A dual-tracer model imaging method and system

Abstract

The present invention discloses an imaging method, system, device and medium of a dual tracer model, wherein the dual tracer model in the imaging method is obtained by the following steps: inputting a first tracer image set into a first mapping network to obtain a first mapping data set and a second mapping data set output by the first mapping network; inputting a second tracer image set into a second mapping network to obtain a third mapping data set and a fourth mapping data set output by the second mapping network; updating the parameters of the initialized dual tracer model according to the first tracer image set, the second tracer image set, the first mapping data set, the second mapping data set, the third mapping data set and the fourth mapping data set to obtain a trained dual tracer model. The method can effectively improve the safety of tracer imaging, as well as improve the imaging experience and convenience of the subject in performing tracer imaging. The present application relates to the technical field of positron emission tomography.

Description

Technical Field

[0001] The invention relates to the technical field of positron emission tomography, in particular to an imaging method and system of a double tracer model. Background Art

[0002] Positron emission computed tomography (PET) is an important clinical examination method in the field of nuclear medicine and is widely used in oncology, brain imaging, cardiology and other fields.

[0003] At present, the traditional dual-tracer imaging method usually uses two PET / CT scans with different tracers on the same subject at different times to obtain two different tracer images of the subject. This method causes the subject to receive a higher radiation dose in a shorter period of time and is not very safe. At the same time, this method also increases the time and economic costs of the subjects, and the imaging experience and convenience of tracer imaging are not satisfactory.

[0004] Therefore, the problems existing in the prior art still need to be solved and optimized. Summary of the invention

[0005] The purpose of the present invention is to solve one of the technical problems existing in the related art to at least a certain extent.

[0006] To this end, an object of an embodiment of the present invention is to provide an imaging method for a dual-tracer model, which can effectively improve the safety of tracer imaging, and improve the imaging experience and convenience of the subject in performing tracer imaging.

[0007] Another object of an embodiment of the present application is to provide an imaging system for a dual tracer model.

[0008] In order to achieve the above technical objectives, the technical solutions adopted by the embodiments of the present application include:

[0009] In a first aspect, an embodiment of the present application provides an imaging method of a dual tracer model, comprising:

[0010] obtaining source tracer images of the subject;

[0011] Inputting the source tracer image into a trained dual-tracer model to obtain a flow tracer image;

[0012] The dual-tracer model includes a first mapping network and a second mapping network, and the trained dual-tracer model is obtained by the following steps:

[0013] acquiring a first tracer data set and a second tracer data set of a subject;

[0014] Preprocessing the first tracer data set and the second tracer data set to obtain a dual-tracer image set, wherein the dual-tracer image set includes a first tracer image set and a second tracer image set;

[0015] Inputting the first tracer image set into the first mapping network, obtaining a first mapping data set and a second mapping data set output by the first mapping network, wherein the first mapping data set is a tracer image set corresponding to the second tracer image set and generated by mapping of the first mapping network; the second mapping data set is a tracer image set restored and generated by the first mapping network based on the first mapping data set, and the second mapping data set corresponds to the first tracer image set;

[0016] Inputting the second tracer image set into the second mapping network, obtaining a third mapping data set and a fourth mapping data set output by the second mapping network, wherein the third mapping data set is a tracer image set corresponding to the first tracer image set and generated by mapping of the second mapping network; the fourth mapping data set is a tracer image set restored and generated by the second mapping network based on the third mapping data set, and the fourth mapping data set corresponds to the second tracer image set;

[0017] The parameters of the initialized dual-tracer model are updated according to the first tracer image set, the second tracer image set, the first mapping data set, the second mapping data set, the third mapping data set and the fourth mapping data set to obtain the trained dual-tracer model.

[0018] In addition, the imaging method of the dual tracer model according to the above embodiment of the present application may also have the following additional technical features:

[0019] Further, in one embodiment of the present application, the preprocessing of the first tracer data set and the second tracer data set to obtain a dual tracer image set includes:

[0020] performing image registration on the second tracer data set according to the first tracer data set to obtain a registered first tracer data set and a registered second tracer data set;

[0021] Data enhancement is performed on the registered first tracer data set and the registered second tracer data set to obtain the dual tracer image set.

[0022] Further, in one embodiment of the present application, performing image registration on the second tracer data set according to the first tracer data set includes:

[0023] Obtaining a registration transformation matrix corresponding to the first tracer data set and the second tracer data set;

[0024] According to the registration transformation matrix, matrix registration is performed on the first tracer data set and the second tracer data set to obtain the registered first tracer data set and the registered second tracer data set.

