112116674A - Image reconstruction method, device, terminal and storage medium - Google Patents
Image reconstruction method, device, terminal and storage medium Download PDFInfo
- Publication number
- CN112116674A CN112116674A CN202010810587.8A CN202010810587A CN112116674A CN 112116674 A CN112116674 A CN 112116674A CN 202010810587 A CN202010810587 A CN 202010810587A CN 112116674 A CN112116674 A CN 112116674A
- Authority
- CN
- China
- Prior art keywords
- frequency domain
- image reconstruction
- image
- image data
- data
- 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
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (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)
- Magnetic Resonance Imaging Apparatus (AREA)
Abstract
The application relates to an image reconstruction method, an image reconstruction device, a terminal and a storage medium, wherein an image reconstruction model is constructed through the training of fully sampled frequency domain space image data, residual image data are output when input data are under-sampled frequency domain space image data, and a reconstructed magnetic resonance image is obtained by adding the input under-sampled frequency domain space image data and the output residual image data. The model for predicting the non-sampled frequency domain spatial data is simple, the training and predicting processes are accelerated, and meanwhile when the proportion of frequency domain spatial data acquisition is too low, the reconstructed image has a better effect.
Description
Technical Field
The embodiments of the present application relate to, but not limited to, the field of magnetic resonance imaging technologies, and in particular, to an image reconstruction method, system, terminal, and storage medium.
Background
Magnetic Resonance Imaging (MRI) is one of the important Imaging modalities in modern medical Imaging. Magnetic resonance imaging uses the phenomenon of magnetic resonance, uses radio frequency excitation to excite hydrogen atoms in a living body, and uses the electromagnetic characteristics thereof to generate clear and accurate images.
Although there are many advantages in magnetic resonance imaging, the data acquisition time of the magnetic resonance image is too long to limit the integrity of the sampled data, for example, by reducing the number of phase encoding to reduce the data sampling time, or acquiring partial echoes to reduce the echo time, can cause partial loss of the frequency domain spatial data. When the inverse fourier transform is used to directly reconstruct the frequency domain spatial data, the resulting image will be affected by the artifacts. Therefore, in the image reconstruction process, the frequency domain spatial data of the non-sampled portion needs to be predicted.
At present, projection on Projection (POCS) is usually used to predict the un-sampled frequency domain spatial data, but this method greatly reduces the image signal-to-noise ratio, and under the condition of severe image phase change, the prediction effect of the un-sampled frequency domain spatial data is not good; meanwhile, when the proportion of frequency domain spatial data acquisition is too low, the acquired data is not enough to predict phase information, and further the image reconstruction effect is influenced.
Disclosure of Invention
The following is a summary of the subject matter described in detail herein. This summary is not intended to limit the scope of the claims.
The embodiment of the application provides an image reconstruction method, an image reconstruction device, a terminal and a storage medium, wherein based on a deep learning model, an image reconstruction model is constructed through frequency domain space image data training of complete sampling, when input data is under-sampled frequency domain space image data, residual image data are output, and then the input under-sampled frequency domain space image data and the output residual image data are added to obtain a reconstructed magnetic resonance image. The model for predicting the non-sampled frequency domain spatial data by adopting the method is simple, the training and predicting processes are accelerated, and meanwhile, when the proportion of the frequency domain spatial data acquisition is too low, the image reconstruction has a better effect.
In a first aspect, an embodiment of the present application provides an image reconstruction method applied to magnetic resonance imaging, including: acquiring completely sampled frequency domain space image data as training data; training and obtaining an image reconstruction model using the training data; and reconstructing to obtain a magnetic resonance image by using the image reconstruction model, wherein the input of the image reconstruction model is under-sampled frequency domain space image data, the output of the image reconstruction model is residual image data, and the magnetic resonance image is the sum of the under-sampled frequency domain space image data and the residual image data.
