Image processing method, device, computer equipment and storage medium

Abstract

The present application relates to an image processing method, device, computer equipment and storage medium. The method comprises: obtaining a first medical image; inputting the first medical image into a preset processing model, and obtaining a second medical image through the processing model; wherein the contrast of the second medical image is greater than the contrast of the first medical image; the processing model is obtained by training based on a first sample medical image and a gold standard image corresponding to the first sample medical image, and the contrast of the gold standard image is greater than the contrast of the first sample medical image. The method can reduce the dose of contrast agent required for medical imaging and reduce the possible harm of contrast agent to human body while ensuring the imaging quality of medical images.

Description

Technical Field

[0001] The present application relates to the field of medical imaging technology, and in particular to an image processing method, apparatus, computer equipment and storage medium. Background Art

[0002] Medical imaging technology can display the internal tissues of the human body. For example, Magnetic Resonance Imaging (MRI) technology can safely and accurately image multiple parts of the human body. Medical imaging technology has become an important imaging examination method and is currently widely used in clinical diagnosis and scientific research in hospitals.

[0003] However, medical imaging technology cannot fully display some specific tissues or smaller lesions (such as tumors, blood vessels, etc.), making it difficult for doctors to distinguish whether human tissues are lesions. Therefore, it is particularly important to improve the quality of medical images. In traditional technology, patients are usually injected or orally given contrast agents before clinical scanning to improve the imaging contrast between normal and lesion areas by changing the relaxation rate of water protons in local tissues in the patient's body, thereby improving the quality of medical images.

[0004] However, contrast agents injected or taken orally before scanning may pose certain hazards to human health and safety, and the contrast agent dosage needs to meet certain standards. If the contrast agent dosage does not meet the standard, it will affect the contrast of medical images. Summary of the invention

[0005] Based on this, it is necessary to provide an image processing method, device, computer equipment and storage medium that can ensure the contrast of medical images when improving the quality of medical images in response to the above technical problems.

[0006] An image processing method, the method comprising:

[0007] acquiring a first medical image;

[0008] The first medical image is input into a preset processing model, and a second medical image is obtained through the processing model; wherein the contrast of the second medical image is greater than the contrast of the first medical image; the processing model is obtained by training based on a first sample medical image and a gold standard image corresponding to the first sample medical image, and the contrast of the gold standard image is greater than the contrast of the first sample medical image.

[0009] In one embodiment, the processing model includes a feature extraction network, a feature learning network and an image reconstruction network, and the first medical image is input into a preset processing model, and a second medical image is obtained through the processing model, including:

[0010] Inputting the first medical image into the feature extraction network, performing feature extraction on the first medical image through the feature extraction network, and obtaining features of the first medical image;

[0011] Inputting the features of the first medical image into the feature learning network, and learning the features of the first medical image through the feature learning network to obtain a feature map; the feature map is a feature map containing the relationship between each pixel of the first medical image;

[0012] The feature map is input into the image reconstruction network, and the second medical image is obtained through the image reconstruction network.

[0013] In one of the embodiments, the feature learning network includes multiple feature learning sub-networks, each of the feature learning sub-networks includes at least three residual structure and attention mechanism combination modules, and each of the residual structure and attention mechanism combination modules is densely connected.

[0014] In one of the embodiments, each of the residual structure and attention mechanism combination modules includes a residual module, a spatial attention module and a channel attention module.

[0015] In one embodiment, the training process of the processing model includes:

[0016] Acquire a first sample medical image and a gold standard image corresponding to the first sample medical image;

[0017] Obtaining a second sample medical image according to the first sample medical image and a preset initial processing model; the contrast of the second sample medical image is greater than the contrast of the first sample medical image;

[0018] Obtaining a value of a loss function of the initial processing model according to the second sample medical image and the gold standard image;

[0019] According to the value of the loss function of the initial processing model, the parameters of the initial processing model are adjusted to obtain the processing model.

[0020] In one embodiment, the second sample medical image includes a plurality of images, and obtaining the value of the loss function of the initial processing model according to the second sample medical image and the gold standard image includes:

[0021] Obtaining a value of a loss function between each of the second sample medical images and the gold standard image;

[0022] The loss function values ​​of each of the second sample medical images and the gold standard image are summed to obtain the loss function value of the initial processing model.

[0023] In one embodiment, obtaining a second sample medical image according to the first sample medical image and a preset initial processing model includes:

[0024] Obtaining an amplitude map corresponding to the first sample medical image;

[0025] The amplitude map corresponding to the first sample medical image is input into the initial processing model, and the second sample medical image is obtained through the initial processing model.

