CN111243052A - Image reconstruction method and device, computer equipment and storage medium - Google Patents
Image reconstruction method and device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN111243052A CN111243052A CN202010052344.2A CN202010052344A CN111243052A CN 111243052 A CN111243052 A CN 111243052A CN 202010052344 A CN202010052344 A CN 202010052344A CN 111243052 A CN111243052 A CN 111243052A
- Authority
- CN
- China
- Prior art keywords
- image
- medical image
- edge
- sample
- medical
- 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
- 238000000034 method Methods 0.000 title claims abstract description 73
- 238000004364 calculation method Methods 0.000 claims abstract description 46
- 230000006870 function Effects 0.000 claims description 64
- 238000012545 processing Methods 0.000 claims description 25
- 238000012549 training Methods 0.000 claims description 21
- 230000008569 process Effects 0.000 claims description 18
- 238000004590 computer program Methods 0.000 claims description 15
- 230000007246 mechanism Effects 0.000 claims description 14
- 230000009466 transformation Effects 0.000 claims description 10
- 238000002059 diagnostic imaging Methods 0.000 description 14
- 238000010586 diagram Methods 0.000 description 11
- 238000002595 magnetic resonance imaging Methods 0.000 description 10
- 230000000694 effects Effects 0.000 description 9
- 238000010606 normalization Methods 0.000 description 9
- 208000037170 Delayed Emergence from Anesthesia Diseases 0.000 description 5
- 238000012360 testing method Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 4
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 3
- 238000004891 communication Methods 0.000 description 3
- 238000002591 computed tomography Methods 0.000 description 3
- 239000012216 imaging agent Substances 0.000 description 3
- 239000003814 drug Substances 0.000 description 2
- 229940079593 drug Drugs 0.000 description 2
- PCHJSUWPFVWCPO-UHFFFAOYSA-N gold Chemical compound [Au] PCHJSUWPFVWCPO-UHFFFAOYSA-N 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 230000004060 metabolic process Effects 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 238000012952 Resampling Methods 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000013170 computed tomography imaging Methods 0.000 description 1
- 239000002872 contrast media Substances 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000008034 disappearance Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
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
- G06T11/005—Specific pre-processing for tomographic reconstruction, e.g. calibration, source positioning, rebinning, scatter correction, retrospective gating
Abstract
The application relates to an image reconstruction method, an image reconstruction device, a computer device and a storage medium. The method comprises the following steps: acquiring a first medical image; the scanning time length corresponding to the first medical image is less than the preset standard scanning time length; performing gradient calculation on the first medical image to obtain an edge image corresponding to the first medical image; the edge image is used for representing edge information of the first medical image; and according to the first medical image, the edge image and a preset image reconstruction model, obtaining a second medical image corresponding to the standard scanning time. The method improves the efficiency of the obtained second medical image corresponding to the standard scanning time.
Description
Technical Field
The present application relates to the field of image reconstruction technologies, and in particular, to an image reconstruction method, an image reconstruction apparatus, a computer device, and a storage medium.
Background
With the development of medical Imaging technology, there are Positron Emission Computed Tomography-electron Computed Tomography (PET-CT) and Positron Emission Computed Tomography/Magnetic Resonance Imaging (PET/MRI), PET-CT is a large functional metabolism and molecular Imaging diagnostic device composed of PET and CT, and has the examination functions of PET and CT, PET-MR is a large functional metabolism molecular Imaging diagnostic device composed of PET and MR, and has the examination functions of PET and MR.
In the conventional technology, when a PET-CT or PET-MR apparatus is used to examine a subject, a standard dose of PET drug is usually injected into the subject, and the subject is scanned for a long time, and data acquired during the long-time scanning process is used to reconstruct an image, so as to obtain a reconstructed PET image.
However, the conventional technique has a problem that it is inefficient to obtain a reconstructed PET image because the time for scanning the subject is long.
Disclosure of Invention
In view of the above, it is necessary to provide an image reconstruction method, an apparatus, a computer device and a storage medium for solving the problem of low efficiency of obtaining a reconstructed PET image in the conventional technology.
In a first aspect, an embodiment of the present invention provides an image reconstruction method, where the method includes:
acquiring a first medical image; the scanning time length corresponding to the first medical image is less than the preset standard scanning time length;
performing gradient calculation on the first medical image to obtain an edge image corresponding to the first medical image; the edge image is used for characterizing edge information of the first medical image;
and obtaining a second medical image corresponding to the standard scanning time according to the first medical image, the edge image and a preset image reconstruction model.
In a second aspect, an embodiment of the present invention provides an image reconstruction apparatus, including:
a first acquisition module for acquiring a first medical image; the scanning time length corresponding to the first medical image is less than the preset standard scanning time length;
the first calculation module is used for performing gradient calculation on the first medical image to obtain an edge image corresponding to the first medical image; the edge image is used for characterizing edge information of the first medical image;
and the reconstruction module is used for reconstructing a model according to the first medical image, the edge image and a preset image to obtain a second medical image corresponding to the standard scanning time.
In a third aspect, an embodiment of the present invention provides a computer device, including a memory and a processor, where the memory stores a computer program, and the processor implements the following steps when executing the computer program:
acquiring a first medical image; the scanning time length corresponding to the first medical image is less than the preset standard scanning time length;
performing gradient calculation on the first medical image to obtain an edge image corresponding to the first medical image; the edge image is used for characterizing edge information of the first medical image;
and obtaining a second medical image corresponding to the standard scanning time according to the first medical image, the edge image and a preset image reconstruction model.
In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the following steps:
acquiring a first medical image; the scanning time length corresponding to the first medical image is less than the preset standard scanning time length;
performing gradient calculation on the first medical image to obtain an edge image corresponding to the first medical image; the edge image is used for characterizing edge information of the first medical image;
and obtaining a second medical image corresponding to the standard scanning time according to the first medical image, the edge image and a preset image reconstruction model.
In the image reconstruction method, the image reconstruction device, the computer device and the storage medium provided by the above embodiments, the computer device acquires a first medical image; the scanning time length corresponding to the first medical image is less than the preset standard scanning time length; performing gradient calculation on the first medical image to obtain an edge image corresponding to the first medical image; the edge image is used for representing edge information of the first medical image; and according to the first medical image, the edge image and a preset image reconstruction model, obtaining a second medical image corresponding to the standard scanning time. In the method, the computer device performs gradient calculation on the acquired first medical image with the scanning duration less than the preset standard scanning duration to obtain an edge image representing edge information of the first medical image, and further reconstructs a second medical image with rich details corresponding to the standard scanning duration according to the first medical image with the scanning duration less than the preset standard scanning duration, the edge image and a preset image reconstruction model, so that the efficiency of the obtained second medical image is improved.