In a second aspect, an embodiment of the present application provides an image reconstruction apparatus for performing the image reconstruction method according to the first aspect.
In a third aspect, an embodiment of the present application further provides a terminal, which at least includes: a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the image reconstruction method as in the first aspect when executing the computer program.
In a fourth aspect, embodiments of the present application further provide a computer-readable storage medium storing computer-executable instructions for performing the image reconstruction method of the first aspect as described above.
The method and the device for reconstructing the magnetic resonance image are based on a deep learning model, an image reconstruction model is constructed through training of completely sampled frequency domain space image data, residual image data are output when input data are under-sampled frequency domain space image data, and the input under-sampled frequency domain space image data and the output residual image data are added to obtain the reconstructed magnetic resonance image. The model for predicting the non-sampled frequency domain spatial data is simple, the training and predicting processes are accelerated, and meanwhile, when the proportion of frequency domain spatial data acquisition is too low, the image reconstruction has a better effect.
Additional features and advantages of the application will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by the practice of the application. The objectives and other advantages of the application may be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
Drawings
The accompanying drawings are included to provide a further understanding of the claimed subject matter and are incorporated in and constitute a part of this specification, illustrate embodiments of the subject matter and together with the description serve to explain the principles of the subject matter and not to limit the subject matter.
Fig. 1 is a schematic flowchart of an image reconstruction method according to an embodiment of the present disclosure;
fig. 2 is a schematic diagram of a magnetic resonance image obtained by an image reconstruction method according to an embodiment of the present application;
FIG. 3 is a diagram of a magnetic resonance image obtained by a convex set projection method according to an embodiment of the present application;
fig. 4 is a schematic diagram of a magnetic resonance image obtained by an image reconstruction method according to another embodiment of the present application;
FIG. 5 is a schematic diagram of a magnetic resonance image obtained by a convex set projection method according to another embodiment of the present application;
fig. 6 is a schematic diagram of a magnetic resonance image obtained by an image reconstruction method according to another embodiment of the present application;
fig. 7 is a schematic diagram of a magnetic resonance image obtained by a convex set projection method according to another embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
It should be noted that although functional blocks are partitioned in a schematic diagram of an apparatus and a logical order is shown in a flowchart, in some cases, the steps shown or described may be performed in a different order than the partitioning of blocks in the apparatus or the order in the flowchart. The terms first, second and the like in the description and in the claims, and the drawings described above, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order.
In the description of the embodiments of the present application, unless otherwise explicitly limited, terms such as setting, installing, connecting and the like should be understood in a broad sense, and those skilled in the art can reasonably determine the specific meanings of the terms in the embodiments of the present application by combining the specific contents of the technical solutions.
In magnetic resonance imaging, due to the physical characteristics of magnetic resonance and the limitation of the scanned human body, it often takes a long time to acquire enough frequency domain spatial data for image reconstruction. The long scanning time results in inefficient scanning diagnosis and is prone to patient discomfort during scanning. In addition, the long scan time makes it difficult for magnetic resonance scanning to deal with imaging of moving tissues of the human body, such as the abdomen, the heart, and the like. Thus, using multiple methods reduces the data acquisition time, which necessarily results in partial loss of the acquired data. For the proportion of data that has been collected, it is called the fourier fraction. Images resulting from direct reconstruction of the frequency domain spatial data with an inverse fourier transform will inevitably suffer from artefacts. Therefore, in the image reconstruction process, the non-acquired frequency domain spatial data needs to be predicted first.
Based on this, the embodiment of the application provides an image reconstruction method, an image reconstruction device, a terminal and a storage medium, which can construct an image reconstruction model through frequency domain space image data training of complete sampling based on a depth learning model, output residual image data when input data is under-sampled frequency domain space image data, and obtain a reconstructed magnetic resonance image by adding the input under-sampled frequency domain space image data and the output residual image data. The model for predicting the non-sampled frequency domain spatial data is simple, the training and predicting processes are accelerated, and meanwhile, when the proportion of frequency domain spatial data acquisition is too low, the image reconstruction has a better effect.