[0025] Further, in one embodiment of the present application, the target mapping network is used to perform the following steps:

[0026] Performing mask generation processing on a starting image set corresponding to the target mapping network to obtain an attention mask image and a content mask image;

[0027] According to the starting image set, performing image fusion on the attention mask image and the content mask image to obtain an end image set;

[0028] The target mapping network is the first mapping network or the second mapping network, and the starting image set is any one of the first tracer image set, the second tracer image set, the first mapping data set, or the third mapping data set.

[0029] Further, in one embodiment of the present application, performing mask generation processing on a starting image set corresponding to the target mapping network to obtain an attention mask image and a content mask image includes:

[0030] Extracting encoding features from the starting image set to obtain a feature data set;

[0031] Performing residual convolution on the feature data set to obtain a residual data set;

[0032] The residual data set is decoded and sampled to obtain the attention mask image and the content mask image.

[0033] Further, in one embodiment of the present application, the target mapping network further includes a discriminator, and the discriminator is used to perform the following steps:

[0034] Performing convolution normalization on the attention mask image and the corresponding endpoint image set to obtain a normalized image set;

[0035] Convolution discrimination is performed on the normalized image set to obtain an attention discrimination result.

[0036] Further, in one embodiment of the present application, updating the parameters of the initialized dual tracer model according to the first tracer image set, the second tracer image set, the first mapping data set, the second mapping data set, the third mapping data set, and the fourth mapping data set to obtain the trained dual tracer model includes:

[0037] calculating a target loss function based on the first tracer image set, the second tracer image set, the first mapping data set, the second mapping data set, the third mapping data set, and the fourth mapping data set;

[0038] The initialized dual-tracer model is trained and updated according to the target loss function to obtain the trained dual-tracer model.

[0039] In a second aspect, an embodiment of the present application provides an imaging system of a dual tracer model, comprising:

[0040] an acquisition module, for acquiring a source tracer image of a subject;

[0041] a processing module, configured to input the source tracer image into a trained dual tracer model to obtain a flow tracer image;

[0042] The dual-tracer model includes a first mapping network and a second mapping network, and the trained dual-tracer model is obtained by the following steps:

[0043] acquiring a first tracer data set and a second tracer data set of a subject;

[0044] Preprocessing the first tracer data set and the second tracer data set to obtain a dual-tracer image set, wherein the dual-tracer image set includes a first tracer image set and a second tracer image set;

[0045] Inputting the first tracer image set into the first mapping network, obtaining a first mapping data set and a second mapping data set output by the first mapping network, wherein the first mapping data set is a tracer image set corresponding to the second tracer image set and generated by mapping of the first mapping network; the second mapping data set is a tracer image set restored and generated by the first mapping network based on the first mapping data set, and the second mapping data set corresponds to the first tracer image set;

[0046] Inputting the second tracer image set into the second mapping network, obtaining a third mapping data set and a fourth mapping data set output by the second mapping network, wherein the third mapping data set is a tracer image set corresponding to the first tracer image set and generated by mapping of the second mapping network; the fourth mapping data set is a tracer image set restored and generated by the second mapping network based on the third mapping data set, and the fourth mapping data set corresponds to the second tracer image set;

[0047] The parameters of the initialized dual-tracer model are updated according to the first tracer image set, the second tracer image set, the first mapping data set, the second mapping data set, the third mapping data set and the fourth mapping data set to obtain the trained dual-tracer model.

[0048] Furthermore, in one embodiment of the present application, the imaging system further includes a target mapping network, and the target mapping network is used to perform the following steps:

[0049] Performing mask generation processing on a starting image set corresponding to the target mapping network to obtain an attention mask image and a content mask image;

[0050] According to the starting image set, performing image fusion on the attention mask image and the content mask image to obtain an end image set;

[0051] The target mapping network is the first mapping network or the second mapping network, and the starting image set is any one of the first tracer image set, the second tracer image set, the first mapping data set, or the third mapping data set.

[0052] Furthermore, in one embodiment of the present application, the target mapping network further includes a discriminator, and the discriminator is used to perform the following steps:

[0053] Performing convolution normalization on the attention mask image and the corresponding endpoint image set to obtain a normalized image set;

[0054] Convolution discrimination is performed on the normalized image set to obtain an attention discrimination result.

[0055] In a third aspect, an embodiment of the present application further provides an electronic device, including:

[0056] at least one processor;

[0057] at least one memory for storing at least one program;

[0058] When the at least one program is executed by the at least one processor, the at least one processor implements the method of the first aspect described above.

[0059] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, which stores a program executable by a processor, and the program executable by the processor is used to implement the method of the first aspect when executed by the processor.