Drawings
FIG. 1 is a schematic diagram of an internal structure of a computer device according to an embodiment;
FIG. 1a is a flowchart illustrating obtaining a prediction graph of a standard scan duration according to an embodiment;
FIG. 2 is a flowchart illustrating an image reconstruction method according to an embodiment;
FIG. 2a is a diagram illustrating a standard scan duration image obtained by using an image reconstruction model according to an embodiment;
FIG. 2b is a diagram illustrating a standard scan duration image obtained using an image reconstruction model in an embodiment;
FIG. 2c is a diagram illustrating a standard scan duration image obtained using an image reconstruction model in an embodiment;
fig. 3 is a schematic flowchart of an image reconstruction method according to another embodiment;
fig. 4 is a schematic flowchart of an image reconstruction method according to another embodiment;
fig. 5 is a schematic flowchart of an image reconstruction method according to another embodiment;
FIG. 5a is a schematic network structure diagram of an initial image reconstruction model according to an embodiment;
fig. 6 is a schematic flowchart of an image reconstruction method according to another embodiment;
fig. 7 is a schematic structural diagram of an image reconstruction apparatus according to an embodiment.
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.
The image reconstruction method provided by the embodiment of the application can be applied to computer equipment shown in fig. 1. The computer device comprises a processor and a memory connected by a system bus, wherein a computer program is stored in the memory, and the steps of the method embodiments described below can be executed when the processor executes the computer program. Optionally, the computer device may further comprise a network interface, a display screen and an input device. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a nonvolatile storage medium storing an operating system and a computer program, and an internal memory. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used for communicating with an external terminal through a network connection. Optionally, the computer device may be a server, a personal computer, a personal digital assistant, other terminal devices such as a tablet computer, a mobile phone, and the like, or a cloud or a remote server, and the specific form of the computer device is not limited in the embodiment of the present application.
It should be noted that, in the image reconstruction method provided in the embodiments of the present application, the execution subject may be an image reconstruction apparatus, and the image reconstruction apparatus may be implemented as part of or all of a computer device by software, hardware, or a combination of software and hardware. In the following method embodiments, the execution subject is a computer device as an example.
It should be noted that there are two main factors affecting the quality of PET images obtained by a Positron Emission Computed Tomography-electron Computed Tomography scanner (PET-CT) or a Positron Emission Computed Tomography/Magnetic Resonance imager (PET/MRI): first, the injected dose of PET drug; second, the scan time of the PET system. Generally, the longer the scanning time of the PET system is, the more sufficient the acquired data is, and the quality of the reconstructed PET image can meet the clinical requirement; on the contrary, if the scanning time is short, the acquired data is insufficient, and the reconstructed PET image is difficult to be applied clinically. Currently, for the acquisition of PET data, it usually takes 3 minutes per bed with the normal dose injected; if half the normal dose is injected this means that at least 6 minutes of data need be acquired to reconstruct a clinically satisfactory PET image. However, the long-time scanning leads to the problem of low efficiency of acquiring the reconstructed image, so that the efficiency of obtaining the PET reconstructed image is improved, the quality of the generated PET reconstructed image is not influenced, and the method has great significance for PET-CT and PET/MRI scanning. Fig. 1a is a flowchart illustrating obtaining a prediction graph of a standard scan duration according to an embodiment. As shown in fig. 1a, a computer device normalizes a large number of short-time scanning sample images, inputs the normalized short-time scanning sample images into an initial image reconstruction model to obtain an output result, trains the initial image reconstruction model according to the output result of the initial image reconstruction model and the normalized standard scanning time sample images to obtain an image reconstruction model, and, when a prediction map of the standard scanning time needs to be reconstructed clinically, normalizes the clinically obtained short-time scanning images, inputs the normalized short-time scanning images into the image reconstruction model, and performs linear transformation on the output result to obtain a standard scanning time prediction map.
The following describes the technical solution of the present invention and how to solve the above technical problems with specific examples. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments.
Fig. 2 is a flowchart illustrating an image reconstruction method according to an embodiment. FIG. 2a is a diagram illustrating a standard scan duration image obtained by using an image reconstruction model according to an embodiment; FIG. 2b is a diagram illustrating a standard scan duration image obtained using an image reconstruction model in an embodiment; FIG. 2c is a diagram illustrating a standard scan duration image obtained using an image reconstruction model in an embodiment; the embodiment relates to a specific implementation process of performing gradient calculation on a first medical image with scanning duration less than preset standard scanning duration by computer equipment to obtain an edge image corresponding to the first medical image, and obtaining a second medical image corresponding to the standard scanning duration according to the first medical image, the edge image and a preset image reconstruction model. As shown in fig. 2, the method may include:
s201, acquiring a first medical image; the scanning duration corresponding to the first medical image is less than the preset standard scanning duration.
It is understood that the standard scan duration herein refers to a scan duration corresponding to the scan of the subject with different doses of contrast agent during the data acquisition process of the subject by the medical imaging device to generate an image satisfying the clinical requirement. Illustratively, taking a PET scan as an example, if the dose of the imaging agent injected to the subject is a normal dose, and the scan duration corresponding to the generation of the PET image satisfying the clinical requirement is 3 minutes, the corresponding standard scan duration is 3 minutes at this time; if the dose of the imaging agent injected to the examinee is half of the normal dose, and the scan time corresponding to the generation of the PET image satisfying the clinical requirement is 6 minutes, the corresponding standard scan time is 6 minutes at this time. Optionally, the computer device may acquire the first medical image from a PACS (Picture Archiving and Communication Systems) server, or may acquire the first medical image in real time from a medical imaging device, and optionally, the medical imaging device may be a PET-CT device, or a PET/MRI device, or a PET device, or a CT device. Optionally, the first medical image may be a PET image generated by a PET-CT or PET/MRI medical imaging device, a PET image generated by a single PET medical imaging device, or a CT image generated by a CT device.
S202, performing gradient calculation on the first medical image to obtain an edge image corresponding to the first medical image; the edge image is used to characterize edge information of the first medical image.
Specifically, the computer device performs gradient calculation on the first medical image to obtain an edge image corresponding to the first medical image. Wherein the edge image is used for characterizing edge information of the first medical image. Optionally, the computer device may perform gradient calculation on the first medical image by using a Sobel operator, or may perform gradient calculation on the first medical image by using a Prewitt operator or a Roberts cross edge detector (Roberts cross operator), so as to obtain an edge image corresponding to the first medical image.
And S203, according to the first medical image, the edge image and a preset image reconstruction model, obtaining a second medical image corresponding to the standard scanning time.