The embodiments of the present application will be further explained with reference to the drawings.
In a first aspect, an embodiment of the present application provides an image reconstruction method.
Fig. 1 is a schematic flowchart of an image reconstruction method according to an embodiment of the present application. As shown in fig. 1, the image reconstruction method provided in this embodiment at least includes:
step S100: and acquiring fully sampled frequency domain space image data as training data.
The fully sampled frequency domain spatial data may be derived not only from data obtained from a real magnetic resonance imaging scanner scan, but also from complete data obtained using a simulation method. And taking the completely sampled frequency domain space image data as training data for constructing an image reconstruction model. It should be noted that, no matter the frequency domain spatial data is fully sampled, or the frequency domain spatial data is undersampled, the parameters of the data sources may be different or even dissimilar, for example, the fully sampled frequency domain spatial data and the undersampled frequency domain spatial data may come from different objects or anatomical structures, and when data is acquired, different magnetic resonance imaging contrast, different magnetic resonance imaging scanner, and different magnetic resonance imaging magnetic field strength may not affect the image reconstruction method disclosed in this embodiment.
Step S200: an image reconstruction model is trained and obtained using the training data.
When training an image reconstruction model using training data, the reconstruction model may be one deep learning model or a combination of two or more deep learning models. Common deep learning models include convolutional neural network models, generative confrontation network models, cyclic neural network models, long-short term memory network models, automatic encoder network models, deep belief network models, deep residual error network models, gate cycle unit network models, echo state network models, and the like. The training process using a single model tends to be faster than the training process using a combined model, but the combined model tends to bring about a more accurate prediction effect.
Therefore, based on the existing deep learning model, an image reconstruction model with excellent performance can be obtained through the training of fully sampled frequency domain space image data.
Step S300: and reconstructing to obtain a magnetic resonance image by using the image reconstruction model.
The input of the image reconstruction model is under-sampled frequency domain space image data, the output of the image reconstruction model is residual image data, and the magnetic resonance image is the sum of the under-sampled frequency domain space image data and the residual image data. The method comprises the steps of obtaining under-sampled frequency domain spatial image data by performing under-sampling processing on fully sampled frequency domain spatial image data, inputting the under-sampled frequency domain spatial image data into an image reconstruction model, predicting and outputting residual image data through the image reconstruction model, performing addition operation on the under-sampled frequency domain spatial image data and the residual image data to obtain image data after the addition operation, converting the image data into an image domain through Fourier transform, then performing interpolation on the image domain, and reconstructing to obtain a magnetic resonance image.
In the embodiment, the undersampled frequency domain spatial image data is used as the input of the image reconstruction model, the residual image data is used as the output of the image reconstruction model, and the input undersampled frequency domain spatial image data and the output residual image data are added to obtain the reconstructed magnetic resonance image, so that the model for predicting the non-sampled frequency domain spatial data is simple, the training and predicting processes are accelerated, and meanwhile, when the ratio of acquiring the frequency domain spatial data is too low, the image reconstruction has a better effect. In addition, in this embodiment, the output of the image reconstruction model is residual map data, which can reduce the requirement of the neural network, so that the structure of the neural network becomes simple, and the effect of training the neural network is further improved.
