AD disease assessment guiding method and device based on brain image generation

Abstract

The invention relates to an AD disease assessment guiding method and device based on brain image generation. The method comprises the following steps: acquiring a brain glucose metabolism diagram of a user, and performing registration processing through an image fusion template based on the brain glucose metabolism diagram to obtain a brain image map corresponding to the brain glucose metabolism diagram; on the basis of the brain image map, synthesizing a brain amyloid protein deposition map of the user through an image synthesis network, and on the basis of the brain image map and the brain amyloid protein deposition map, evaluating the brain amyloid protein deposition map of the user through an evaluation model. Generating a first evaluation result corresponding to the brain image map and a second evaluation result corresponding to the brain amyloid protein deposition map; and based on the first assessment result and the second assessment result, generating disease assessment guidance information of the user. By adopting the method, the assessment guidance efficiency of the AD disease can be improved.

Description

Technical Field

[0001] The present application relates to the field of medical image processing technology, and in particular to a method and device for AD disease assessment guidance based on brain image generation. Background Art

[0002] Alzheimer's Disease (AD) is the most common neurodegenerative disease, and the deposition of β-amyloid protein (Aβ) in the brain is one of its earliest biomarkers. Aβ-PET (positron emission tomography) is currently the gold standard imaging technology for detecting pathological Aβ deposition. Aβ-PET uses radioactive tracers to label β-amyloid protein, which can visually display its distribution and perform quantitative analysis, providing an important basis for the early diagnosis and efficacy evaluation of AD. However, the cost of Aβ tracers is high, and they require expensive and complex cyclotron-assisted optimization, which makes the auxiliary guidance for the early diagnosis and efficacy evaluation of AD diseases costly and inefficient. Therefore, how to improve the efficiency of evaluation guidance for AD diseases is the current research focus.

[0003] The traditional technology is to use 18F-fluorodeoxyglucose (18F-FDG) PET to evaluate cellular glucose metabolism, reflect neuronal synaptic activity, and thus indicate AD-related neuronal loss and degeneration (i.e., assessing the condition of AD disease). However, this method is not only costly and inefficient, but also requires multiple examinations, which has a huge time cost, resource cost, and physical consumption for patients, resulting in low efficiency in the assessment and guidance of AD disease. Summary of the invention

[0004] Based on this, it is necessary to provide a method, apparatus, computer device, computer-readable storage medium and computer program product for AD disease assessment guidance generated based on brain images to address the above technical problems.

[0005] In a first aspect, the present application provides an AD disease assessment guidance method based on brain image generation, comprising:

[0006] Acquire a user's brain glucose metabolism map, and perform registration processing based on the brain glucose metabolism map through an image fusion template to obtain a brain image map corresponding to the brain glucose metabolism map;

[0007] Based on the brain image, synthesize the user's brain amyloid protein deposition map through an image synthesis network, and based on the brain image and the brain amyloid protein deposition map, generate a first evaluation result corresponding to the brain image and a second evaluation result corresponding to the brain amyloid protein deposition map through an evaluation model;

[0008] Based on the first assessment result and the second assessment result, disease assessment guidance information for the user is generated.

[0009] Optionally, before obtaining the user's brain glucose metabolism map, the method further includes:

[0010] Acquire sample brain glucose metabolism maps of multiple users and sample brain amyloid protein deposition maps of each user, and perform registration processing on each sample brain glucose metabolism map and each sample brain amyloid protein deposition map respectively through the image fusion template to obtain a first brain image map corresponding to each sample brain glucose metabolism map and a second brain image map corresponding to each sample brain amyloid protein deposition map;

[0011] Based on each of the first brain image maps and each of the second brain image maps, generating a synthesized second brain image map corresponding to each of the first brain image maps and a synthesized first brain image map corresponding to each of the second brain image maps through an initial image synthesis network;

[0012] Inputting each of the first brain image maps and each of the synthesized first brain image maps into a first diagnostic network to obtain first synthesis deviation information of the initial image synthesis network, and inputting each of the second brain image maps and each of the synthesized second brain image maps into a second diagnostic network to obtain second synthesis deviation information of the initial image synthesis network;

[0013] When the first synthesis deviation information is greater than the first deviation threshold, or the second synthesis deviation information is greater than the second deviation threshold, adjusting the initial image synthesis network based on each of the first synthesis deviation information and each of the second synthesis deviation information to obtain a new image synthesis network;

[0014] The new image synthesis network replaces the initial image synthesis network, and the step of generating a synthetic second brain image corresponding to each of the first brain image images and a synthetic first brain image corresponding to each of the second brain image images through the initial image synthesis network based on each of the first brain image images and each of the second brain image images is returned to the step of generating a synthetic second brain image corresponding to each of the first brain image images and a synthetic first brain image corresponding to each of the second brain image images through the initial image synthesis network, until all the first synthesis deviation information is not greater than the first deviation threshold and all the second synthesis deviation information is not greater than the second deviation threshold, and the initial image synthesis network obtained by the last iteration is used as the image synthesis network.

[0015] Optionally, the step of inputting each of the first brain image maps and each of the synthesized first brain image maps into a first diagnostic network to obtain first synthesis deviation information of the initial image synthesis network, and inputting each of the second brain image maps and each of the synthesized second brain image maps into a second diagnostic network to obtain second synthesis deviation information of the initial image synthesis network includes:

[0016] Inputting each of the first brain image maps and each of the synthesized first brain image maps into a first diagnostic network, identifying metabolic deviation information between each of the first brain image maps and each of the synthesized first brain image maps, and based on each of the metabolic deviation information, identifying anabolic metabolic deviation abnormal range of the initial image synthesis network and anabolic metabolic sensitivity value of the initial image synthesis network;

[0017] Inputting each of the second brain image maps and each of the synthesized second brain image maps into a second diagnostic network, identifying a deposition deviation value between each of the second brain image maps and each of the synthesized second brain image maps, and based on each of the deposition deviation values, identifying a synthetic deposition deviation abnormal range of the initial image synthesis network and a synthetic deposition sensitivity value of the initial image synthesis network;

[0018] The anabolic deviation abnormal range and the anabolic sensitivity value are used as first synthetic deviation information of the initial image synthesis network, and the synthetic deposition deviation abnormal range and the synthetic deposition sensitivity value are used as second synthetic deviation information of the initial image synthesis network.

[0019] Optionally, the registering process is performed based on the brain glucose metabolism map by using an image fusion template to obtain a brain image map corresponding to the brain glucose metabolism map, including:

[0020] Based on the brain parenchyma mask information of the image fusion template, extracting the brain parenchyma image data of the user in the brain glucose metabolism map through a threshold screening strategy;

[0021] Identifying the voxel value of each sub-brain parenchyma image data in the brain parenchyma image data, and adjusting each of the sub-brain parenchyma image data according to a preset voxel value range to obtain each target sub-brain parenchyma image data;

[0022] The image data corresponding to all target sub-brain parenchyma image data are used as the brain image map corresponding to the brain glucose metabolism map.