Specifically, the computer device obtains a second medical image corresponding to the standard scanning duration according to the first medical image, the edge image and a preset image reconstruction model. Optionally, the computer device may input the first medical image into a preset image reconstruction model to obtain a prediction graph corresponding to the first medical image, and obtain a second medical image corresponding to the standard scanning duration according to the prediction graph and the edge image corresponding to the first medical image; the first medical image and the edge image can also be simultaneously input into a preset image reconstruction model to obtain a second medical image corresponding to the standard scanning time. Illustratively, taking a PET scan as an example, if the dose of the imaging agent injected into the subject is a normal dose, the corresponding standard scan duration is 3 minutes, and if the scan duration corresponding to the acquired first medical image is 1.5 minutes, the scan duration corresponding to the acquired second medical image is 3 minutes. Optionally, the network structure of the preset image reconstruction model may be a full convolution network (fully convolutional Networks), such as a V-Net network, a U-Net network, and the like, or may be a generation countermeasure network (generic adaptive Networks), such as a pix2pix network, a WGAN network, and the like. Referring to fig. 2a, 2b, and 2c, as shown in fig. 2a, 2b, and 2c, the images of the short-time scan, the standard scan duration image obtained by the image reconstruction model of the present application, and the gold standard scan duration image are sequentially from left to right in the drawings, it can be seen that the standard scan duration image obtained by the present application is almost the same as the gold standard scan duration image, that is, the accuracy of the standard scan duration image obtained by the image reconstruction model of the present application is sufficient, and meanwhile, the image of the standard scan duration is reconstructed by the images of the short-time scan and the image reconstruction model, so that the efficiency of the obtained reconstructed image is improved, and the problem of low efficiency of obtaining the reconstructed image of the subject through the long-time scan is solved.
In this embodiment, the computer device performs gradient calculation on the acquired first medical image with the scanning duration less than the preset standard scanning duration to obtain an edge image representing edge information of the first medical image, and further reconstructs a second medical image with rich details corresponding to the standard scanning duration according to the first medical image with the scanning duration less than the preset standard scanning duration, the edge image and the preset image reconstruction model, thereby improving the efficiency of the obtained second medical image.
Fig. 3 is a schematic flowchart of an image reconstruction method according to another embodiment. The embodiment relates to a specific implementation process of obtaining a second medical image corresponding to a standard scanning time length by computer equipment according to a first medical image, an edge image and a preset image reconstruction model. As shown in fig. 3, on the basis of the foregoing embodiment, as an optional implementation manner, the foregoing S203 includes:
s301, obtaining a prediction graph and an edge prediction graph according to the first medical image, the edge image and the image reconstruction model; the prediction graph comprises less edge feature information than the edge prediction graph.
Specifically, the computer device obtains a prediction map and an edge prediction map according to the first medical image, the edge image and the image reconstruction model. And the edge characteristic information included in the prediction graph is smaller than that included in the edge prediction graph. It can be understood that, in the process of obtaining the prediction graph according to the first medical image and the reconstruction model, due to reasons such as incomplete extraction of the edge feature information of the first medical image, the obtained prediction graph has loss of the edge feature information, so that the edge feature information included in the prediction graph is smaller than the edge feature information included in the edge prediction graph. Optionally, the computer device may input the first medical image and the edge image into the image reconstruction model to obtain a prediction graph, and obtain an edge prediction graph corresponding to the edge image according to feature information generated in the process of obtaining the prediction graph.
And S302, obtaining a second medical image according to the prediction image and the edge prediction image.
Specifically, the computer device may obtain the second medical image according to the obtained prediction map and the edge prediction map. Optionally, the computer device may fuse the features of the edge prediction graph into the obtained prediction graph, and recover the detail information in the prediction graph according to the edge prediction graph to obtain the second medical image.
In the embodiment, the computer device can rapidly obtain the prediction image and the edge prediction image according to the first medical image, the edge image and the image reconstruction model, and further rapidly obtain the second medical image with rich detail information according to the prediction image and the edge prediction image, so that the efficiency of the obtained second medical image is improved; in addition, according to the first medical image, the edge image and the image reconstruction model, the prediction image and the edge prediction image can be accurately obtained, and then the second medical image rich in detail information can be accurately obtained according to the prediction image and the edge prediction image, so that the accuracy of the obtained second medical image is improved.
Fig. 4 is a schematic flowchart of an image reconstruction method according to another embodiment. The embodiment relates to a specific implementation process for obtaining a prediction graph and an edge prediction graph by computer equipment according to a first medical image, an edge image and a preset image reconstruction model. As shown in fig. 4, on the basis of the foregoing embodiment, as an optional implementation manner, the foregoing S301 includes:
s401, inputting the first medical image and the edge image into an image reconstruction model to obtain the characteristic information of the first medical image.
Specifically, the computer device inputs the first medical image and the edge image into the image reconstruction model to obtain the feature information of the first medical image. Optionally, the image reconstruction model may perform down-sampling operation and up-sampling operation on the input first medical image to obtain feature information of the first medical image. Optionally, before the computer device inputs the first medical image into the image reconstruction model, the computer device may compare the resolution of the first medical image with the resolution of a sample image used in training the image reconstruction model, and if the resolution of the first medical image is different from the resolution of the sample image, the computer device may perform resampling processing on the first medical image, resample the first medical image into an image in the same resolution space as the sample image, and then input the resampled first medical image into the image reconstruction model.
And S402, obtaining a prediction graph according to the characteristic information.
Specifically, the computer device inputs the first medical image and the edge image into the image reconstruction model, and obtains the feature information of the first medical image, and then the image reconstruction model obtains the prediction graph according to the obtained feature information of the first medical image. Optionally, the image reconstruction model may predict the feature information of the first medical image according to the obtained feature information of the first medical image to obtain predicted feature information, and then generate a prediction graph according to the obtained predicted feature information.
And S403, combining the characteristic information and the edge image by using an edge attention mechanism to obtain an edge prediction graph.
Specifically, after the computer device inputs the first medical image and the edge image into the image reconstruction model to obtain the feature information of the first medical image, the image reconstruction model combines the obtained feature information of the first medical image and the input edge image by using an edge attention mechanism to obtain an edge prediction map. The edge attention mechanism can exclude irrelevant information in the feature information of the first medical image under the guidance of the feature information of the first medical image, intensively learn and extract features relevant to edges, and obtain an edge prediction graph by combining the edge images.
In this embodiment, the computer device inputs the first medical image and the edge image into the image reconstruction model to obtain the feature information of the first medical image, and then the prediction graph can be accurately obtained according to the feature information of the first medical image.
In some scenarios, in order to facilitate convergence of the image reconstruction model, the first medical image and the edge image of the input image reconstruction model need to be normalized, since the edge image is obtained by performing gradient calculation on the first medical image, that is, in order to facilitate convergence of the image reconstruction model, the first medical image needs to be normalized. On the basis of the foregoing embodiment, as an optional implementation manner, before the foregoing S202, the method further includes: according to the mean value of the first medical image pixels and the standard deviation of the first medical image pixels, carrying out standardization processing on the first medical image to obtain a standardized first medical image; s202, comprising: and performing gradient calculation on the first medical image after the standardization processing to obtain an edge image corresponding to the first medical image.