[0060] The advantages and benefits of the present application will be partially given in the following description, and partially become apparent from the following description, or be understood through the practice of the present application:

[0061] An imaging method, system, device and medium of a dual tracer model disclosed in an embodiment of the present application, wherein the imaging method obtains a source tracer image of a subject; inputs the source tracer image into a trained dual tracer model to obtain a flow tracer image; the dual tracer model includes a first mapping network and a second mapping network, and the trained dual tracer model is obtained by the following steps: obtaining a first tracer data set and a second tracer data set of the subject; preprocessing the first tracer data set and the second tracer data set to obtain a dual tracer image set, the dual tracer image set including a first tracer image set and a second tracer image set; inputting the first tracer image set into the first mapping network to obtain a first mapping data set and a second mapping data set output by the first mapping network, the first mapping data set being a tracer image set corresponding to the second tracer image set and generated by mapping the first mapping network; The second mapping data set is a tracer image set restored and generated by the first mapping network based on the first mapping data set, and the second mapping data set corresponds to the first tracer image set; the second tracer image set is input into the second mapping network to obtain a third mapping data set and a fourth mapping data set output by the second mapping network, the third mapping data set is a tracer image set corresponding to the first tracer image set and generated by the second mapping network; the fourth mapping data set is a tracer image set restored and generated by the second mapping network based on the third mapping data set, and the fourth mapping data set corresponds to the second tracer image set; according to the first tracer image set, the second tracer image set, the first mapping data set, the second mapping data set, the third mapping data set and the fourth mapping data set, the parameters of the initialized dual tracer model are updated to obtain the trained dual tracer model. The imaging method realizes mutual generation between the first tracer and the second tracer based on the first mapping network and the second mapping network, which can effectively improve the safety of tracer imaging, and improve the imaging experience and convenience of the subject in tracer imaging. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the embodiments of the present application or the drawings of the related technical solutions in the prior art are introduced below. It should be understood that the drawings introduced below are only for the convenience of clearly expressing some embodiments of the technical solutions of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0063] Figure 1 A schematic flow chart of an imaging method for a dual tracer model provided in an embodiment of the present application;

[0064] Figure 2 A schematic diagram of a training process of a dual tracer model provided in an embodiment of the present application;

[0065] Figure 3 A flow chart of the conversion of a tracer image by a target mapping network provided in an embodiment of the present application;

[0066] Figure 4 A schematic diagram of the structure of an imaging system of a dual tracer model provided in an embodiment of the present application;

[0067] Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0068] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present application, and are not to be construed as limitations on the present application. For the step numbers in the following embodiments, they are only provided for the convenience of explanation, and the order between the steps is not limited in any way, and the execution order of each step in the embodiment can be adaptively adjusted according to the understanding of those skilled in the art.

[0069] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0070] At present, the traditional dual-tracer imaging method usually uses two PET / CT scans with different tracers on the same subject at different times to obtain two different tracer images of the subject. This method causes the subject to receive a higher radiation dose in a shorter period of time and is not very safe. At the same time, this method also increases the time and economic costs of the subjects, and the imaging experience and convenience of tracer imaging are not satisfactory.

[0071] In view of this, an embodiment of the present invention provides an imaging method of a dual-tracer model, which realizes the mutual generation between the first tracer and the second tracer, that is, based on any tracer image, another tracer image is generated, thereby achieving a dual-tracer imaging effect, reducing the radiation dose received by the subject, and thereby effectively improving the safety of tracer imaging, as well as improving the imaging experience and convenience of the subject for tracer imaging. In addition, the imaging method is also based on attention mask generation, which can end-point learn the difference content of different tracers, so as to better obtain the specific performance of different tracers and effectively improve the imaging effect of tracer imaging.

[0072] Reference Figure 1 and Figure 2 In an embodiment of the present application, an imaging method of a dual tracer model includes:

[0073] Step 110, obtaining a source tracer image of the subject;

[0074] Step 120, inputting the source tracer image into the trained dual tracer model to obtain a flow tracer image;

[0075] In an embodiment of the present application, a tracer image corresponding to any tracer received by a subject can be used as a source tracer image, and input into a trained dual-tracer model for image generation to generate a tracer image corresponding to another tracer, and the tracer image can be used as a flow tracer image.

[0076] The dual-tracer model includes a first mapping network and a second mapping network, and the trained dual-tracer model is obtained by the following steps:

[0077] Step 130, obtaining a first tracer data set and a second tracer data set of the subject;

[0078] In an embodiment of the present application, the tracer corresponding to the first tracer data set is different from the tracer corresponding to the second tracer data set, which may be a dual tracer data set publicly available on the Internet, or, with the consent of the subject, the PET images of the subject under different tracers may be divided into the first tracer data set and the second tracer data set.