[0023] Optionally, the generating, based on the brain image map and the brain amyloid protein deposition map, a first evaluation result corresponding to the brain image map and a second evaluation result corresponding to the brain amyloid protein deposition map by an evaluation model includes:

[0024] Based on the brain image, identifying the metabolic distribution information of the user, and based on the metabolic distribution information, identifying each abnormal metabolic range in the metabolic distribution information and the abnormal metabolic information corresponding to each abnormal metabolic range through the metabolic abnormality identification network of the evaluation model;

[0025] Using each of the abnormal metabolic ranges and each of the abnormal metabolic information as a first evaluation result corresponding to the brain image;

[0026] Based on the brain amyloid protein deposition map, identifying the amyloid protein deposition distribution information of the user, and based on the amyloid protein deposition distribution information, identifying each abnormal deposition range in the amyloid protein deposition distribution information and the abnormal deposition information corresponding to the abnormal deposition range through the deposition abnormality identification network of the evaluation model;

[0027] Each of the abnormal deposition ranges and each of the abnormal deposition information is used as a second evaluation result corresponding to the brain amyloid protein deposition map.

[0028] Optionally, generating disease assessment guidance information for the user based on the first assessment result and the second assessment result includes:

[0029] identifying a first brain region corresponding to each of the abnormal metabolic ranges, and identifying a degree of metabolic abnormality of each of the first brain regions based on abnormal metabolic information of each of the abnormal metabolic ranges;

[0030] Identifying a second brain region corresponding to each of the abnormal deposition ranges, and identifying a deposition abnormality degree of each of the second brain regions based on abnormal deposition information corresponding to each of the abnormal deposition ranges;

[0031] Based on the metabolic abnormality level of each of the first brain regions and the deposition abnormality level of each of the second brain regions, the abnormal brain regions of the user and the target abnormality level of each of the abnormal brain regions are identified, and based on the abnormal range corresponding to each of the abnormal brain regions and the target abnormality level of each of the abnormal brain regions, disease assessment guidance information for the user is generated through a preset disease assessment guidance template.

[0032] In a second aspect, the present application also provides an AD disease assessment guidance device based on brain image generation, comprising:

[0033] An acquisition module is used to acquire a user's brain glucose metabolism map, and based on the brain glucose metabolism map, perform registration processing through an image fusion template to obtain a brain image map corresponding to the brain glucose metabolism map;

[0034] a synthesis module, configured to synthesize a brain amyloid protein deposition map of the user based on the brain image map through an image synthesis network, and to generate a first evaluation result corresponding to the brain image map and a second evaluation result corresponding to the brain amyloid protein deposition map through an evaluation model based on the brain image map and the brain amyloid protein deposition map;

[0035] A generation module is used to generate disease assessment guidance information for the user based on the first assessment result and the second assessment result.

[0036] Optionally, the device further comprises:

[0037] A registration module is used to obtain sample brain glucose metabolism maps of multiple users and sample brain amyloid protein deposition maps of each user, and to perform registration processing on each sample brain glucose metabolism map and each sample brain amyloid protein deposition map respectively through the image fusion template to obtain a first brain image map corresponding to each sample brain glucose metabolism map and a second brain image map corresponding to each sample brain amyloid protein deposition map;

[0038] An image synthesis module, configured to generate, based on each of the first brain image images and each of the second brain image images, a synthesized second brain image image corresponding to each of the first brain image images and a synthesized first brain image image corresponding to each of the second brain image images through an initial image synthesis network;

[0039] a deviation identification module, for inputting each of the first brain image maps and each of the synthesized first brain image maps into a first diagnostic network to obtain first synthesis deviation information of the initial image synthesis network, and for inputting each of the second brain image maps and each of the synthesized second brain image maps into a second diagnostic network to obtain second synthesis deviation information of the initial image synthesis network;

[0040] An adjustment module, configured to adjust the initial image synthesis network based on each of the first synthesis deviation information and each of the second synthesis deviation information to obtain a new image synthesis network when the first synthesis deviation information is greater than the first deviation threshold or the second synthesis deviation information is greater than the second deviation threshold;

[0041] An iteration module is used to replace the initial image synthesis network with the new image synthesis network, and return to execute the step of generating a synthesized second brain image corresponding to each of the first brain image images and a synthesized first brain image corresponding to each of the second brain image images through the initial image synthesis network based on each of the first brain image images and each of the second brain image images, until all the first synthesis deviation information is not greater than the first deviation threshold and all the second synthesis deviation information is not greater than the second deviation threshold, and the initial image synthesis network obtained by the last iteration is used as the image synthesis network.

[0042] Optionally, the deviation identification module is specifically used to:

[0043] Inputting each of the first brain image maps and each of the synthesized first brain image maps into a first diagnostic network, identifying metabolic deviation information between each of the first brain image maps and each of the synthesized first brain image maps, and based on each of the metabolic deviation information, identifying anabolic metabolic deviation abnormal range of the initial image synthesis network and anabolic metabolic sensitivity value of the initial image synthesis network;

[0044] Inputting each of the second brain image maps and each of the synthesized second brain image maps into a second diagnostic network, identifying a deposition deviation value between each of the second brain image maps and each of the synthesized second brain image maps, and based on each of the deposition deviation values, identifying a synthetic deposition deviation abnormal range of the initial image synthesis network and a synthetic deposition sensitivity value of the initial image synthesis network;

[0045] The anabolic deviation abnormal range and the anabolic sensitivity value are used as first synthetic deviation information of the initial image synthesis network, and the synthetic deposition deviation abnormal range and the synthetic deposition sensitivity value are used as second synthetic deviation information of the initial image synthesis network.

[0046] Optionally, the acquisition module is specifically used to:

[0047] Based on the brain parenchyma mask information of the image fusion template, extracting the brain parenchyma image data of the user in the brain glucose metabolism map through a threshold screening strategy;

[0048] Identifying the voxel value of each sub-brain parenchyma image data in the brain parenchyma image data, and adjusting each of the sub-brain parenchyma image data according to a preset voxel value range to obtain each target sub-brain parenchyma image data;

[0049] The image data corresponding to all target sub-brain parenchyma image data are used as the brain image map corresponding to the brain glucose metabolism map.

[0050] Optionally, the synthesis module is specifically used for:

[0051] Based on the brain image, identifying the metabolic distribution information of the user, and based on the metabolic distribution information, identifying each abnormal metabolic range in the metabolic distribution information and the abnormal metabolic information corresponding to each abnormal metabolic range through the metabolic abnormality identification network of the evaluation model;

[0052] Using each of the abnormal metabolic ranges and each of the abnormal metabolic information as a first evaluation result corresponding to the brain image;

[0053] Based on the brain amyloid protein deposition map, protein deposition distribution information of the user is identified, and based on the protein deposition distribution information, each abnormal deposition range in the protein deposition distribution information and abnormal deposition information corresponding to the abnormal deposition range are identified through the deposition abnormality identification network of the evaluation model;

[0054] Each of the abnormal deposition ranges and each of the abnormal deposition information is used as a second evaluation result corresponding to the brain amyloid protein deposition map.