Specifically, the computer device performs normalization processing on the first medical image according to the mean value of the pixels of the first medical image and the standard deviation of the pixels of the first medical image to obtain the normalized first medical image, and then performs gradient calculation on the normalized first medical image to obtain the edge image corresponding to the first medical image. Alternatively, the computer device may be based on a formula
Normalizing the first medical image to obtain a normalized first medical image, wherein I'stDenotes the first medical image, I' denotes the first medical image after normalization processing, μ denotes the mean of the first medical image pixels, and σ denotes the standard deviation of the first medical image pixels.
In this embodiment, the computer device performs normalization processing on the first medical image according to the mean value of the pixels of the first medical image and the standard deviation of the pixels of the first medical image to obtain the normalized first medical image, and then the obtained edge image corresponding to the first medical image is also the edge image after the normalization processing, so that the normalized first medical image and the normalized edge image are input into the image reconstruction model, convergence of the image reconstruction model can be facilitated, and the efficiency of obtaining the second medical image corresponding to the standard scanning duration is improved.
In some scenarios, if the first medical image is normalized, a linear transformation process is required to obtain the second medical image. On the basis of the foregoing embodiment, as an optional implementation manner, after S203, the method further includes: and performing linear transformation processing on the second medical image according to the mean value of the pixels of the first medical image and the standard deviation of the pixels of the first medical image to obtain a processed second medical image.
Specifically, the computer device performs linear transformation processing on the second medical image according to the mean value of the pixels of the first medical image and the standard deviation of the pixels of the first medical image to obtain a processed second medical image. Alternatively, the computer device may be according to formula I ═ IDLCarrying out linear transformation processing on the second medical image to obtain a processed second medical image, wherein IDLRepresenting the second medical image, I representing the processed second medical image, μ representing the mean of the pixels of the first medical image, and σ representing the standard deviation of the pixels of the first medical image.
In this embodiment, the computer device performs linear transformation processing on the second medical image according to the mean value of the pixels of the first medical image and the standard deviation of the pixels of the first medical image, so that the size and the number of channels of the obtained processed second medical image can be kept consistent with those of the first medical image, and the accuracy of the obtained processed second medical image is improved.
Fig. 5 is a schematic flowchart of an image reconstruction method according to another embodiment. Fig. 5a is a schematic network structure diagram of an initial image reconstruction model according to an embodiment. The embodiment relates to a specific implementation process of training an image reconstruction model by computer equipment. As shown in fig. 5, the training process of the image reconstruction model may include:
s501, performing gradient calculation on the first sample medical image to obtain a first sample edge image; the first sample edge image is used for representing edge information of the first sample medical image; the scanning time corresponding to the first sample medical image is less than the standard scanning time.
Specifically, the computer device performs gradient calculation on the first sample medical image to obtain a first sample edge image. The first sample edge image is used for representing edge information of the first sample medical image, and the scanning duration corresponding to the first sample medical image is smaller than the standard scanning duration. For example, in this embodiment, if the standard scanning duration is 3 minutes, the scanning duration corresponding to the first sample medical image may be 1.5 minutes, or may also be 1 minute, and without loss of generality, a medical image whose scanning duration is less than half of the standard scanning duration may be taken as the first sample medical image in this embodiment. Optionally, the computer device may acquire the first medical image from a PACS (Picture Archiving and Communication Systems) server, or may acquire the first medical image in real time from a medical imaging device, and optionally, the medical imaging device may be a PET-CT device, a PET/MRI device, or a CT device. Optionally, the first sample medical image may be a PET image generated by a PET-CT or PET/MRI medical imaging device, a PET image generated by a single PET medical imaging device, or a CT image generated by a CT device. Optionally, the computer device may perform gradient calculation on the first sample medical image by using a Sobel operator, or may perform gradient calculation on the first sample medical image by using a Prewitt operator or a Roberts cross edge detector (Roberts cross detector), so as to obtain an edge image corresponding to the first sample medical image.
S502, performing gradient calculation on the second sample medical image to obtain a second sample edge image; the second sample edge image is used for representing edge information of the second sample medical image; the scanning duration corresponding to the second sample medical image is the standard scanning duration.
Specifically, the computer device performs gradient calculation on the second sample medical image to obtain a second sample edge image. The second sample edge image is used for representing edge information of the second sample medical image, and the scanning duration corresponding to the second sample medical image is standard scanning duration. Optionally, the computer device may obtain the second sample medical image from a PACS (Picture Archiving and communication Systems) server, or may obtain the second sample medical image in real time from a medical imaging device, and optionally, the medical imaging device may be a PET-CT device, a PET/MRI device, or a CT device. Optionally, the second sample medical image may be a PET image generated by a PET-CT or PET/MRI medical imaging device, a PET image generated by a single PET medical imaging device, or a CT image generated by a CT device, and it should be noted that the second sample medical image is an image corresponding to the image type of the first sample medical image, for example, if the first sample medical image is a PET image whose scanning duration is less than the standard scanning duration, the second sample medical image is a PET image corresponding to the standard scanning duration; and if the first sample medical image is a CT image with the scanning duration less than the standard scanning duration, the second sample medical image is a CT image with the marking scanning duration corresponding to. Optionally, the computer device may perform gradient calculation on the second sample medical image by using a Sobel operator, or may perform gradient calculation on the second sample medical image by using a Prewitt operator or a Roberts cross edge detector (Roberts cross operator), so as to obtain an edge image corresponding to the second sample medical image.
S503, obtaining a sample edge prediction image and a reconstructed sample image according to the first sample medical image, the first edge image and a preset initial image reconstruction model; the reconstructed sample image is an image with standard scanning duration corresponding to the first sample image.