The input of the image reconstruction model corresponds to two or more channels, and the output of the image reconstruction model corresponds to two or more channels. In an embodiment, the image reconstruction model is a convolutional neural network model, and the undersampled frequency domain spatial image data is fourier two-dimensional frequency domain spatial data, which are taken as an example to explain the structure of the convolutional neural network model and the image reconstruction method. Specifically, the input to the convolutional neural network model is complex image data, e.g., 256 × 256 × 2 in size, which is input through two channels, i.e., a real channel and an imaginary channel. The convolutional neural network model comprises five convolutional layers, each convolutional layer further comprises a batch normalization layer and an activation function, the input of the convolutional neural network model corresponds to the real part and the imaginary part of Fourier two-dimensional frequency domain space data respectively, the output of the convolutional neural network model corresponds to the real part and the imaginary part of residual image data respectively, the sizes of convolutional kernels of the five convolutional layers are 9 x 9, 7 x 7, 5 x 5 and 3 x 3 respectively, the number of convolutional kernels of each convolutional layer is 128, 64, 32 and 2 respectively, and residual image data output in prediction is a three-dimensional matrix of 256 x 2. Training the convolutional neural network through the completely sampled frequency domain space image data to obtain a convolutional neural network model, adding the undersampled frequency domain space image data and the residual image data, and reconstructing to obtain an image of magnetic resonance imaging. In this embodiment, the number of channels of the image reconstruction model and the dimensionality of the spatial data are not specifically limited, the number of channels may be two or more, and the spatial data may be three-dimensional spatial data, four-dimensional spatial data, or even higher dimensionality.
In another embodiment, the sampling may be performed in cartesian frequency domain space by a frequency encoding method, and the undersampled frequency domain spatial image data may be obtained by partial sampling in one or more directions, e.g., reconstructing a partial fourier along the frequency encoding direction (vertical direction). Referring to fig. 2, in this embodiment, an under-sampled data is obtained by sampling in a cartesian frequency domain space, the under-sampled data is input to a trained image reconstruction model, residual image data is output, and the input under-sampled image data and the residual image data are added to obtain a reconstructed image. Figure 3 is a magnetic resonance image obtained by a conventional convex set projection method. Fig. 2 and fig. 3 respectively include three cases of undersampling, the fourier scores of which are 0.51, 0.55 and 0.65, respectively, and when the fourier score is 0.51, the undersampling degree is the most serious, and the prediction difficulty of the non-acquired data is the greatest. As can be seen from comparing fig. 2 with fig. 3, compared with the convex set projection method in the prior art, when the fourier score is 0.65, the image reconstruction effects of the two methods are similar; when the fourier fraction is 0.55, the magnetic resonance image obtained by the image reconstruction method provided by the embodiment is obviously superior to the magnetic resonance image obtained by the convex set projection method in the prior art; when the fourier score is 0.51, the convex set projection method suffers from significant loss of high frequency information, while the image reconstruction method provided by the present embodiment retains clear edge information and has no significant noise.
In an embodiment, sampling may be performed in a cartesian frequency domain space through phase encoding, the under-sampled frequency domain space image data is obtained by partial sampling in one or more directions, referring to fig. 4 and 5, fig. 4 is a magnetic resonance image obtained by using the image reconstruction method disclosed in this embodiment, and fig. 5 is a magnetic resonance image obtained by using a conventional convex set projection method, it can be seen that when a fourier fraction is 0.60 or 0.55 along a phase encoding direction (horizontal direction), an effect of a reconstructed image obtained by using the image reconstruction method provided in this application is superior to that of the conventional convex set projection method, that is, the image reconstruction method can maintain a high signal-to-noise ratio while recovering high-frequency information.
In other embodiments, the sampling may be performed in polar frequency domain space by a projection coding method, and the under-sampled frequency domain spatial image data obtained by partial sampling in an angular direction may be also performed by a helical sampling method, and the under-sampled frequency domain spatial image data obtained by trajectory partial sampling of non-cartesian and non-polar frequency domain spaces.