[0055] Optionally, the generating module is specifically used for:

[0056] identifying a first brain region corresponding to each of the abnormal metabolic ranges, and identifying a degree of metabolic abnormality of each of the first brain regions based on abnormal metabolic information of each of the abnormal metabolic ranges;

[0057] Identifying a second brain region corresponding to each of the abnormal deposition ranges, and identifying a deposition abnormality degree of each of the second brain regions based on abnormal deposition information corresponding to each of the abnormal deposition ranges;

[0058] Based on the metabolic abnormality level of each of the first brain regions and the deposition abnormality level of each of the second brain regions, the abnormal brain regions of the user and the target abnormality level of each of the abnormal brain regions are identified, and based on the abnormal range corresponding to each of the abnormal brain regions and the target abnormality level of each of the abnormal brain regions, disease assessment guidance information for the user is generated through a preset disease assessment guidance template.

[0059] In a third aspect, the present application provides a computer device, wherein the computer device comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of any one of the methods in the first aspect are implemented.

[0060] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the steps of any one of the methods in the first aspect are implemented.

[0061] In a fifth aspect, the present application provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the steps of any one of the methods in the first aspect are implemented.

[0062] The above-mentioned AD disease assessment guidance method and device generated based on brain images obtains the user's brain glucose metabolism map, and based on the brain glucose metabolism map, performs registration processing through an image fusion template to obtain a brain image map corresponding to the brain glucose metabolism map; based on the brain image map, synthesizes the user's brain amyloid protein deposition map through an image synthesis network, and based on the brain image map and the brain amyloid protein deposition map, generates a first assessment result corresponding to the brain image map and a second assessment result corresponding to the brain amyloid protein deposition map through an assessment model; based on the first assessment result and the second assessment result, generates disease assessment guidance information for the user. This solution uses an image synthesis network to synthesize the brain glucose metabolism map into a brain amyloid protein deposition map, avoiding the high cost and low efficiency of using Aβ tracers and cyclotrons. It then comprehensively evaluates and guides the user's disease information from two perspectives: the brain imaging map corresponding to the brain glucose metabolism map and the brain amyloid protein deposition map, thereby generating the user's disease assessment guidance information without the need for multiple inspections. It also ensures the accuracy and reliability of the generated disease assessment guidance information for medical staff, thereby comprehensively improving the efficiency of assessment guidance for AD diseases. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0064] Figure 1 A flowchart of a method for generating AD disease assessment guidance based on brain images in one embodiment;

[0065] Figure 2 A schematic diagram of the model structure of a 3D-Uet network model in one embodiment;

[0066] Figure 3 A schematic diagram of a discriminator structure of a PatchGAN discriminator in one embodiment;

[0067] Figure 4 A schematic diagram showing a comparison before and after the normalization process of space normalization in one embodiment;

[0068] Figure 5 A flowchart of an example of AD disease assessment guidelines generated based on brain images in one embodiment;

[0069] Figure 6It is a structural block diagram of an AD disease assessment guidance device generated based on brain images in one embodiment;

[0070] Figure 7 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0071] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0072] The AD disease assessment guidance method based on brain image generation provided in the embodiment of the present application can be applied to the application environment of AD disease assessment guidance. Among them, the terminal can be but not limited to various personal computers, laptops, mid-size computers, etc. Among them, the terminal synthesizes the brain glucose metabolism map into a brain amyloid protein deposition map through an image synthesis network, avoiding the high cost and low efficiency of Aβ tracers and cyclotrons, and then comprehensively evaluates and guides the user's disease information from two perspectives: the brain image map corresponding to the brain glucose metabolism map and the brain amyloid protein deposition map, thereby generating the user's disease assessment guidance information without multiple inspections, and ensuring the accuracy and reliability of the generated disease assessment guidance information for medical staff, thereby comprehensively improving the efficiency of AD disease assessment guidance.

[0073] In an exemplary embodiment, Figure 1 As shown, a method for AD disease assessment guidance based on brain image generation is provided, and the method is applied to a terminal as an example for explanation, including the following steps S101 to S103. Among them:

[0074] Step S101, obtaining a user's brain glucose metabolism map, and based on the brain glucose metabolism map, performing registration processing through an image fusion template to obtain a brain image map corresponding to the brain glucose metabolism map.

[0075] In this embodiment, the terminal collects the user's brain glucose metabolism map through PET examination technology. The brain glucose metabolism map is based on 18 F-deoxyglucose ( 18 F-fluorodeoxyglucose, 18 F-FDG) PET principle 18F-FDG PET image. Then, the terminal performs registration processing based on the brain glucose metabolism map through the image fusion template to obtain the brain image corresponding to the brain glucose metabolism map. Among them, the brain image is the image data containing only brain parenchyma obtained after preprocessing the brain glucose metabolism map. The specific registration process will be described in detail later.

[0076] Step S102, based on the brain image, synthesize the user's brain amyloid protein deposition map through the image synthesis network, and based on the brain image and the brain amyloid protein deposition map, generate a first evaluation result corresponding to the brain image and a second evaluation result corresponding to the brain amyloid protein deposition map through the evaluation model.

[0077] In this embodiment, the terminal synthesizes the user's brain amyloid protein deposition map based on the brain image map through an image synthesis network, and generates a first evaluation result corresponding to the brain image map and a second evaluation result corresponding to the brain amyloid protein deposition map through an evaluation model based on the brain image map and the brain amyloid protein deposition map. The image synthesis network is a neural network that realizes bidirectional synthesis between two modalities based on an improved 3D-UNet model. Specifically, Figure 2 As shown in Figure 2, the 3D-UNet model is used to generate pseudo-paired PET images. It is adjusted by the Adaptive Instance Normalization (AdaIN) module to achieve a bidirectional mapping between the two modalities during the training phase (from 18 F-FDG PET to Aβ PET and vice versa 18 F-FDG PET), that is, synthesizing the brain glucose metabolism map into the brain amyloid protein deposition map, and synthesizing the brain amyloid protein deposition map into the brain glucose metabolism map. The brain amyloid protein deposition map is based on the deposition of β-amyloid protein (Aβ) in the brain, and the Aβ PET image is obtained by Aβ-PET (positron emission tomography) technology. Among them, the evaluation model includes two abnormal recognition networks (i.e., discriminators), which are discriminators based on PatchGAN, such as Figure 3 As shown in the figure, they are 18F-FDG PET discriminator and AβPET discriminator, both of which have the same network structure. The specific evaluation process will be described in detail later.

[0078] Step S103, generating disease assessment guidance information for the user based on the first assessment result and the second assessment result.

[0079] In this embodiment, the terminal generates disease assessment guidance information for the user based on the first assessment result and the second assessment result. The disease assessment guidance information includes the abnormal range corresponding to each abnormal brain region and the target abnormal degree of each abnormal brain region. The specific generation process will be described in detail later.