Specifically, the computer device reconstructs a model according to the first sample medical image, the first edge image and a preset initial image to obtain a sample edge prediction image and a reconstructed sample image. The reconstructed sample image is an image with standard scanning duration corresponding to the first sample image. Optionally, the computer device may input the first sample medical image and the first edge image into a preset initial image reconstruction model to obtain feature information of the first sample medical image, obtain a sample prediction map of the first sample medical image according to the feature information of the first sample medical image, obtain a sample edge prediction map by using an edge attention mechanism in combination with the feature information of the first sample medical image and the first edge image, and obtain a reconstructed sample image according to the obtained sample prediction map and the sample edge prediction map. As shown in fig. 5a, the main function of the edge attention mechanism is to eliminate irrelevant information with the help of the feature information of the first sample medical image, intensively learn and extract features related to the boundary, connect the obtained sample edge prediction map to the initial image reconstruction model, assist the initial image reconstruction model to recover the detail information of the first sample medical image, fuse the first sample medical image and the first edge image, sequentially pass through 3 Gate control mechanisms (Gate modules), and finally output the prediction of the first edge image through a layer of convolution operation, wherein the Gate control mechanism is a basic unit for extracting the most core of the boundary related region, and the Gate control mechanism performs bilinear (bilinear) upsampling on a network layer (providing high-level semantic feature information, in this embodiment, the second, third, and fourth downsampling output results) of the initial image reconstruction model, Convolution, connection (linkage) and sigmoid processing are carried out, then point multiplication of element-wise products is carried out on the feature graph after the sigmoid processing and the output (the network is shallow) of the edge attention mechanism residual block, the result of the point multiplication is used as the input of a next gating mechanism, and the edge attention mechanism is locally supervised by introducing the value of the loss function of the first edge image and the sample edge prediction image, so that the reconstruction of the boundary area is better completed. Optionally, in this embodiment, in order to generate a high-quality reconstructed sample image with rich details, a network structure of the preset initial image reconstruction model may be a deepedgeanswer network, where the network structure of the network is as shown in fig. 5a, a top layer of the deepedgeanswer network is a long residual connection (long residual connection) from input to output, so that the network can directly learn a residual between the reconstructed sample image and the first sample medical image, and accelerate convergence of the depth network, and from top to bottom, the network has a large dimension and a large resolution in a shallow image, and adopts dense connection to reduce information loss; the lower the dimension of the bottom image is, the lower the resolution is, the less the information is, and sparse connection is adopted; in addition, the DeepEdgeResUNet network largely adopts a residual block (residual block) composed of convolution, bn batch normalization and relu linear rectification functions, so that the problem of gradient disappearance is solved while the network performance is improved, and the convergence of the network is ensured. In order to avoid that the network parameters influence the generalization capability and the running time too much, a depth separable residual block (depthwise separable convolution) consisting of a bn batch normalization and a relu linear rectification function is particularly used in the deepedgeResunt network, and the use of the depth separable residual block greatly reduces the network parameter number and the running cost, so that the network can be time-consuming and lighter while having strong fitting capability.
In this embodiment, in order to further ensure the quantitative accuracy of the obtained reconstructed sample images, 2.5D processing may be performed on the first sample medical image, the first sample edge image, the second sample medical image, and the second sample edge image, where the 2.5D processing refers to performing, during training of the initial image reconstruction model, each training process in an image block instead of performing in the entire volume of the input image, for example, 5 first sample medical images in 2D are input into the initial image reconstruction model, so as to obtain 1 reconstructed sample image in 2D. Optionally, in this embodiment, the normalized first sample medical image and the processed second sample medical image may be respectively cropped randomly, where the size of the cropping may be any fixed size that meets the requirement of the initial image reconstruction model and the image dimension, but it is necessary to ensure that the positions of the image cropping are in one-to-one correspondence, for example: the cropped size may be [64,64,5], and thus corresponds to a 2.5D process, the input to the initial image reconstruction model may be represented as [64,64,5], the corresponding second sample medical image may be represented as [64,64,1], and the input to the edge attention mechanism may be represented as [64,64,10 ].
And S504, training the initial image reconstruction model according to the sample edge prediction image, the second sample edge image, the reconstructed sample image and the second sample medical image to obtain an image reconstruction model.
Specifically, the computer device trains the initial image reconstruction model according to the obtained sample edge prediction image, the second sample edge image, the reconstructed sample image and the second sample medical image, so as to obtain an image reconstruction model. Optionally, the computer device may obtain a loss function value of the initial image reconstruction model according to the sample edge prediction map, the second sample edge image, the reconstructed sample image, and the second sample medical image, train the initial image reconstruction model according to the loss function value of the initial image reconstruction model, and determine the corresponding initial image reconstruction model as the image reconstruction model when the loss function value of the initial image reconstruction model reaches a stable value. Optionally, when the initial image reconstruction model is trained, the optimizer can be selected as an Adam optimizer, and the Adam optimizer can rapidly converge and has good generalization capability. Optionally, the learning rate of the initial image reconstruction model training may adopt a learning rate Range Test (LR Range Test) technique to select an optimal learning rate, and the process may be described as follows: firstly, setting the learning rate to a small value, then simply iterating the model and the data for several times, increasing the learning rate after each iteration is finished, recording the training loss of each iteration, and then drawing an LRRange Test graph, wherein the general LRRange Test graph comprises three regions: if the first region learning rate is too small, the loss is basically unchanged, the second region loss is reduced and converges quickly, and the last region learning rate is too large, so that the loss begins to diverge, the optimal learning rate can be determined as the learning rate corresponding to the lowest point in the LR Range Test graph, and the optimal learning rate is used as the initial learning rate of the Adam optimizer.
In this embodiment, the computer device performs gradient calculation on the second sample image in the standard scanning duration, so as to obtain an accurate second sample edge image, and according to the first sample medical image, the first edge image and the preset initial image reconstruction model, a sample edge prediction image and a reconstructed sample image can be obtained, so that the initial image reconstruction model can be accurately trained according to the sample edge prediction image, the second sample edge image, the reconstructed sample image and the second sample medical image, and the accuracy of the obtained image reconstruction model is improved.
Fig. 6 is a schematic flowchart of an image reconstruction method according to another embodiment. The embodiment relates to a specific implementation process of training an initial image reconstruction model by computer equipment according to a sample edge prediction graph, a second sample edge image, a reconstructed sample image and a second sample medical image to obtain the image reconstruction model. As shown in fig. 6, on the basis of the foregoing embodiment, as an optional implementation manner, the foregoing S504 includes:
s601, obtaining a value of a first loss function of the initial image reconstruction model according to the sample edge prediction image and the second sample edge image.
Specifically, the computer device obtains a value of a first loss function of the initial image reconstruction model according to the sample edge prediction map and the second sample edge image. Optionally, the computer device may compare the obtained pixel value of the sample edge prediction image with the pixel value of the second sample edge image to obtain a value of the first loss function of the initial image reconstruction model.
S602, obtaining a value of a second loss function of the initial image reconstruction model according to the pixel value of the reconstructed sample image and the pixel value of the second sample medical image.
Specifically, the computer device obtains a value of a second loss function of the initial image reconstruction model according to the obtained pixel value of the reconstructed sample image and the pixel value of the second sample medical image. Alternatively, the computer device may reconstruct the sample image based on the pixel values of the second sample medical image, the pixel values of the reconstructed sample medical image, and the formula loss (x)i,yi)＝|xi-yiObtaining a second loss function of the initial image reconstruction modelValue of a number, in which xiPixel values, y, representing a reconstructed sample imageiPixel value, loss (x) representing the second sample medical imagei,yi) A value of a second loss function representing an initial image reconstruction model, wherein loss (x)i,yi) For calculating pixel-by-pixel intensity differences between the reconstructed sample image and the second sample medical image to ensure that the resulting quantitative error between the reconstructed sample image and the second sample image is as small as possible.