In an embodiment, the image reconstruction model is obtained by training frequency domain spatial image data of different organs or parts, and the image reconstruction model is suitable for images of other organs with different contrasts. Referring to fig. 6, fig. 6 is a magnetic resonance image obtained by sampling in a cartesian frequency domain space through phase encoding (vertical direction) in the image reconstruction method of the present application, and fig. 7 is a magnetic resonance image obtained by a conventional convex set projection method, for example, a convolutional neural network is trained using image data of a knee joint image as training data to obtain a convolutional neural network model, and partial fourier reconstruction is performed on a brain T2 weighted FSE (Fast Spin Echo) and T1 weighted GE (Gradient Echo) image through the convolutional neural network model. Here, the partial fourier fraction of the T2 weighted image in the phase encoding direction (vertical direction) is 0.55, and the partial fourier fraction of the T1 weighted image in the frequency encoding direction (horizontal direction) is 0.55. Compared with the existing convex set projection method, the convex set projection method cannot adapt to the rapid local phase change of the image, so that the obtained magnetic resonance model can generate artifacts, as shown in fig. 7, a, b, c and d in the figure are the artifact parts in the magnetic resonance image obtained by the convex set projection method, and the magnetic resonance image obtained by the image reconstruction method of the application can not generate artifacts.
In a second aspect, an embodiment of the present application further provides an image reconstruction apparatus, configured to perform the image reconstruction method of the first aspect.
In a third aspect, an embodiment of the present application further provides a terminal, including: a memory, a processor and a computer program stored on the memory and executable on the processor, the processor performing the image reconstruction method of the first aspect.
In a fourth aspect, embodiments of the present application further provide a computer-readable storage medium storing computer-executable instructions for performing the image reconstruction method according to the first aspect.
The above-described embodiments of the apparatus are merely illustrative, wherein the units illustrated as separate components may or may not be physically separate, i.e. may be located in one place, or may also be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment.
It will be understood by those of ordinary skill in the art that all or some of the steps of the methods, systems, functional modules/units in the devices disclosed above may be implemented as software, firmware, hardware, and suitable combinations thereof. In a hardware implementation, the division between functional modules/units mentioned in the above description does not necessarily correspond to the division of physical components; for example, one physical component may have multiple functions, or one function or step may be performed by several physical components in cooperation. Some or all of the physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application specific integrated circuit. Such software may be distributed on computer readable media, which may include computer storage media (or non-transitory media) and communication media (or transitory media). The term computer storage media includes volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data, as is well known to those of ordinary skill in the art. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, Digital Versatile Disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to store the desired information and which can accessed by a computer. In addition, communication media typically embodies computer readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery media as known to those skilled in the art. The mobile terminal equipment can be a mobile phone, a tablet computer, a notebook computer, a palm computer, vehicle-mounted terminal equipment, wearable equipment, a super mobile personal computer, a netbook, a personal digital assistant, CPE, UFI (wireless hotspot equipment) and the like; the embodiments of the present application are not particularly limited.
While the preferred embodiments of the present invention have been described, the present invention is not limited to the above embodiments, and those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention, and such equivalent modifications or substitutions are included in the scope of the present invention defined by the claims.
Claims (10)
1. An image reconstruction method applied to magnetic resonance imaging comprises the following steps:
acquiring completely sampled frequency domain space image data as training data;
training and obtaining an image reconstruction model using the training data;
and reconstructing and obtaining a magnetic resonance image by using the image reconstruction model, wherein the input of the image reconstruction model is under-sampled frequency domain space image data, the output of the image reconstruction model is residual image data, and the magnetic resonance image is the sum of the under-sampled frequency domain space image data and the residual image data.
2. The image reconstruction method of claim 1, wherein the input of the image reconstruction model corresponds to two or more channels and the output of the image reconstruction model corresponds to two or more channels.
3. The image reconstruction method of claim 1, wherein the undersampled frequency domain spatial image data comprises:
the sampling is performed in a cartesian frequency domain space by a phase encoding or/and frequency encoding method, and the under-sampled frequency domain space image data is obtained by partial sampling in one or more directions.
4. The image reconstruction method of claim 1, wherein the undersampled frequency domain spatial image data comprises:
and sampling in a polar frequency domain space by a projection coding method, wherein the undersampled frequency domain space image data is obtained by partial sampling in an angle direction.