[0080] Based on the above scheme, the brain glucose metabolism map is synthesized into a brain amyloid protein deposition map through an image synthesis network, avoiding the high cost and low efficiency of Aβ tracers and cyclotrons. Then, the user's disease information is comprehensively evaluated and guided from two perspectives: the brain imaging map corresponding to the brain glucose metabolism map and the brain amyloid protein deposition map, thereby generating the user's disease assessment guidance information without the need for multiple inspections, and ensuring the accuracy and reliability of the generated disease assessment guidance information for medical staff, thereby comprehensively improving the efficiency of AD disease assessment guidance.

[0081] Optionally, before obtaining the user's brain glucose metabolism map, the method further includes: obtaining sample brain glucose metabolism maps of multiple users and sample brain amyloid protein deposition maps of each user, and registering each sample brain glucose metabolism map and each sample brain amyloid protein deposition map through an image fusion template to obtain a first brain image map corresponding to each sample brain glucose metabolism map and a second brain image map corresponding to each sample brain amyloid protein deposition map; based on each first brain image map and each second brain image map, generating a synthetic second brain image map corresponding to each first brain image map and a synthetic first brain image map corresponding to each second brain image map through an initial image synthesis network; inputting each first brain image map and each synthetic first brain image map into a first diagnostic network to obtain first synthetic deviation information of the initial image synthesis network, and inputting each second brain image map and each synthetic first brain image map into a first diagnostic network to obtain first synthetic deviation information of the initial image synthesis network; The second brain image map is input into the second diagnostic network to obtain the second synthesis deviation information of the initial image synthesis network; when the first synthesis deviation information is greater than the first deviation threshold, or the second synthesis deviation information is greater than the second deviation threshold, the initial image synthesis network is adjusted based on each first synthesis deviation information and each second synthesis deviation information to obtain a new image synthesis network; the new image synthesis network is used to replace the initial image synthesis network, and the step of generating a synthetic second brain image map corresponding to each first brain image map and a synthetic first brain image map corresponding to each second brain image map through the initial image synthesis network based on each first brain image map and each second brain image map is returned to execute, until all the first synthesis deviation information are not greater than the first deviation threshold, and all the second synthesis deviation information are not greater than the second deviation threshold, and the initial image synthesis network obtained by the last iteration is used as the image synthesis network.

[0082] In this embodiment, the terminal obtains sample brain glucose metabolism maps of multiple users and sample brain amyloid protein deposition maps of each user, and performs registration processing on each sample brain glucose metabolism map and each sample brain amyloid protein deposition map through an image fusion template to obtain a first brain image map corresponding to each sample brain glucose metabolism map and a second brain image map corresponding to each sample brain amyloid protein deposition map. Specifically, the terminal uses the syn registration method to register individual 18 The F-FDG PET image data and Aβ PET image data were registered with the standard PET template to achieve spatial standardization (spacing = 1mm*1mm*1mm, resolution 181*217*181). Figure 4 As shown in the figure, the brain parenchyma mask of the standard meningeal version is used in combination with a simple threshold method to remove the skull and background to achieve brain parenchyma extraction; at the same time, for the brain parenchyma data within the mask, the 99.7% quantile and 0 in the non-negative value are used as the upper and lower thresholds, respectively, and the threshold is used to replace the voxels that exceed the threshold.

[0083] Then, based on each first brain image map and each second brain image map, the terminal generates a synthesized second brain image map corresponding to each first brain image map and a synthesized first brain image map corresponding to each second brain image map through the initial image synthesis network. Then, the terminal inputs each first brain image map and each synthesized first brain image map into the first diagnostic network to obtain the first synthesis deviation information of the initial image synthesis network, and inputs each second brain image map and each synthesized second brain image map into the second diagnostic network to obtain the second synthesis deviation information of the initial image synthesis network. Among them, the first diagnostic network and the second diagnostic network are neural networks based on a discriminant model, and the discriminant model is composed of a Conv3D module (convolution kernel size = 7, step size = 2, padding = 3), a 3D maximum pooling layer (convolution kernel size = 3, step size = 2, padding = 1), followed by 4 dense blocks and 3 transition blocks between them. The 4 dense blocks contain 3, 6, 12 and 8 dense layers (Dense block) respectively. The growth rate (k) in the dense block is 12, and the compression factor (θ) in the transition block is 0.6. The training parameters of the identification model are as follows, and the loss function uses the cross-entropy loss. The model is optimized using the Adam optimizer with the optimization parameters: β1 = 0.5, β2 = 0.999. The weight initialization uses the Kaiming initialization strategy. The training was performed for a total of 200 epochs, and the learning rate was initially set to 1 × 10^-4 and decreased every 10 epochs with a decrease factor of 2, and the minimum learning rate was 1 × 10^-6.

[0084] The terminal presets a first deviation threshold and a second deviation threshold, and when the first synthesis deviation information is greater than the first deviation threshold, or the second synthesis deviation information is greater than the second deviation threshold, the initial image synthesis network is adjusted based on each first synthesis deviation information and each second synthesis deviation information to obtain a new image synthesis network. The specific process of identifying the deviation information will be described in detail later, and the method of adjusting the initial image synthesis network based on the deviation information is:

[0085] The network input is a 3-layer 181×181 FDG image, and the output is the corresponding ABeta image. The network structure consists of two parts: encoder and decoder. The encoder consists of 4 downsampling blocks, each of which contains two 3×3 convolutional layers, batch normalization layers and ReLU activation functions, and is downsampled by 2 times through maximum pooling. The decoder contains 4 upsampling blocks, each of which uses transposed convolution to perform 2 times upsampling, and then jumps to the feature map of the corresponding layer of the encoder. After the connection, it passes through two 3×3 convolutional layers, batch normalization layers and ReLU activation functions. The last layer of the network uses 1×1 convolution to map the number of channels to the target number of channels, and the output layer uses Sigmoid activation function. The training uses the Adam optimizer, with an initial learning rate of 0.001, which decays to the original 0.1 every 50 epochs. The loss function uses mean square error loss (MSE Loss). To prevent overfitting, dropout (ratio 0.5) and L2 regularization (weight 1e-4) are used. During the training process, the validation set is used to monitor the model performance, and the early stopping strategy is adopted to avoid overfitting.

[0086] Afterwards, the summary sheet replaces the initial image synthesis network with the new image synthesis network, and returns to execute the steps of generating, based on each first brain image map and each second brain image map, a synthesized second brain image map corresponding to each first brain image map, and a synthesized first brain image map corresponding to each second brain image map through the initial image synthesis network, until all first synthesis deviation information is not greater than the first deviation threshold, and all second synthesis deviation information is not greater than the second deviation threshold, and the initial image synthesis network obtained in the last iteration is used as the image synthesis network.