[0079] Step 140: preprocess the first tracer data set and the second tracer data set to obtain a dual-tracer image set, wherein the dual-tracer image set includes a first tracer image set and a second tracer image set;

[0080] In some embodiments, preprocessing the first tracer data set and the second tracer data set to obtain a dual tracer image set includes:

[0081] A1. performing image registration on the second tracer data set according to the first tracer data set to obtain a registered first tracer data set and a registered second tracer data set;

[0082] Further, performing image registration on the second tracer data set according to the first tracer data set includes:

[0083] A11, obtaining a registration transformation matrix corresponding to the first tracer data set and the second tracer data set;

[0084] A12. Perform matrix registration on the first tracer dataset and the second tracer dataset according to the registration transformation matrix to obtain the registered first tracer dataset and the registered second tracer dataset.

[0085] In an embodiment of the present application, the preprocessing in step 140 includes image registration and data enhancement, wherein the image registration in step A1 can obtain a registration transformation matrix corresponding to the first tracer data set and the second tracer data set, and the registration transformation matrix is ​​used to map the coordinate points of the first tracer data set to the coordinate system of the second tracer data set, or to map the coordinate points of the second tracer data set to the coordinate system of the first tracer data set.

[0086] Exemplarily, the registration transformation matrix can be obtained by the following steps: first, a first CT image corresponding to the first tracer data set and a second CT image corresponding to the second tracer data set are obtained; then, the bed is removed from the first CT image and the second CT image are resampled to obtain the removed first CT image and the removed second CT image; then, the ElasticSyN transformation in the ANTspy (Advanced normalization tools in Python) library in Python is called to perform corresponding processing on the removed first CT image and the removed second CT image to obtain the registration transformation matrix.

[0087] It can be understood that after obtaining the registration transformation matrix, the coordinate points of the corresponding tracer data set can be mapped according to the mapping relationship of the registration transformation matrix. The tracer data set can be the first tracer data set or the second tracer data set, so as to obtain the registered first tracer data set and the registered second tracer data set.

[0088] A2. Performing data enhancement on the registered first tracer data set and the registered second tracer data set to obtain the dual tracer image set.

[0089] In an embodiment of the present application, the data enhancement of step A2 may be at least one of rotation transformation, translation transformation, flip transformation, offset transformation, scaling transformation, scale transformation, contrast transformation, noise disturbance and the like, performed on the aligned first tracer data set and the aligned second tracer data set, which will not be described in detail in the present application.

[0090] Step 150: input the first tracer image set into the first mapping network to obtain a first mapping data set and a second mapping data set output by the first mapping network, wherein the first mapping data set is a tracer image set corresponding to the second tracer image set and generated by mapping of the first mapping network; the second mapping data set is a tracer image set restored and generated by the first mapping network based on the first mapping data set, and the second mapping data set corresponds to the first tracer image set;

[0091] Step 160: input the second tracer image set into the second mapping network to obtain a third mapping data set and a fourth mapping data set output by the second mapping network, wherein the third mapping data set is a tracer image set corresponding to the first tracer image set and generated by mapping of the second mapping network; the fourth mapping data set is a tracer image set restored and generated by the second mapping network based on the third mapping data set, and the fourth mapping data set corresponds to the second tracer image set;

[0092] In an embodiment of the present application, the first mapping network and the second mapping network can be constructed based on a 3D attention-guided cycle-consistent generative adversarial network, wherein the first mapping network is used to receive a first tracer image in a first tracer image set, and generate a first mapping image according to the input first tracer image mapping, and the tracer corresponding to the first mapping image is the same as the tracer corresponding to the second tracer image set.

[0093] It can be understood that the first mapping network also generates a second mapping image by restoring the first mapping image generated according to the mapping. The second mapping image can be a tracer image in the second mapping data set, and the tracer corresponding to the second mapping image is the same as the tracer corresponding to the first tracer image set. It should be noted that the second mapping network is similar to the content of the first mapping network, and can be simply deduced by analogy, so this application will not repeat it here.

[0094] In some embodiments, the target mapping network is used to perform the following steps:

[0095] B1. Performing mask generation processing on the starting image set corresponding to the target mapping network to obtain an attention mask image and a content mask image;

[0096] Furthermore, the step B1, performing mask generation processing on the starting image set corresponding to the target mapping network to obtain an attention mask image and a content mask image, includes:

[0097] B11, extracting encoding features from the starting image set to obtain a feature data set;

[0098] B12, performing residual convolution on the feature data set to obtain a residual data set;

[0099] B13. Decode and sample the residual data set to obtain the attention mask image and the content mask image.

[0100] In an embodiment of the present application, the embodiment of the present application takes the target mapping network as the first mapping network and the starting image set as the first tracer image set as an example. The first mapping network can be composed of two attention-guided generator networks and an attention-guided discriminator network, wherein the output of the first attention-guided generator network is connected to the input of the second attention-guided generator network, the first attention-guided generator network is used to generate a first mapping data set according to the first mapping network mapping, and the second attention-guided generator network is used to restore and generate a second mapping data set according to the first mapping data set provided by the previous generator network.