5. The image reconstruction method of claim 1, wherein the undersampled frequency domain spatial image data comprises:
sampling is carried out through a spiral sampling method, and the under-sampled frequency domain space image data is obtained through sampling of the track part of a non-Cartesian and non-polar frequency domain space.
6. The image reconstruction method according to any one of claims 1 to 5, wherein the image reconstruction model includes one or a combination of two or more of the following deep learning models:
a convolutional neural network model, a generative confrontation network model, a cyclic neural network model, a long-short term memory network model, an automatic encoder network model, a deep belief network model, a deep residual error network model, a gate cycle unit network model and an echo state network model.
7. The image reconstruction method of claim 6, wherein an input of the convolutional neural network model corresponds to real and imaginary parts of the undersampled frequency domain spatial image data, the convolutional neural network model includes five convolutional layers, and an output of the convolutional neural network model corresponds to real and imaginary parts of the residual image data.
8. An image reconstruction apparatus for performing the image reconstruction method of any one of claims 1 to 7.
9. A terminal, comprising: a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein: the processor, when executing the computer program, implements the image reconstruction method of any one of claims 1 to 7.
10. A computer-readable storage medium storing computer-executable instructions for performing the image reconstruction method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010810587.8A CN112116674A (en) | 2020-08-13 | 2020-08-13 | Image reconstruction method, device, terminal and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010810587.8A CN112116674A (en) | 2020-08-13 | 2020-08-13 | Image reconstruction method, device, terminal and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112116674A true CN112116674A (en) | 2020-12-22 |
Family
ID=73804906
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010810587.8A Pending CN112116674A (en) | 2020-08-13 | 2020-08-13 | Image reconstruction method, device, terminal and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112116674A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113240617A (en) * | 2021-07-09 | 2021-08-10 | 同方威视技术股份有限公司 | Scan image reconstruction method, inspection apparatus, and computer-readable storage medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109741256A (en) * | 2018-12-13 | 2019-05-10 | 西安电子科技大学 | Image super-resolution rebuilding method based on rarefaction representation and deep learning |
| CN109785279A (en) * | 2018-12-28 | 2019-05-21 | 江苏师范大学 | A kind of image co-registration method for reconstructing based on deep learning |
| CN110288672A (en) * | 2019-06-28 | 2019-09-27 | 闽江学院 | A kind of compressed sensing MR image reconstruction method based on the dense network of ultra-deep |
| CN111047660A (en) * | 2019-11-20 | 2020-04-21 | 深圳先进技术研究院 | Image reconstruction method, device, equipment and storage medium |
| CN111157935A (en) * | 2019-12-31 | 2020-05-15 | 上海联影智能医疗科技有限公司 | Magnetic resonance imaging method, magnetic resonance imaging device, storage medium and computer equipment |
| CN111383742A (en) * | 2018-12-27 | 2020-07-07 | 深圳先进技术研究院 | Method, device, equipment and storage medium for establishing medical imaging model |
| CN111383741A (en) * | 2018-12-27 | 2020-07-07 | 深圳先进技术研究院 | Method, device and equipment for establishing medical imaging model and storage medium |
-
2020
- 2020-08-13 CN CN202010810587.8A patent/CN112116674A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109741256A (en) * | 2018-12-13 | 2019-05-10 | 西安电子科技大学 | Image super-resolution rebuilding method based on rarefaction representation and deep learning |
| CN111383742A (en) * | 2018-12-27 | 2020-07-07 | 深圳先进技术研究院 | Method, device, equipment and storage medium for establishing medical imaging model |