[0087] Based on the above scheme, by first using a set of paired 18 A modified 3D-UNet model was trained on F-FDG PET and Aβ PET images to generate pseudo-paired PET images. The model was adjusted by the Adaptive Instance Normalization (AdaIN) module to achieve a bidirectional mapping between the two modalities during the training phase (from 18 F-FDG PET to Aβ PET and vice versa 18F-FDG PET). For this bidirectional mapping, two discriminators based on PatchGAN are designed, namely 18 The F-FDG PET discriminator and the Aβ PET discriminator have the same network structure. The generated outputs (synthesized PET images and original PET images) are input into the diagnostic network, and the performance of the synthesized PET images is improved by back propagation of the classification loss, thereby generating images with higher positive and negative discrimination, thereby improving the accuracy of image synthesis between the two modalities.

[0088] Optionally, each first brain image map and each synthesized first brain image map are input into a first diagnostic network to obtain first synthesis deviation information of the initial image synthesis network, and each second brain image map and each synthesized second brain image map are input into a second diagnostic network to obtain second synthesis deviation information of the initial image synthesis network, including: inputting each first brain image map and each synthesized first brain image map into the first diagnostic network, identifying metabolic deviation information between each first brain image map and each synthesized first brain image map, and based on each metabolic deviation information, identifying the abnormal range of synthetic metabolic deviation of the initial image synthesis network and the initial image synthesis network. The method comprises the following steps: inputting each second brain image map and each synthesized second brain image map into the second diagnostic network, identifying the deposition deviation value between each second brain image map and each synthesized second brain image map, and based on each deposition deviation value, identifying the abnormal range of the synthesis deposition deviation of the initial image synthesis network and the synthetic deposition sensitivity value of the initial image synthesis network; using the abnormal range of the synthesis metabolism deviation and the synthetic metabolism sensitivity value as the first synthesis deviation information of the initial image synthesis network, and using the abnormal range of the synthesis deposition deviation and the synthetic deposition sensitivity value as the second synthesis deviation information of the initial image synthesis network.

[0089] In this embodiment, the terminal inputs each first brain image map and each synthesized first brain image map into the first diagnostic network, identifies the metabolic deviation information between each first brain image map and each synthesized first brain image map, and identifies the abnormal range of the synthetic metabolic deviation of the initial image synthesis network and the synthetic metabolic sensitivity value of the initial image synthesis network based on each metabolic deviation information. The abnormal deviation range is the image range where the image deviation exists, and the synthetic metabolic sensitivity value is calculated by dividing the image range by the ratio value obtained by all ranges of the first image map.

[0090] Then, the terminal inputs each second brain image map and each synthesized second brain image map into the second diagnostic network, identifies the deposition deviation value between each second brain image map and each synthesized second brain image map, and identifies the synthetic deposition deviation abnormal range of the initial image synthesis network and the synthetic deposition sensitivity value of the initial image synthesis network based on each deposition deviation value. The synthetic deposition sensitivity value is calculated by dividing the image range by the ratio value obtained by all ranges of the second brain image map.

[0091] Afterwards, the terminal uses the abnormal range of synthetic metabolic deviation and the synthetic metabolic sensitivity value as the first synthetic deviation information of the initial image synthesis network, and uses the abnormal range of synthetic deposition deviation and the synthetic deposition sensitivity value as the second synthetic deviation information of the initial image synthesis network.

[0092] Based on the above scheme, the deviation information between the image and the synthetic image is identified through the diagnostic network corresponding to the divided image, thereby improving the accuracy of identifying the deviation information.

[0093] Optionally, based on the brain glucose metabolism map, registration processing is performed through an image fusion template to obtain a brain image map corresponding to the brain glucose metabolism map, including: extracting the user's brain parenchyma image data from the brain glucose metabolism map through a threshold screening strategy based on the brain parenchyma mask information of the image fusion template; identifying the voxel value of each sub-brain parenchyma image data in the brain parenchyma image data, and adjusting each sub-brain parenchyma image data through a preset voxel value range to obtain each target sub-brain parenchyma image data; using the image data corresponding to all target sub-brain parenchyma image data as the brain image map corresponding to the brain glucose metabolism map.

[0094] In this embodiment, the terminal extracts the user's brain parenchyma image data from the brain glucose metabolism map based on the brain parenchyma mask information of the image fusion template through a threshold screening strategy. Specifically, the terminal uses the syn registration method to 18 The F-FDG PET image data were registered with the standard PET template to achieve spatial standardization (spacing = 1 mm * 1 mm * 1 mm, resolution 181 * 217 * 181). The brain parenchyma mask of the standard meningeal version was used in combination with a simple threshold method to remove the skull and background to achieve brain parenchyma extraction.

[0095] Then, the terminal identifies the voxel value of each sub-brain parenchyma image data in the brain parenchyma image data, and adjusts each sub-brain parenchyma image data through a preset voxel value range to obtain each target sub-brain parenchyma image data. Specifically, the terminal uses the 99.7% quantile and 0 in the non-negative value as the upper and lower thresholds for the voxel value of each sub-brain parenchyma image data in the parenchyma image data, and replaces the voxel value of the sub-brain parenchyma image data that exceeds the threshold with the threshold to obtain each target sub-brain parenchyma image data.

[0096] Finally, the terminal uses the image data corresponding to all target sub-brain parenchyma image data as the brain image map corresponding to the brain glucose metabolism map.

[0097] Based on the above scheme, by extracting brain images of brain parenchyma excluding the skull and background, the influence of the skull and background on the evaluation can be avoided when actually analyzing the deviation information and evaluating the evaluation, thereby improving the accuracy of the evaluation.

[0098] Optionally, based on the brain image map and the brain amyloid protein deposition map, an evaluation model is used to generate a first evaluation result corresponding to the brain image map and a second evaluation result corresponding to the brain amyloid protein deposition map, including: based on the brain image map, identifying the user's metabolic distribution information, and based on the metabolic distribution information, identifying each abnormal metabolic range in the metabolic distribution information and the abnormal metabolic information corresponding to each abnormal metabolic range through the metabolic abnormality recognition network of the evaluation model; using each abnormal metabolic range and each abnormal metabolic information as the first evaluation result corresponding to the brain image map; based on the brain amyloid protein deposition map, identifying the user's protein deposition distribution information, and based on the protein deposition distribution information, identifying each abnormal deposition range in the protein deposition distribution information and the abnormal deposition information corresponding to the abnormal deposition range through the deposition abnormality recognition network of the evaluation model; using each abnormal deposition range and each abnormal deposition information as the second evaluation result corresponding to the brain amyloid protein deposition map.

[0099] In this embodiment, the terminal identifies the user's metabolic distribution information based on the brain image, and based on the metabolic distribution information, identifies each abnormal metabolic range in the metabolic distribution information and the abnormal metabolic information corresponding to each abnormal metabolic range through the metabolic abnormality recognition network of the evaluation model. The metabolic abnormality recognition network is used to identify the metabolic value of each position point in the brain image, and takes the range corresponding to the metabolic value greater than the preset range of the terminal as the abnormal metabolic range, wherein the abnormal metabolic information corresponding to each abnormal metabolic range is the metabolic value distribution information in the abnormal metabolic range.

[0100] Then, the terminal uses each abnormal metabolic range and each abnormal metabolic information as a first evaluation result corresponding to the brain image.