S603, obtaining a value of a region loss function of the initial image reconstruction model according to the mean value of each region of interest of the reconstructed sample image and the mean value of each region of interest of the second sample medical image.
Specifically, the computer device obtains a value of a region loss function of the initial image reconstruction model according to the mean value of each region of interest of the reconstructed sample image and the mean value of each region of interest of the second sample medical image. Optionally, the computer device may reconstruct the mean value of the regions of interest of the sample image, the mean value of the regions of interest of the second sample medical image, and the formula
The value of the regional loss function of the initial image reconstruction model is obtained, where,
represents the mean value of each Region of interest (ROI) of the reconstructed sample image,
the average value of each Region of interest (ROI) of the second sample medical image is represented, i represents the index of each Region of interest, mean represents the averaging, and loss (x, y) represents the value of the Region loss function of the initial image reconstruction model, wherein the loss (x, y) is used for calculating the relative error between each Region of interest of the reconstructed sample image and each Region of interest of the second sample medical image so as to ensure the quantitative accuracy of the obtained reconstructed sample image.
And S604, obtaining a value of a gradient loss function of the initial image reconstruction model according to the gradient amplitude of each pixel of the reconstructed sample image and the gradient amplitude of each pixel of the second sample medical image.
Specifically, the computer device obtains a value of a gradient loss function of the initial image reconstruction model according to the gradient amplitude of each pixel of the reconstructed sample image and the gradient amplitude of each pixel of the second sample medical image. Optionally, the computer device may reconstruct the gradient magnitude of each pixel of the sample image, the gradient magnitude of each pixel of the second sample medical image, and the formula loss (x)Gi,yGi)＝|G(xi)-G(yi) Obtaining the value of the gradient loss function of the initial image reconstruction model, wherein loss (x) isGi,yGi) Value of gradient loss function, G (x), representing initial image reconstruction modeli) Representing the magnitude of the gradient, G (y), of each pixel of the reconstructed sample imagei) Representing the gradient magnitude of each pixel of the second sample medical image, wherein loss (x)Gi,yGi) The method is used for calculating the gradient amplitude of each pixel of the reconstructed sample image and the gradient amplitude of each pixel of the second sample medical image so as to ensure that the obtained reconstructed sample image has rich detail information.
And S605, training the initial image reconstruction model according to the sum of the value of the first loss function, the value of the second loss function, the value of the regional loss function and the value of the gradient loss function to obtain the image reconstruction model.
Specifically, the computer device trains the initial image reconstruction model according to the sum of the value of the first loss function, the value of the second loss function, the value of the region loss function and the value of the gradient loss function of the initial image reconstruction model, and determines the corresponding initial image reconstruction model as the image reconstruction model when the sum of the value of the first loss function, the value of the second loss function, the value of the region loss function and the value of the gradient loss function of the initial image reconstruction model reaches a stable value. Optionally, the computer device may further train the initial image reconstruction model according to a weighted sum of a value of a first loss function, a value of a second loss function, a value of a region loss function, and a value of a gradient loss function of the initial image reconstruction model, so as to obtain the image reconstruction model.
In this embodiment, the value of the first loss function of the initial image reconstruction model obtained by the computer device can ensure that the quantitative error between the obtained sample edge prediction map and the second sample edge image is as small as possible, the value of the second loss function can ensure that the quantitative error between the obtained reconstructed sample image and the second sample image is as small as possible, the value of the regional loss function can ensure that the obtained reconstructed sample image is quantitative accurately, the value of the gradient loss function can ensure that the obtained reconstructed sample image has rich detail information, and the computer device can accurately train the initial image reconstruction model according to the sum of the obtained value of the first loss function, the obtained value of the second loss function, the obtained value of the regional loss function, and the obtained value of the gradient loss function, thereby improving the accuracy of the obtained image reconstruction model.
It should be understood that although the various steps in the flow charts of fig. 2-6 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least some of the steps in fig. 2-6 may include multiple steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, which are not necessarily performed in sequence, but may be performed in turn or alternately with other steps or at least some of the other steps.
Fig. 7 is a schematic structural diagram of an image reconstruction apparatus according to an embodiment. As shown in fig. 7, the apparatus may include: a first acquisition module 10, a first calculation module 11 and a reconstruction module 12.
Specifically, the first acquiring module 10 is configured to acquire a first medical image; the scanning time length corresponding to the first medical image is less than the preset standard scanning time length;
the first calculation module 11 is configured to perform gradient calculation on the first medical image to obtain an edge image corresponding to the first medical image; the edge image is used for representing edge information of the first medical image;
and the reconstruction module 12 is configured to reconstruct a model according to the first medical image, the edge image, and a preset image, and obtain a second medical image corresponding to the standard scanning duration.
The image reconstruction apparatus provided in this embodiment may implement the method embodiments described above, and the implementation principle and the technical effect are similar, which are not described herein again.
On the basis of the foregoing embodiment, optionally, the reconstruction module 12 includes: a first acquisition unit and a second acquisition unit.
Specifically, the first obtaining unit is configured to obtain a prediction graph and an edge prediction graph according to the first medical image, the edge image and the image reconstruction model; the prediction graph comprises less edge feature information than the edge prediction graph.
And the second acquisition unit is used for obtaining a second medical image according to the prediction image and the edge prediction image.
The image reconstruction apparatus provided in this embodiment may implement the method embodiments described above, and the implementation principle and the technical effect are similar, which are not described herein again.
On the basis of the foregoing embodiment, optionally, the first obtaining unit is specifically configured to input the first medical image and the edge image into an image reconstruction model to obtain feature information of the first medical image; obtaining a prediction graph according to the characteristic information; and obtaining an edge prediction graph by combining the characteristic information and the edge image by using an edge attention mechanism.
The image reconstruction apparatus provided in this embodiment may implement the method embodiments described above, and the implementation principle and the technical effect are similar, which are not described herein again.
On the basis of the foregoing embodiment, optionally, the apparatus further includes: and a standardization processing module.
Specifically, the normalization processing module is configured to perform normalization processing on the first medical image according to a mean value of pixels of the first medical image and a standard deviation of pixels of the first medical image, so as to obtain a normalized first medical image;
the first calculating module 11 is specifically configured to perform gradient calculation on the first medical image after the normalization processing to obtain an edge image corresponding to the first medical image.
The image reconstruction apparatus provided in this embodiment may implement the method embodiments described above, and the implementation principle and the technical effect are similar, which are not described herein again.