[0101] It can be understood that the attention-guided generator network in the embodiment of the present application includes an encoding module, a residual convolution module and a decoder module, the encoding module is used to implement step B11, the residual convolution module is used to implement step B12, and the decoder module is used to implement step B13. Specifically, the encoding module in the embodiment of the present application includes a convolution layer with a convolution kernel size of 7×7×7, a batch normalization (BatchNorm) layer and a rectified linear unit (ReLU); in the residual convolution module, it includes 8 residual convolutions, each residual convolution includes a convolution layer with a stride of 1 and a convolution kernel of 3×3×3; in the decoding module, it includes a convolution layer with a stride of 1 / 2 and a convolution kernel of 3×3×3, a BatchNorm layer and a ReLU layer, and the ReLU layer is used for upsampling.

[0102] B2. According to the starting image set, the attention mask image and the content mask image are fused to obtain an end image set;

[0103] The target mapping network is the first mapping network or the second mapping network, and the starting image set is any one of the first tracer image set, the second tracer image set, the first mapping data set, or the third mapping data set.

[0104] In an embodiment of the present application, a certain first tracer image in the first tracer image set can be fused with the corresponding attention mask image and content mask image to obtain a first mapping image. The same is true for the remaining first tracer images. The first mapping data set can be obtained by integrating all the first mapping images.

[0105] It can be understood that the end image set corresponds to the starting image set, and when the starting image set is the first tracer image set input to the first mapping network, the end image set can be the first mapping data set; when the starting image set is the first mapping data set generated by mapping the first mapping network, the end image set can be the second mapping data set. Alternatively, when the starting image set is the second tracer image set input to the second mapping network, the end image set can be the third mapping data set; when the starting image set is the third mapping data set generated by mapping the second mapping network, the end image set can be the fourth mapping data set.

[0106] For example, refer to Figure 3 , Figure 3 : is a flow chart of the conversion of the target mapping network to the tracer image. The embodiment of the present application takes a second tracer image in the second tracer image set input to the second mapping network as an example. At this time, the images corresponding to the second tracer image are the third mapping image and the fourth mapping image, respectively. The third mapping image is an image in the third mapping data set, and the fourth mapping image is an image in the fourth mapping data set. Specifically, after receiving the second tracer image, the second mapping network generates an attention mask image and a content mask image through the first attention-guided generator network, and then fuses the input image (i.e., the second tracer image), the content mask image, and the attention mask image to generate a third mapping image. The third mapping image is used as the input of the second attention-guided generator network, and the above process is repeated to generate a fourth mapping image.

[0107] In some embodiments, the target mapping network further comprises a discriminator, the discriminator being configured to perform the following steps:

[0108] C1, performing convolution normalization on the attention mask image and the corresponding endpoint image set to obtain a normalized image set;

[0109] C2. Perform convolution discrimination on the normalized image set to obtain an attention discrimination result.

[0110] In an embodiment of the present application, the discriminator may be the aforementioned attention-guided discriminator network, which is used to discriminate the attention mask image and the corresponding tracer image in the endpoint image set. Specifically, when the endpoint image set is a first mapping data set, the discriminator is used to discriminate the attention mask image and the corresponding first mapping image; or, when the endpoint image set is a third mapping data set, the discriminator is used to discriminate the attention mask image and the corresponding third mapping image.

[0111] It can be understood that the discriminator in the embodiment of the present application includes 5 convolutional layers, specifically, it can be composed of 4 convolutional layers with a 3D stride of 2 and a convolution kernel size of 4×4×4, a BatchNorm layer, a ReLU layer, and the last one is a convolutional layer with a stride of 1, wherein the 4 convolutional layers, the BatchNorm layer and the ReLU layer are used to realize the convolution normalization of the attention mask image and the corresponding tracer image, and the convolutional layer with a stride of 1 is used to realize image discrimination of the normalized image set, thereby obtaining the attention discrimination result, which can be used to calculate the adversarial loss function of the target mapping network.

[0112] Step 170: Update the parameters of the initialized dual tracer model according to the first tracer image set, the second tracer image set, the first mapping data set, the second mapping data set, the third mapping data set and the fourth mapping data set to obtain the trained dual tracer model.

[0113] In some embodiments, the step 170, updating the parameters of the initialized dual tracer model according to the first tracer image set, the second tracer image set, the first mapping data set, the second mapping data set, the third mapping data set, and the fourth mapping data set to obtain the trained dual tracer model, includes:

[0114] D1, calculating a target loss function according to the first tracer image set, the second tracer image set, the first mapping data set, the second mapping data set, the third mapping data set and the fourth mapping data set;

[0115] D2. According to the target loss function, the initialized dual-tracer model is trained and updated to obtain the trained dual-tracer model.