[0026] An image processing device, comprising:

[0027] A first acquisition module, used for acquiring a first medical image;

[0028] A first processing module is used to input the first medical image into a preset processing model, and obtain a second medical image through the processing model; wherein the contrast of the second medical image is greater than the contrast of the first medical image; the processing model is obtained by training based on the first sample medical image and the gold standard image corresponding to the first sample medical image, and the contrast of the gold standard image is greater than the contrast of the first sample medical image.

[0029] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0030] acquiring a first medical image;

[0031] The first medical image is input into a preset processing model, and a second medical image is obtained through the processing model; wherein the contrast of the second medical image is greater than the contrast of the first medical image; the processing model is obtained by training based on a first sample medical image and a gold standard image corresponding to the first sample medical image, and the contrast of the gold standard image is greater than the contrast of the first sample medical image.

[0032] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the following steps:

[0033] acquiring a first medical image;

[0034] The first medical image is input into a preset processing model, and a second medical image is obtained through the processing model; wherein the contrast of the second medical image is greater than the contrast of the first medical image; the processing model is obtained by training based on a first sample medical image and a gold standard image corresponding to the first sample medical image, and the contrast of the gold standard image is greater than the contrast of the first sample medical image.

[0035] The above-mentioned image processing method, device, computer equipment and storage medium input the acquired first medical image into a preset processing model, and obtain a second medical image through the processing model, wherein the contrast of the obtained second medical image is greater than the contrast of the first medical image; the processing model is obtained by training based on the first sample medical image and the gold standard image corresponding to the first sample medical image, and the contrast of the gold standard image is greater than the contrast of the first sample medical image. Since the processing model is obtained by training a large number of first sample medical images and the gold standard images corresponding to the first sample medical images, a processing model with a relatively stable network structure can be obtained. In addition, the contrast of the gold standard image is greater than the contrast of the first sample medical image. In this way, the first medical image is input into the trained processing model, and the processing model can be used to obtain a second medical image with a contrast greater than the contrast of the first medical image. Under the premise of ensuring the imaging quality of the medical image, the dose of contrast agent required for medical imaging can be reduced, and the possible harm of the contrast agent to the human body can be reduced; in addition, while reducing the dose of contrast agent, a second medical image with a contrast greater than the first medical image can be obtained, and while ensuring the image quality of the obtained medical image, the cost of medical imaging is also reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 A diagram showing an application environment of an image processing method in an embodiment;

[0037] Figure 2 is a schematic flow chart of an image processing method in one embodiment;

[0038] Figure 3 is a schematic flow chart of an image processing method in one embodiment;

[0039] Figure 3a A schematic diagram of the structure of a processing model in one embodiment;

[0040] Figure 3b A schematic diagram of the structure of a feature learning subnetwork in one embodiment;

[0041] Figure 3c A schematic diagram of the structure of a residual structure and attention mechanism combination module in one embodiment;

[0042] Figure 3d is a schematic diagram of the structure of a spatial attention module in one embodiment;

[0043] Figure 3e Schematic diagram of the structure of a channel attention module in one embodiment;

[0044] Figure 4 is a schematic flow chart of an image processing method in one embodiment;

[0045] Figure 5 is a schematic flow chart of an image processing method in one embodiment;

[0046] Figure 6 FIG. 4 is a structural block diagram of an image processing device in an embodiment. DETAILED DESCRIPTION

[0047] 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.

[0048] The image processing method provided in the embodiment of the present application can be applied to Figure 1 The computer device shown. The computer device includes a processor and a memory connected by a system bus, and a computer program is stored in the memory. When the processor executes the computer program, the steps of the following method embodiment can be executed. Optionally, the computer device may also include a network interface, a display screen and an input device. 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, and 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 network interface of the computer device is used to communicate with an external terminal through a network connection. Optionally, the computer device can be a server, a personal computer, a personal digital assistant, or other terminal devices, such as a tablet computer, a mobile phone, etc., or a cloud or remote server. The embodiment of the present application does not limit the specific form of the computer device.

[0049] Take Magnetic Resonance Imaging (MRI) technology as an example. MRI can safely and accurately image multiple parts of the human body. It is an important imaging examination method and is currently widely used in clinical diagnosis and scientific research in hospitals. However, MRI technology cannot fully display some specific tissues or smaller lesions (such as tumors, blood vessels, etc.), making it difficult for doctors to distinguish whether human tissues are lesions. Therefore, it is particularly important to improve the quality of magnetic resonance images. Usually, before clinical MRI scanning, a contrast agent will be injected into the patient to improve the imaging contrast between normal and lesion sites by changing the relaxation rate of water protons in the local tissues of the patient's body, and it can also shorten the imaging time. Common contrast agents include T1 contrast agents, T2 contrast agents, CEST contrast agents, etc. However, contrast agents injected or taken orally before magnetic resonance scanning may cause certain harm to human health and safety, and the contrast dosage needs to meet certain standards, otherwise it will affect the contrast of magnetic resonance imaging and reduce the quality of the obtained magnetic resonance images. Based on this, it is necessary to provide an image processing method that can reduce the contrast dosage injected into the patient while ensuring high contrast between target tissue organs and surrounding tissue organs in medical images.