On the basis of the foregoing embodiment, optionally, the apparatus further includes: and a linear transformation processing module.
Specifically, the linear transformation processing module is configured to perform linear transformation processing on the second medical image according to the mean value of the pixels of the first medical image and the standard deviation of the pixels of the first medical image, so as to obtain a processed second medical image.
The image reconstruction apparatus provided in this embodiment may implement the method embodiments described above, and the implementation principle and the technical effect are similar, which are not described herein again.
On the basis of the foregoing embodiment, optionally, the apparatus further includes: the device comprises a second calculation module, a third calculation module, a second acquisition module and a training module.
Specifically, the second calculation module is configured to perform gradient calculation on the first sample medical image to obtain a first sample edge image; the first sample edge image is used for representing edge information of the first sample medical image; the scanning time corresponding to the first sample medical image is less than the standard scanning time;
the third calculation module is used for performing gradient calculation on the second sample medical image to obtain a second sample edge image; the second sample edge image is used for representing edge information of the second sample medical image; the scanning time corresponding to the second sample medical image is standard scanning time;
the second acquisition module is used for reconstructing a model according to the first sample medical image, the first edge image and a preset initial image to obtain a sample edge prediction image and a reconstructed sample image; reconstructing the sample image into an image with standard scanning duration corresponding to the first sample image;
and the training module is used for training the initial image reconstruction model according to the sample edge prediction image, the second sample edge image, the reconstructed sample image and the second sample medical image to obtain the image reconstruction model.
The image reconstruction apparatus provided in this embodiment may implement the method embodiments described above, and the implementation principle and the technical effect are similar, which are not described herein again.
On the basis of the foregoing embodiment, optionally, the training module includes: the device comprises a third acquisition unit, a fourth acquisition unit, a fifth acquisition unit, a sixth acquisition unit and a training unit.
Specifically, the third obtaining unit is configured to obtain a value of a first loss function of the initial image reconstruction model according to the sample edge prediction map and the second sample edge image;
the fourth obtaining unit is used for obtaining a value of a second loss function of the initial image reconstruction model according to the pixel value of the reconstructed sample image and the pixel value of the second sample medical image;
the fifth obtaining unit is used for obtaining the value of the area loss function of the initial image reconstruction model according to the mean value of each interested area of the reconstructed sample image and the mean value of each interested area of the second sample medical image;
a sixth obtaining unit, configured to obtain a value of a gradient loss function of the initial image reconstruction model according to the gradient amplitude of each pixel of the reconstructed sample image and the gradient amplitude of each pixel of the second sample medical image;
and the training unit is used for training the initial image reconstruction model according to the sum of the value of the first loss function, the value of the second loss function, the value of the regional loss function and the value of the gradient loss function to obtain the image reconstruction model.
The image reconstruction apparatus provided in this embodiment may implement the method embodiments described above, and the implementation principle and the technical effect are similar, which are not described herein again.
For specific limitations of the image reconstruction apparatus, reference may be made to the above limitations of the image reconstruction method, which are not described herein again. The modules in the image reconstruction device can be wholly or partially realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, a computer device is provided, comprising a memory and a processor, the memory having a computer program stored therein, the processor implementing the following steps when executing the computer program:
acquiring a first medical image; the scanning time length corresponding to the first medical image is less than the preset standard scanning time length;
performing gradient calculation on the first medical image to obtain an edge image corresponding to the first medical image; the edge image is used for representing edge information of the first medical image;
and according to the first medical image, the edge image and a preset image reconstruction model, obtaining a second medical image corresponding to the standard scanning time.
The implementation principle and technical effect of the computer device provided by the above embodiment are similar to those of the above method embodiment, and are not described herein again.
In one embodiment, a computer-readable storage medium is provided, having a computer program stored thereon, which when executed by a processor, performs the steps of:
acquiring a first medical image; the scanning time length corresponding to the first medical image is less than the preset standard scanning time length;
performing gradient calculation on the first medical image to obtain an edge image corresponding to the first medical image; the edge image is used for representing edge information of the first medical image;
and according to the first medical image, the edge image and a preset image reconstruction model, obtaining a second medical image corresponding to the standard scanning time.
The implementation principle and technical effect of the computer-readable storage medium provided by the above embodiments are similar to those of the above method embodiments, and are not described herein again.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database or other medium used in the embodiments provided herein can include at least one of non-volatile and volatile memory. Non-volatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical storage, or the like. Volatile Memory can include Random Access Memory (RAM) or external cache Memory. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), among others.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A method of image reconstruction, the method comprising:
acquiring a first medical image; the scanning time length corresponding to the first medical image is less than the preset standard scanning time length;
performing gradient calculation on the first medical image to obtain an edge image corresponding to the first medical image; the edge image is used for characterizing edge information of the first medical image;
and obtaining a second medical image corresponding to the standard scanning time according to the first medical image, the edge image and a preset image reconstruction model.
2. The method according to claim 1, wherein the obtaining a second medical image corresponding to the standard scan duration according to the first medical image, the edge image and a preset image reconstruction model comprises:
obtaining a prediction graph and an edge prediction graph according to the first medical image, the edge image and the image reconstruction model; wherein the prediction graph comprises less edge feature information than the edge prediction graph;
and obtaining the second medical image according to the prediction image and the edge prediction image.
3. The method of claim 2, wherein the reconstructing a model from the first medical image, the edge image, and the image to obtain a prediction map and an edge prediction map comprises:
inputting the first medical image and the edge image into the image reconstruction model to obtain the characteristic information of the first medical image;
obtaining the prediction graph according to the characteristic information;
and obtaining the edge prediction graph by combining the characteristic information and the edge image by utilizing an edge attention mechanism.
4. The method according to claim 1, wherein before performing the gradient calculation on the first medical image to obtain the edge image corresponding to the first medical image, the method further comprises:
according to the mean value of the first medical image pixels and the standard deviation of the first medical image pixels, carrying out standardization processing on the first medical image to obtain a standardized first medical image;
the gradient calculation of the first medical image to obtain an edge image corresponding to the first medical image includes:
and performing gradient calculation on the first medical image after the standardization processing to obtain an edge image corresponding to the first medical image.
5. The method according to claim 4, wherein after obtaining the second medical image corresponding to the standard scan duration according to the first medical image, the edge image and a preset image reconstruction model, the method further comprises:
and performing linear transformation processing on the second medical image according to the mean value of the first medical image pixels and the standard deviation of the first medical image pixels to obtain a processed second medical image.