[0116] In an embodiment of the present application, for the first mapping network, step D1 can be to calculate the pixel loss function of the first mapping network during mapping generation based on the first tracer image set and the first mapping data set; and to calculate the cycle consistency loss function of the first mapping network based on the first tracer image set and the second mapping data set.

[0117] Exemplarily, the equivalent expression of the cycle consistency loss function of the first mapping network can be expressed as:

[0118] L cycle (G A→B , G B→A )

[0119] =E A～pdata (A)[||G B→A (G A→B (A))-A||1]+E B～pdata(B) [||G A→B (GB→A(B))-B||1]

[0120] Among them, L cycle (G A→B , G B→A ) is the cycle consistency loss function of the first mapping network, A is the tracer image corresponding to the first tracer; B is the tracer image corresponding to the second tracer; G A→B Used to characterize the first attention-guided generator network in the first mapping network; G B→A Used to characterize the second attention-guided generator network in the first mapping network; ||·||1 is the L1 norm; E A～pdata(A) is the cross entropy probability distribution of the tracer image corresponding to the first tracer; E B～pdata(B) is the cross entropy probability distribution of the tracer image corresponding to the second tracer.

[0121] The equivalent expression of the pixel loss function of the first mapping network can be expressed as:

[0122] L pixel (G A→B , G B→A )=E A～pdata(A) [||G A→B (A)-A||1]+E B～pdata(B) [||G B→A (B)-B||1]

[0123] Among them, L pixel (G A→B , G B→A ) is the pixel loss function of the first mapping network.

[0124] It can be understood that the pixel loss function and cycle consistency loss function of the second mapping network are similar to those of the first mapping network, and can be simply deduced by analogy. After obtaining the pixel loss function and cycle consistency loss function of the dual tracer model, the adversarial loss function of each mapping network can be combined, where the adversarial loss function of each mapping network is an adversarial loss function based on attention guidance.

[0125] The equivalent expression of the attention-guided adversarial loss function of the first mapping network can be:

[0126] L AGGAN (G A→B , D BX )

[0127] =E B～pdata(B) [logD BX ([M B ,B])]+E A～pdata(A) [log(1-D BX ([M B , G A→B (A)]))]

[0128] Among them, L AGGAN (G A→B , D BX ) is the attention-guided adversarial loss function of the first mapping network; D BX is the attention-guided discriminator network in the first mapping network; M B It is the attention mask image generated by A mapping to B in the first attention-guided generator network of the first mapping network, that is, the attention mask image corresponding to the first mapping image.

[0129] The attention-guided loss function of the second mapping network is similar to that of the first mapping network. It can be simply deduced that the target loss function can be simply calculated based on the above related loss function, and then the training loss value is determined based on the target loss function, so as to update the parameters of the dual tracer model and obtain a trained dual tracer model.

[0130] An imaging system of a dual tracer model proposed according to an embodiment of the present application is described in detail below with reference to the accompanying drawings.

[0131] Reference Figure 4 , an imaging system of a dual tracer model proposed in an embodiment of the present application comprises:

[0132] An acquisition module 101 is used to acquire a source tracer image of a subject;

[0133] A processing module 102 is used to input the source tracer image into a trained dual tracer model to obtain a flow tracer image;

[0134] The dual-tracer model includes a first mapping network and a second mapping network, and the trained dual-tracer model is obtained by the following steps:

[0135] acquiring a first tracer data set and a second tracer data set of a subject;

[0136] Preprocessing the first tracer data set and the second tracer data set to obtain a dual-tracer image set, wherein the dual-tracer image set includes a first tracer image set and a second tracer image set;

[0137] Inputting the first tracer image set into the first mapping network, obtaining a first mapping data set and a second mapping data set output by the first mapping network, wherein the first mapping data set is a tracer image set corresponding to the second tracer image set and generated by mapping of the first mapping network; the second mapping data set is a tracer image set restored and generated by the first mapping network based on the first mapping data set, and the second mapping data set corresponds to the first tracer image set;

[0138] Inputting the second tracer image set into the second mapping network, obtaining a third mapping data set and a fourth mapping data set output by the second mapping network, wherein the third mapping data set is a tracer image set corresponding to the first tracer image set and generated by mapping of the second mapping network; the fourth mapping data set is a tracer image set restored and generated by the second mapping network based on the third mapping data set, and the fourth mapping data set corresponds to the second tracer image set;

[0139] The parameters of the initialized dual-tracer model are updated according to the first tracer image set, the second tracer image set, the first mapping data set, the second mapping data set, the third mapping data set and the fourth mapping data set to obtain the trained dual-tracer model.