[0101] The terminal identifies the user's protein deposition distribution information based on the brain amyloid deposition map, and based on the protein deposition distribution information, identifies each abnormal deposition range in the protein deposition distribution information and the abnormal deposition information corresponding to the abnormal deposition range through the deposition abnormality recognition network of the evaluation model. Among them, the deposition abnormality recognition network has the same network structure as the metabolic abnormality recognition network. The difference between the two networks is that the deposition abnormality recognition network is a network trained by a brain amyloid deposition map with a large number of labeled protein deposition values, while the metabolic abnormality recognition network is a network trained by a brain glucose metabolism map with a large number of labeled metabolic values. The abnormality recognition process of the brain amyloid deposition map is the same as the abnormality recognition process of the brain imaging map.

[0102] Finally, the terminal uses each abnormal deposition range and each abnormal deposition information as the second evaluation result corresponding to the brain amyloid protein deposition map.

[0103] Based on the above solution, abnormal information of different images is identified through different networks, which improves the accuracy of recognition.

[0104] Optionally, based on the first assessment result and the second assessment result, disease assessment guidance information for the user is generated, including: identifying the first brain region corresponding to each abnormal metabolic range, and identifying the degree of metabolic abnormality of each first brain region based on the abnormal metabolic information of each abnormal metabolic range; identifying the second brain region corresponding to each abnormal deposition range, and identifying the degree of deposition abnormality of each second brain region based on the abnormal deposition information corresponding to each abnormal deposition range; identifying each abnormal brain region of the user and a target abnormal degree of each abnormal brain region based on the degree of metabolic abnormality of each first brain region and the degree of deposition abnormality of each second brain region, and generating the user's disease assessment guidance information through a preset disease assessment guidance template based on the abnormal range corresponding to each abnormal brain region and the target abnormal degree of each abnormal brain region.

[0105] In this embodiment, the terminal identifies the first brain region corresponding to each abnormal metabolic range, and identifies the metabolic abnormality degree of each first brain region based on the abnormal metabolic information of each abnormal metabolic range. The first brain region is the brain region involved in the abnormal metabolic range, and the brain region includes but is not limited to the brainstem region, the cerebellum region, the intercerebral region, and the telencephalon region. The method of identifying the metabolic abnormality degree of the first brain region through abnormal metabolic information is to identify the distribution range of different metabolic values ​​based on the metabolic value distribution information of the first brain region, and the range of the first brain region. Then, the terminal presets the abnormality degree corresponding to the distribution range of each metabolic value of different brain regions. Then, the terminal performs weighted summation processing on the abnormality degree corresponding to the distribution range of each metabolic value of the first brain region according to the preset weight value of each brain region, and obtains the metabolic abnormality degree of the first brain region.

[0106] Then, the terminal identifies the second brain region corresponding to each abnormal deposition range, and identifies the abnormal deposition degree of each second brain region based on the abnormal deposition information corresponding to each abnormal deposition range. The method for identifying the abnormal deposition degree is the same as the method for identifying the metabolic abnormality degree.

[0107] Afterwards, the terminal identifies each abnormal brain region of the user and the target abnormal degree of each abnormal brain region based on the abnormal degree of metabolism of each first brain region and the abnormal degree of deposition of each second brain region. Specifically, when the brain region is the first brain region or the second brain region, the abnormal degree value of the brain region is the abnormal degree of metabolism of the first brain region or the abnormal degree of deposition of the second brain region. When the brain region is the first brain region and the second brain region, the abnormal degree value of the brain region is the weighted sum of the abnormal degree of metabolism of the first brain region and the abnormal degree of deposition of the second brain region, wherein the weight value corresponding to the abnormal degree of metabolism of the first brain region and the weight value corresponding to the abnormal degree of deposition are weight values ​​preset in the terminal by the staff.

[0108] Finally, the terminal generates disease assessment guidance information for the user based on the abnormal range corresponding to each abnormal brain area and the target abnormal degree of each abnormal brain area through a preset disease assessment guidance template. The disease assessment guidance template is a template preset in the terminal by the staff, and the template can change with the staff's adjustment.

[0109] This application also provides an example of AD disease assessment guidelines generated based on brain imaging, such as Figure 5 As shown, the specific processing process includes the following steps:

[0110] Step S501, obtaining a brain glucose metabolism map of the user.

[0111] Step S502 , based on the brain parenchyma mask information of the image fusion template, a threshold screening strategy is used to extract the user's brain parenchyma image data from the brain glucose metabolism map.

[0112] Step S503, identifying the voxel value of each sub-brain parenchyma image data in the brain parenchyma image data, and adjusting each sub-brain parenchyma image data according to a preset voxel value range to obtain each target sub-brain parenchyma image data.

[0113] Step S504: The image data corresponding to all target sub-brain parenchyma image data are used as the brain image map corresponding to the brain glucose metabolism map.

[0114] Step S505, based on the brain image, synthesize the user's brain amyloid protein deposition map through the image synthesis network.

[0115] Step S506, based on the brain image, identifying the user's metabolic distribution information, and based on the metabolic distribution information, identifying each abnormal metabolic range in the metabolic distribution information and the abnormal metabolic information corresponding to each abnormal metabolic range through the metabolic abnormality recognition network of the evaluation model.

[0116] Step S507: taking each abnormal metabolic range and each abnormal metabolic information as a first evaluation result corresponding to the brain image.

[0117] Step S508, based on the brain amyloid protein deposition map, identify the user's protein deposition distribution information, and based on the protein deposition distribution information, identify each abnormal deposition range in the protein deposition distribution information and the abnormal deposition information corresponding to the abnormal deposition range through the deposition abnormality recognition network of the evaluation model.

[0118] Step S509: taking each abnormal deposition range and each abnormal deposition information as a second evaluation result corresponding to the brain amyloid protein deposition map.

[0119] Step S510, identifying the first brain regions corresponding to each abnormal metabolic range, and identifying the degree of metabolic abnormality of each first brain region based on the abnormal metabolic information of each abnormal metabolic range.

[0120] Step S511, identifying the second brain regions corresponding to each abnormal deposition range, and identifying the abnormal deposition degree of each second brain region based on the abnormal deposition information corresponding to each abnormal deposition range.

[0121] Step S512, based on the metabolic abnormality level of each first brain area and the deposition abnormality level of each second brain area, identify the user's abnormal brain areas and the target abnormality level of each abnormal brain area, and based on the abnormal range corresponding to each abnormal brain area and the target abnormality level of each abnormal brain area, generate the user's disease assessment guidance information through a preset disease assessment guidance template.

[0122] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0123] Based on the same inventive concept, the embodiment of the present application also provides a brain image-based AD disease assessment guidance device for implementing the above-mentioned brain image-based AD disease assessment guidance method. The implementation solution provided by the device to solve the problem is similar to the implementation solution recorded in the above-mentioned method, so the specific limitations in one or more brain image-based AD disease assessment guidance device embodiments provided below can refer to the above limitations on the brain image-based AD disease assessment guidance method, which will not be repeated here.