6. The method of claim 1, wherein the training process of the image reconstruction model comprises:
performing gradient calculation on the first sample medical image to obtain a first sample edge image; the first sample edge image is used for representing edge information of the first sample medical image; the scanning time corresponding to the first sample medical image is less than the standard scanning time;
performing gradient calculation on the second sample medical image to obtain a second sample edge image; the second sample edge image is used for characterizing edge information of the second sample medical image; the scanning time corresponding to the second sample medical image is the standard scanning time;
obtaining a sample edge prediction image and a reconstructed sample image according to the first sample medical image, the first edge image and a preset initial image reconstruction model; the reconstructed sample image is an image with standard scanning duration corresponding to the first sample image;
and training the initial image reconstruction model according to the sample edge prediction image, the second sample edge image, the reconstructed sample image and the second sample medical image to obtain the image reconstruction model.
7. The method of claim 6, wherein the training the initial image reconstruction model according to the sample edge prediction map, the second sample edge image, the reconstructed sample image, and the second sample medical image to obtain the image reconstruction model comprises:
obtaining a value of a first loss function of the initial image reconstruction model according to the sample edge prediction graph and the second sample edge image;
obtaining a value of a second loss function of the initial image reconstruction model according to the pixel value of the reconstructed sample image and the pixel value of the second sample medical image;
obtaining a value of a region loss function of the initial image reconstruction model according to the mean value of each region of interest of the reconstructed sample image and the mean value of each region of interest of the second sample medical image;
obtaining a value of a gradient loss function of the initial image reconstruction model according to the gradient amplitude of each pixel of the reconstructed sample image and the gradient amplitude of each pixel of the second sample medical image;
and training the initial image reconstruction model according to the sum of the value of the first loss function, the value of the second loss function, the value of the regional loss function and the value of the gradient loss function to obtain the image reconstruction model.
8. An image reconstruction apparatus, characterized in that the apparatus comprises:
a first acquisition module for acquiring a first medical image; the scanning time length corresponding to the first medical image is less than the preset standard scanning time length;
the first calculation module is used for performing gradient calculation on the first medical image to obtain an edge image corresponding to the first medical image; the edge image is used for characterizing edge information of the first medical image;
and the reconstruction module is used for reconstructing a model according to the first medical image, the edge image and a preset image to obtain a second medical image corresponding to the standard scanning time.
9. A computer device comprising a memory and a processor, the memory storing a computer program, wherein the processor implements the steps of the method of any one of claims 1 to 7 when executing the computer program.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010052344.2A CN111243052A (en) | 2020-01-17 | 2020-01-17 | Image reconstruction method and device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010052344.2A CN111243052A (en) | 2020-01-17 | 2020-01-17 | Image reconstruction method and device, computer equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111243052A true CN111243052A (en) | 2020-06-05 |
Family
ID=70865644
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010052344.2A Pending CN111243052A (en) | 2020-01-17 | 2020-01-17 | Image reconstruction method and device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111243052A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113128583A (en) * | 2021-04-15 | 2021-07-16 | 重庆邮电大学 | Medical image fusion method and medium based on multi-scale mechanism and residual attention |
| WO2021248749A1 (en) * | 2020-06-10 | 2021-12-16 | 四川大学华西医院 | Diagnosis aid model for acute ischemic stroke, and image processing method |
| WO2022217984A1 (en) * | 2021-04-14 | 2022-10-20 | 浙江大学 | Low-dose double-tracer pet reconstruction method based on attention mechanism |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107610195A (en) * | 2017-07-28 | 2018-01-19 | 上海联影医疗科技有限公司 | The system and method for image conversion |
-
2020
- 2020-01-17 CN CN202010052344.2A patent/CN111243052A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107610195A (en) * | 2017-07-28 | 2018-01-19 | 上海联影医疗科技有限公司 | The system and method for image conversion |
| CN107633540A (en) * | 2017-07-28 | 2018-01-26 | 上海联影医疗科技有限公司 | The system and method for image conversion |
Non-Patent Citations (2)
| Title |
|---|
| WENHAN YANG 等: ""Deep Edge Guided Recurrent Residual Learning for Image Super-Resolution"", 《ARXIV》, 29 April 2016 (2016-04-29), pages 1 - 4 * |
| 宋玉龙: "基于反馈残差网络的矿井图像超分辨率重建算法研究", no. 02, pages 1 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021248749A1 (en) * | 2020-06-10 | 2021-12-16 | 四川大学华西医院 | Diagnosis aid model for acute ischemic stroke, and image processing method |
| WO2022217984A1 (en) * | 2021-04-14 | 2022-10-20 | 浙江大学 | Low-dose double-tracer pet reconstruction method based on attention mechanism |
| CN113128583A (en) * | 2021-04-15 | 2021-07-16 | 重庆邮电大学 | Medical image fusion method and medium based on multi-scale mechanism and residual attention |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109993726B (en) | Medical image detection method, device, equipment and storage medium | |
| CN109961491B (en) | Multi-mode image truncation compensation method, device, computer equipment and medium | |
| CN109697741B (en) | PET image reconstruction method, device, equipment and medium | |
| CN109697740B (en) | Image reconstruction method and device and computer equipment | |
| CN111161270B (en) | Vascular segmentation method for medical image, computer device and readable storage medium | |
| US11250543B2 (en) | Medical imaging using neural networks | |
| WO2020253886A1 (en) | Supplementary diagnosis method for pathology | |
| CN111179372B (en) | Image attenuation correction method, image attenuation correction device, computer equipment and storage medium | |
| CN110809782A (en) | Attenuation correction system and method | |
| CN111325714B (en) | Method for processing region of interest, computer device and readable storage medium | |
| US20230131722A1 (en) | Systems and methods for image registration | |
| CN111243052A (en) | Image reconstruction method and device, computer equipment and storage medium | |
| CN111161269B (en) | Image segmentation method, computer device, and readable storage medium | |
| CN111179231A (en) | Image processing method, device, equipment and storage medium | |
| JP2023540910A (en) | Connected Machine Learning Model with Collaborative Training for Lesion Detection | |
| CN111080584A (en) | Quality control method for medical image, computer device and readable storage medium | |
| CN110084868B (en) | Image correction method, apparatus, computer device, and readable storage medium | |
| CN110063742B (en) | Scattering correction method, scattering correction device, computer equipment and storage medium | |
| CN110415310A (en) | Medical scanning imaging method, device, storage medium and computer equipment | |
| CN111223158B (en) | Artifact correction method for heart coronary image and readable storage medium | |
| CN113159040A (en) | Method, device and system for generating medical image segmentation model | |
| CN111951272A (en) | Method and device for segmenting brain image, computer equipment and readable storage medium | |
| CN113989231A (en) | Method and device for determining kinetic parameters, computer equipment and storage medium | |
| US8805122B1 (en) | System, method, and computer-readable medium for interpolating spatially transformed volumetric medical image data | |
| CN111681297B (en) | Image reconstruction method, computer device, and storage medium |
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 |