[0140] Furthermore, in one embodiment of the present application, the imaging system further includes a target mapping network, and the target mapping network is used to perform the following steps:

[0141] Performing mask generation processing on a starting image set corresponding to the target mapping network to obtain an attention mask image and a content mask image;

[0142] According to the starting image set, performing image fusion on the attention mask image and the content mask image to obtain an end image set;

[0143] The target mapping network is the first mapping network or the second mapping network, and the starting image set is any one of the first tracer image set, the second tracer image set, the first mapping data set, or the third mapping data set.

[0144] Furthermore, in one embodiment of the present application, the target mapping network further includes a discriminator, and the discriminator is used to perform the following steps:

[0145] Performing convolution normalization on the attention mask image and the corresponding endpoint image set to obtain a normalized image set;

[0146] Convolution discrimination is performed on the normalized image set to obtain an attention discrimination result.

[0147] It can be understood that the contents of the above method embodiments are all applicable to the present system embodiments, the functions specifically implemented by the present system embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0148] Reference Figure 5 , the embodiment of the present application further provides an electronic device, including:

[0149] at least one processor 201;

[0150] At least one memory 202, used to store at least one program;

[0151] When the at least one program is executed by the at least one processor 201, the at least one processor 201 implements the above method embodiment.

[0152] Similarly, it can be understood that the contents of the above method embodiments are all applicable to the present device embodiments, the functions specifically implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0153] The embodiment of the present application further provides a computer-readable storage medium, in which a program executable by the processor 201 is stored. The program executable by the processor 201 is used to implement the above-mentioned method embodiment when executed by the processor 201.

[0154] Similarly, the contents of the above method embodiments are all applicable to the computer-readable storage medium embodiments. The functions specifically implemented by the computer-readable storage medium embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0155] In some selectable embodiments, the function / operation mentioned in the block diagram may not occur in the order mentioned in the operation diagram. For example, depending on the function / operation involved, the two boxes shown in succession can actually be executed substantially simultaneously or the boxes can sometimes be executed in reverse order. In addition, the embodiment presented and described in the flow chart of the application is provided by way of example, for the purpose of providing a more comprehensive understanding of technology. The disclosed method is not limited to the operation and logic flow presented herein. Selectable embodiments are expected, wherein the order of various operations is changed and the sub-operation of a part described as a larger operation is performed independently.

[0156] In addition, although the present application is described in the context of functional modules, it should be understood that, unless otherwise specified, one or more of the functions and / or features can be integrated into a single physical device and / or software module, or one or more functions and / or features can be implemented in a separate physical device or software module. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding the present application. More specifically, in view of the properties, functions, and internal relationships of the various functional modules in the device disclosed herein, the actual implementation of the module will be understood within the conventional techniques of the engineer. Therefore, those skilled in the art can implement the present application set forth in the claims without excessive experimentation using ordinary techniques. It is also understood that the specific concepts disclosed are merely illustrative and are not intended to limit the scope of the present application, which is determined by the full scope of the attached claims and their equivalents.

[0157] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method of the embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0158] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0159] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0160] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0161] In the above description of this specification, the description with reference to the terms "one embodiment / example", "another embodiment / example" or "certain embodiments / examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0162] Although the embodiments of the present application have been shown and described, those skilled in the art will appreciate that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present application, and that the scope of the present application is defined by the claims and their equivalents.

[0163] The above is a specific description of the preferred implementation of the present application, but the present application is not limited to the embodiments. Technical personnel familiar with the field can make various equivalent modifications or substitutions without violating the spirit of the present application. These equivalent modifications or substitutions are all included in the scope defined by the claims of the present application.

Claims

1. An imaging method of a dual tracer model, characterized in that: include: obtaining source tracer images of the subject; Inputting the source tracer image into a trained dual-tracer model to obtain a flow tracer image; The dual-tracer model includes a first mapping network and a second mapping network, and the trained dual-tracer model is obtained by the following steps: acquiring a first tracer data set and a second tracer data set of a subject; Preprocessing the first tracer data set and the second tracer data set to obtain a dual-tracer image set, wherein the dual-tracer image set includes a first tracer image set and a second tracer image set; Inputting the first tracer image set into the first mapping network to obtain a first mapping data set and a second mapping data set output by the first mapping network, wherein the first mapping data set is a tracer image set corresponding to the second tracer image set and generated by mapping of the first mapping network; The second mapping data set is a tracer image set restored and generated by the first mapping network based on the first mapping data set, and the second mapping data set corresponds to the first tracer image set; Inputting the second tracer image set into the second mapping network to obtain a third mapping data set and a fourth mapping data set output by the second mapping network, wherein the third mapping data set is a tracer image set corresponding to the first tracer image set and generated by mapping of the second mapping network; The fourth mapping data set is a tracer image set restored and generated by the second mapping network based on the third mapping data set, and the fourth mapping data set corresponds to the second tracer image set; The parameters of the initialized dual-tracer model are updated according to the first tracer image set, the second tracer image set, the first mapping data set, the second mapping data set, the third mapping data set and the fourth mapping data set to obtain the trained dual-tracer model.