[0124] In an exemplary embodiment, Figure 6 As shown, an AD disease assessment guidance device based on brain image generation is provided, comprising: an acquisition module 610, a synthesis module 620 and a generation module 630, wherein:

[0125] An acquisition module 610 is used to acquire a brain glucose metabolism map of a user, and based on the brain glucose metabolism map, perform registration processing through an image fusion template to obtain a brain image map corresponding to the brain glucose metabolism map;

[0126] A synthesis module 620 is used to synthesize the user's brain amyloid protein deposition map based on the brain image map through an image synthesis network, and generate a first evaluation result corresponding to the brain image map and a second evaluation result corresponding to the brain amyloid protein deposition map through an evaluation model based on the brain image map and the brain amyloid protein deposition map;

[0127] The generation module 630 is used to generate disease assessment guidance information for the user based on the first assessment result and the second assessment result.

[0128] Optionally, the device further comprises:

[0129] A registration module is used to obtain sample brain glucose metabolism maps of multiple users and sample brain amyloid protein deposition maps of each user, and to perform registration processing on each sample brain glucose metabolism map and each sample brain amyloid protein deposition map respectively through the image fusion template to obtain a first brain image map corresponding to each sample brain glucose metabolism map and a second brain image map corresponding to each sample brain amyloid protein deposition map;

[0130] An image synthesis module, configured to generate, based on each of the first brain image images and each of the second brain image images, a synthesized second brain image image corresponding to each of the first brain image images and a synthesized first brain image image corresponding to each of the second brain image images through an initial image synthesis network;

[0131] a deviation identification module, for inputting each of the first brain image maps and each of the synthesized first brain image maps into a first diagnostic network to obtain first synthesis deviation information of the initial image synthesis network, and for inputting each of the second brain image maps and each of the synthesized second brain image maps into a second diagnostic network to obtain second synthesis deviation information of the initial image synthesis network;

[0132] An adjustment module, configured to adjust the initial image synthesis network based on each of the first synthesis deviation information and each of the second synthesis deviation information to obtain a new image synthesis network when the first synthesis deviation information is greater than the first deviation threshold or the second synthesis deviation information is greater than the second deviation threshold;

[0133] An iteration module is used to replace the initial image synthesis network with the new image synthesis network, and return to execute the step of generating a synthesized second brain image corresponding to each of the first brain image images and a synthesized first brain image corresponding to each of the second brain image images through the initial image synthesis network based on each of the first brain image images and each of the second brain image images, until all the first synthesis deviation information is not greater than the first deviation threshold and all the second synthesis deviation information is not greater than the second deviation threshold, and the initial image synthesis network obtained by the last iteration is used as the image synthesis network.

[0134] Optionally, the deviation identification module is specifically used to:

[0135] Inputting each of the first brain image maps and each of the synthesized first brain image maps into a first diagnostic network, identifying metabolic deviation information between each of the first brain image maps and each of the synthesized first brain image maps, and based on each of the metabolic deviation information, identifying anabolic metabolic deviation abnormal range of the initial image synthesis network and anabolic metabolic sensitivity value of the initial image synthesis network;

[0136] Inputting each of the second brain image maps and each of the synthesized second brain image maps into a second diagnostic network, identifying a deposition deviation value between each of the second brain image maps and each of the synthesized second brain image maps, and based on each of the deposition deviation values, identifying a synthetic deposition deviation abnormal range of the initial image synthesis network and a synthetic deposition sensitivity value of the initial image synthesis network;

[0137] The anabolic deviation abnormal range and the anabolic sensitivity value are used as first synthetic deviation information of the initial image synthesis network, and the synthetic deposition deviation abnormal range and the synthetic deposition sensitivity value are used as second synthetic deviation information of the initial image synthesis network.

[0138] Optionally, the acquisition module 610 is specifically configured to:

[0139] Based on the brain parenchyma mask information of the image fusion template, extracting the brain parenchyma image data of the user in the brain glucose metabolism map through a threshold screening strategy;

[0140] Identifying the voxel value of each sub-brain parenchyma image data in the brain parenchyma image data, and adjusting each of the sub-brain parenchyma image data according to a preset voxel value range to obtain each target sub-brain parenchyma image data;

[0141] The image data corresponding to all target sub-brain parenchyma image data are used as the brain image map corresponding to the brain glucose metabolism map.

[0142] Optionally, the synthesis module 620 is specifically used for:

[0143] Based on the brain image, identifying the metabolic distribution information of the user, and based on the metabolic distribution information, identifying each abnormal metabolic range in the metabolic distribution information and the abnormal metabolic information corresponding to each abnormal metabolic range through the metabolic abnormality identification network of the evaluation model;

[0144] Using each of the abnormal metabolic ranges and each of the abnormal metabolic information as a first evaluation result corresponding to the brain image;

[0145] Based on the brain amyloid protein deposition map, protein deposition distribution information of the user is identified, and based on the protein deposition distribution information, each abnormal deposition range in the protein deposition distribution information and abnormal deposition information corresponding to the abnormal deposition range are identified through the deposition abnormality identification network of the evaluation model;

[0146] Each of the abnormal deposition ranges and each of the abnormal deposition information is used as a second evaluation result corresponding to the brain amyloid protein deposition map.

[0147] Optionally, the generating module 630 is specifically configured to:

[0148] identifying a first brain region corresponding to each of the abnormal metabolic ranges, and identifying a degree of metabolic abnormality of each of the first brain regions based on abnormal metabolic information of each of the abnormal metabolic ranges;

[0149] Identifying a second brain region corresponding to each of the abnormal deposition ranges, and identifying a deposition abnormality degree of each of the second brain regions based on abnormal deposition information corresponding to each of the abnormal deposition ranges;

[0150] Based on the metabolic abnormality level of each of the first brain regions and the deposition abnormality level of each of the second brain regions, the abnormal brain regions of the user and the target abnormality level of each of the abnormal brain regions are identified, and based on the abnormal range corresponding to each of the abnormal brain regions and the target abnormality level of each of the abnormal brain regions, disease assessment guidance information for the user is generated through a preset disease assessment guidance template.

[0151] Each module in the above-mentioned AD disease assessment guidance device generated based on brain images can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0152] In an exemplary embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as shown in FIG. Figure 7As shown. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and the external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be realized through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a method for guiding AD disease assessment based on brain image generation is implemented. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.

[0153] Those skilled in the art will understand that Figure 7 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0154] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, steps corresponding to a method for AD disease assessment guidance generated based on brain images are implemented.

[0155] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps corresponding to the AD disease assessment guidance method generated based on brain images are implemented.

[0156] In one embodiment, a computer program product is provided, including a computer program, which, when executed by a processor, implements the steps corresponding to the AD disease assessment guidance method generated based on brain images.