| CN111383741A (en) * | 2018-12-27 | 2020-07-07 | 深圳先进技术研究院 | Method, device and equipment for establishing medical imaging model and storage medium |
| CN109785279A (en) * | 2018-12-28 | 2019-05-21 | 江苏师范大学 | A kind of image co-registration method for reconstructing based on deep learning |
| CN110288672A (en) * | 2019-06-28 | 2019-09-27 | 闽江学院 | A kind of compressed sensing MR image reconstruction method based on the dense network of ultra-deep |
| CN111047660A (en) * | 2019-11-20 | 2020-04-21 | 深圳先进技术研究院 | Image reconstruction method, device, equipment and storage medium |
| CN111157935A (en) * | 2019-12-31 | 2020-05-15 | 上海联影智能医疗科技有限公司 | Magnetic resonance imaging method, magnetic resonance imaging device, storage medium and computer equipment |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113240617A (en) * | 2021-07-09 | 2021-08-10 | 同方威视技术股份有限公司 | Scan image reconstruction method, inspection apparatus, and computer-readable storage medium |
| CN113240617B (en) * | 2021-07-09 | 2021-11-02 | 同方威视技术股份有限公司 | Scan image reconstruction method, inspection apparatus, and computer-readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhou et al. | DuDoRNet: learning a dual-domain recurrent network for fast MRI reconstruction with deep T1 prior | |
| Dar et al. | Prior-guided image reconstruction for accelerated multi-contrast MRI via generative adversarial networks | |
| CN109523584B (en) | Image processing method and device, multi-modality imaging system, storage medium and equipment | |
| Chen et al. | Deep learning for image enhancement and correction in magnetic resonance imaging—state-of-the-art and challenges | |
| US11250543B2 (en) | Medical imaging using neural networks | |
| CN115135237A (en) | Magnetic resonance image processing device and method thereof | |
| CN111047660B (en) | Image reconstruction method, device, equipment and storage medium | |
| CN112036506A (en) | Image recognition method and related device and equipment | |
| KR102584166B1 (en) | MAGNETIC RESONANCE IMAGE PROCESSING APPARATUS AND METHOD USING ARTIFICIAL NEURAL NETWORK AND RESCAlING | |
| CN111340903A (en) | Method and system for generating synthetic PET-CT image based on non-attenuation correction PET image | |
| CN111157935B (en) | Magnetic resonance imaging method, magnetic resonance imaging device, storage medium and computer equipment | |
| Jiang et al. | Respiratory motion correction in abdominal MRI using a densely connected U-Net with GAN-guided training | |
| Hosseinpour et al. | Temporal super resolution of ultrasound images using compressive sensing | |
| Cui et al. | Motion artifact reduction for magnetic resonance imaging with deep learning and k-space analysis | |
| CN113534031A (en) | Image domain data generating method, computer device and readable storage medium | |
| US11948288B2 (en) | Motion artifacts simulation | |
| CN112116674A (en) | Image reconstruction method, device, terminal and storage medium | |
| CN109633500B (en) | Transverse relaxation map determination method and device and magnetic resonance imaging equipment | |
| CN114167334A (en) | Magnetic resonance image reconstruction method and device and electronic equipment | |
| US11941732B2 (en) | Multi-slice MRI data processing using deep learning techniques | |
| US20160054420A1 (en) | Compensated magnetic resonance imaging system and method for improved magnetic resonance imaging and diffusion imaging | |
| CN115222628A (en) | Image processing method, device, equipment and storage medium | |
| CN114494014A (en) | Magnetic resonance image super-resolution reconstruction method and device | |
| KR102572311B1 (en) | Magnetic resonance image processing apparatus and method using artificial neural network and substitute map | |
| Goossens et al. | Objectively measuring signal detectability, contrast, blur and noise in medical images using channelized joint observers |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20220928 Address after: Room 4, Cyberport 100, Cyberport Road, Cyberport, Hongkong,, China Applicant after: VERSITECH Ltd. Address before: Pokfulam University of Hong Kong, Hong Kong, China Applicant before: THE University OF HONG KONG |
|
| TA01 | Transfer of patent application right |