2. The imaging method according to claim 1, characterized in that: The preprocessing of the first tracer data set and the second tracer data set to obtain a dual-tracer image set includes: performing image registration on the second tracer data set according to the first tracer data set to obtain a registered first tracer data set and a registered second tracer data set; Data enhancement is performed on the registered first tracer data set and the registered second tracer data set to obtain the dual tracer image set.

3. The imaging method according to claim 2, characterized in that: The performing image registration on the second tracer data set according to the first tracer data set comprises: Obtaining a registration transformation matrix corresponding to the first tracer data set and the second tracer data set; According to the registration transformation matrix, matrix registration is performed on the first tracer data set and the second tracer data set to obtain the registered first tracer data set and the registered second tracer data set.

4. The imaging method according to claim 1, characterized in that: The target mapping network is used to perform the following steps: Performing mask generation processing on a starting image set corresponding to the target mapping network to obtain an attention mask image and a content mask image; According to the starting image set, performing image fusion on the attention mask image and the content mask image to obtain an end image set; The target mapping network is the first mapping network or the second mapping network, and the starting image set is any one of the first tracer image set, the second tracer image set, the first mapping data set, or the third mapping data set.

5. The imaging method according to claim 4, characterized in that: The mask generation process is performed on the starting image set corresponding to the target mapping network to obtain an attention mask image and a content mask image, including: Extracting encoding features from the starting image set to obtain a feature data set; Performing residual convolution on the feature data set to obtain a residual data set; The residual data set is decoded and sampled to obtain the attention mask image and the content mask image.

6. The imaging method according to claim 4, characterized in that: The target mapping network also includes a discriminator, which is used to perform the following steps: Performing convolution normalization on the attention mask image and the corresponding endpoint image set to obtain a normalized image set; Convolution discrimination is performed on the normalized image set to obtain an attention discrimination result.

7. The imaging method according to claim 1, characterized in that: The step of updating parameters of the initialized dual-tracer model according to the first tracer image set, the second tracer image set, the first mapping data set, the second mapping data set, the third mapping data set, and the fourth mapping data set to obtain the trained dual-tracer model comprises: calculating a target loss function based on the first tracer image set, the second tracer image set, the first mapping data set, the second mapping data set, the third mapping data set, and the fourth mapping data set; The initialized dual-tracer model is trained and updated according to the target loss function to obtain the trained dual-tracer model.

8. An imaging system of a dual tracer model, characterized in that: include: an acquisition module, for acquiring a source tracer image of a subject; a processing module, configured to input the source tracer image into a trained dual tracer model to obtain a flow tracer image; The dual-tracer model includes a first mapping network and a second mapping network, and the trained dual-tracer model is obtained by the following steps: acquiring a first tracer data set and a second tracer data set of a subject; Preprocessing the first tracer data set and the second tracer data set to obtain a dual-tracer image set, wherein the dual-tracer image set includes a first tracer image set and a second tracer image set; Inputting the first tracer image set into the first mapping network to obtain a first mapping data set and a second mapping data set output by the first mapping network, wherein the first mapping data set is a tracer image set corresponding to the second tracer image set and generated by mapping of the first mapping network; The second mapping data set is a tracer image set restored and generated by the first mapping network based on the first mapping data set, and the second mapping data set corresponds to the first tracer image set; Inputting the second tracer image set into the second mapping network to obtain a third mapping data set and a fourth mapping data set output by the second mapping network, wherein the third mapping data set is a tracer image set corresponding to the first tracer image set and generated by mapping of the second mapping network; The fourth mapping data set is a tracer image set restored and generated by the second mapping network based on the third mapping data set, and the fourth mapping data set corresponds to the second tracer image set; The parameters of the initialized dual-tracer model are updated according to the first tracer image set, the second tracer image set, the first mapping data set, the second mapping data set, the third mapping data set and the fourth mapping data set to obtain the trained dual-tracer model.

9. The imaging system according to claim 8, characterized in that The imaging system further comprises a target mapping network, wherein the target mapping network is configured to perform the following steps: Performing mask generation processing on a starting image set corresponding to the target mapping network to obtain an attention mask image and a content mask image; According to the starting image set, performing image fusion on the attention mask image and the content mask image to obtain an end image set; The target mapping network is the first mapping network or the second mapping network, and the starting image set is any one of the first tracer image set, the second tracer image set, the first mapping data set, or the third mapping data set.

10. The imaging system according to claim 9, characterized in that The target mapping network also includes a discriminator, which is used to perform the following steps: Performing convolution normalization on the attention mask image and the corresponding endpoint image set to obtain a normalized image set; Convolution discrimination is performed on the normalized image set to obtain an attention discrimination result.

Images