[0157] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0158] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0159] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0160] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for AD disease assessment guidance based on brain image generation, characterized in that: The method comprises: Acquire a user's brain glucose metabolism map, and perform registration processing based on the brain glucose metabolism map through an image fusion template to obtain a brain image map corresponding to the brain glucose metabolism map; Based on the brain image, synthesize the user's brain amyloid protein deposition map through an image synthesis network, and based on the brain image and the brain amyloid protein deposition map, generate a first evaluation result corresponding to the brain image and a second evaluation result corresponding to the brain amyloid protein deposition map through an evaluation model; Based on the first assessment result and the second assessment result, disease assessment guidance information for the user is generated.

2. The method according to claim 1, characterized in that Before obtaining the user's brain glucose metabolism map, the method further includes: Acquire sample brain glucose metabolism maps of multiple users and sample brain amyloid protein deposition maps of each user, and perform registration processing on each sample brain glucose metabolism map and each sample brain amyloid protein deposition map respectively through the image fusion template to obtain a first brain image map corresponding to each sample brain glucose metabolism map and a second brain image map corresponding to each sample brain amyloid protein deposition map; Based on each of the first brain image maps and each of the second brain image maps, generating a synthesized second brain image map corresponding to each of the first brain image maps and a synthesized first brain image map corresponding to each of the second brain image maps through an initial image synthesis network; Inputting each of the first brain image maps and each of the synthesized first brain image maps into a first diagnostic network to obtain first synthesis deviation information of the initial image synthesis network, and inputting each of the second brain image maps and each of the synthesized second brain image maps into a second diagnostic network to obtain second synthesis deviation information of the initial image synthesis network; When the first synthesis deviation information is greater than the first deviation threshold, or the second synthesis deviation information is greater than the second deviation threshold, adjusting the initial image synthesis network based on each of the first synthesis deviation information and each of the second synthesis deviation information to obtain a new image synthesis network; The new image synthesis network replaces the initial image synthesis network, and the step of generating a synthetic second brain image corresponding to each of the first brain image images and a synthetic first brain image corresponding to each of the second brain image images through the initial image synthesis network based on each of the first brain image images and each of the second brain image images is returned to the step of generating a synthetic second brain image corresponding to each of the first brain image images and a synthetic first brain image corresponding to each of the second brain image images through the initial image synthesis network, until all the first synthesis deviation information is not greater than the first deviation threshold and all the second synthesis deviation information is not greater than the second deviation threshold, and the initial image synthesis network obtained by the last iteration is used as the image synthesis network.

3. The method according to claim 2, characterized in that The step of inputting each of the first brain image maps and each of the synthesized first brain image maps into a first diagnostic network to obtain first synthesis deviation information of the initial image synthesis network, and inputting each of the second brain image maps and each of the synthesized second brain image maps into a second diagnostic network to obtain second synthesis deviation information of the initial image synthesis network comprises: Inputting each of the first brain image maps and each of the synthesized first brain image maps into a first diagnostic network, identifying metabolic deviation information between each of the first brain image maps and each of the synthesized first brain image maps, and based on each of the metabolic deviation information, identifying anabolic metabolic deviation abnormal range of the initial image synthesis network and anabolic metabolic sensitivity value of the initial image synthesis network; Inputting each of the second brain image maps and each of the synthesized second brain image maps into a second diagnostic network, identifying a deposition deviation value between each of the second brain image maps and each of the synthesized second brain image maps, and based on each of the deposition deviation values, identifying a synthetic deposition deviation abnormal range of the initial image synthesis network and a synthetic deposition sensitivity value of the initial image synthesis network; The anabolic deviation abnormal range and the anabolic sensitivity value are used as first synthetic deviation information of the initial image synthesis network, and the synthetic deposition deviation abnormal range and the synthetic deposition sensitivity value are used as second synthetic deviation information of the initial image synthesis network.

4. The method according to claim 1, characterized in that The method of performing registration processing based on the brain glucose metabolism map by using an image fusion template to obtain a brain image map corresponding to the brain glucose metabolism map includes: Based on the brain parenchyma mask information of the image fusion template, extracting the brain parenchyma image data of the user in the brain glucose metabolism map through a threshold screening strategy; Identifying the voxel value of each sub-brain parenchyma image data in the brain parenchyma image data, and adjusting each of the sub-brain parenchyma image data according to a preset voxel value range to obtain each target sub-brain parenchyma image data; The image data corresponding to all target sub-brain parenchyma image data are used as the brain image map corresponding to the brain glucose metabolism map.

5. The method according to claim 1, characterized in that The step of generating a first evaluation result corresponding to the brain image and a second evaluation result corresponding to the brain amyloid protein deposition map by using an evaluation model based on the brain image and the brain amyloid protein deposition map comprises: Based on the brain image, identifying the metabolic distribution information of the user, and based on the metabolic distribution information, identifying each abnormal metabolic range in the metabolic distribution information and the abnormal metabolic information corresponding to each abnormal metabolic range through the metabolic abnormality identification network of the evaluation model; Using each of the abnormal metabolic ranges and each of the abnormal metabolic information as a first evaluation result corresponding to the brain image; Based on the brain amyloid protein deposition map, identifying the amyloid protein deposition distribution information of the user, and based on the amyloid protein deposition distribution information, identifying each abnormal deposition range in the amyloid protein deposition distribution information and the abnormal deposition information corresponding to the abnormal deposition range through the deposition abnormality identification network of the evaluation model; Each of the abnormal deposition ranges and each of the abnormal deposition information is used as a second evaluation result corresponding to the brain amyloid protein deposition map.

6. The method according to claim 5, characterized in that The generating disease assessment guidance information of the user based on the first assessment result and the second assessment result includes: identifying a first brain region corresponding to each of the abnormal metabolic ranges, and identifying a degree of metabolic abnormality of each of the first brain regions based on abnormal metabolic information of each of the abnormal metabolic ranges; Identifying a second brain region corresponding to each of the abnormal deposition ranges, and identifying a deposition abnormality degree of each of the second brain regions based on abnormal deposition information corresponding to each of the abnormal deposition ranges; Based on the metabolic abnormality level of each of the first brain regions and the deposition abnormality level of each of the second brain regions, the abnormal brain regions of the user and the target abnormality level of each of the abnormal brain regions are identified, and based on the abnormal range corresponding to each of the abnormal brain regions and the target abnormality level of each of the abnormal brain regions, disease assessment guidance information for the user is generated through a preset disease assessment guidance template.

7. An AD disease assessment guidance device based on brain image generation, characterized in that: The device comprises: An acquisition module is used to acquire a user's brain glucose metabolism map, and based on the brain glucose metabolism map, perform registration processing through an image fusion template to obtain a brain image map corresponding to the brain glucose metabolism map; a synthesis module, configured to synthesize a brain amyloid protein deposition map of the user based on the brain image map through an image synthesis network, and to generate a first evaluation result corresponding to the brain image map and a second evaluation result corresponding to the brain amyloid protein deposition map through an evaluation model based on the brain image map and the brain amyloid protein deposition map; A generation module is used to generate disease assessment guidance information for the user based on the first assessment result and the second assessment result.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images