[0050] In one embodiment, Figure 2 As shown, an image processing method is provided, which is applied to Figure 1 The computer device in the example is used to illustrate, including the following steps:

[0051] S201, acquiring a first medical image.

[0052] The first medical image is obtained by injecting a low dose of contrast agent into the scanned object for scanning. Optionally, the first medical image may be a magnetic resonance imaging (MRI) image, or other medical imaging images. Optionally, the first medical image may be a brain image of the scanned object, or a knee joint image of the scanned object, or a medical image of other parts of the scanned object.

[0053] Specifically, the computer device acquires the first medical image. Optionally, the computer device may acquire the first medical image from a PACS (Picture Archiving and Communication Systems) server, or may acquire the first medical image in real time from a corresponding medical imaging device.

[0054] S202, inputting the first medical image into a preset processing model, and obtaining a second medical image through the processing model; wherein the contrast of the second medical image is greater than the contrast of the first medical image; the processing model is obtained by training based on the first sample medical image and a gold standard image corresponding to the first sample medical image, and the contrast of the gold standard image is greater than the contrast of the first sample medical image.

[0055] Specifically, the computer device inputs the acquired first medical image into a preset processing model, and obtains a second medical image through the processing model; wherein the contrast of the acquired second medical image is greater than the contrast of the acquired first medical image; the preset processing model is obtained by training according to the first sample medical image and the gold standard image corresponding to the first sample medical image, and the contrast of the gold standard image corresponding to the first sample medical image is greater than the contrast of the first sample medical image. Optionally, the contrast of the second medical image refers to the contrast between the target tissue organ and the surrounding tissue organ in the second medical image, and the contrast of the first medical image refers to the contrast between the target tissue organ and the surrounding tissue organ in the first medical image. Accordingly, it can be understood that the contrast of the first sample medical image refers to the contrast between the target tissue organ and the surrounding tissue organ in the first sample medical image, and the contrast of the gold standard image corresponding to the first sample medical image refers to the contrast between the target tissue organ and the surrounding tissue organ in the gold standard image. Optionally, the network structure corresponding to the preset processing model can be a convolutional neural network, a generative adversarial network, or other neural networks. It should be noted that the network structure of the preset processing model provided in this embodiment is a LRDNet (Long range dependency network) convolutional neural network.

[0056] In the above-mentioned image processing method, since the processing model is obtained by training a large number of first sample medical images and gold standard images corresponding to the first sample medical images, a processing model with a relatively stable network structure can be obtained. In addition, the contrast of the gold standard image is greater than the contrast of the first sample medical image. In this way, the first medical image is input into the trained processing model, and a second medical image with a contrast greater than the contrast of the first medical image can be obtained through the processing model. Under the premise of ensuring the imaging quality of the medical image, the dose of contrast agent required for medical imaging can be reduced, and the possible harm of the contrast agent to the human body can be reduced; in addition, while reducing the dose of contrast agent, a second medical image with a contrast greater than the first medical image can be obtained, while ensuring the image quality of the obtained medical image, the cost of medical imaging is also reduced.

[0057] In the above-mentioned scenario where the acquired first medical image is input into a preset processing model and a second medical image is obtained through the processing model, in one embodiment, the above-mentioned processing model is an LRDNet convolutional neural network model, which includes a feature extraction network, a feature learning network and an image reconstruction network, such as Figure 3 As shown, the above S202 includes:

[0058] S301, inputting a first medical image into a feature extraction network, performing feature extraction on the first medical image through the feature extraction network, and obtaining features of the first medical image.

[0059] Specifically, Figure 3a As shown, the processing model includes a feature extraction network, a feature learning network and an image reconstruction network. The computer device inputs the acquired first medical image into the feature extraction network of the processing model, and extracts features of the first medical image through the feature extraction network to obtain features of the first medical image. Optionally, the feature extraction network can be composed of a 3×3 convolution layer.

[0060] S302, inputting the features of the first medical image into a feature learning network, learning the features of the first medical image through the feature learning network, and obtaining a feature map; the feature map is a feature map that includes the relationship between pixels of the first medical image.

[0061] Specifically, the computer device inputs the features of the first medical image obtained above into the feature learning network of the processing model, and learns the features of the first medical image through the feature learning network to obtain a feature map, wherein the feature map is a feature map containing the relationship between each pixel of the first medical image. Optionally, the feature learning network includes multiple feature learning subnetworks, such as Figure 3b As shown, each feature learning subnetwork includes at least three RA (Residual Attention, a combination of residual structure and attention mechanism) modules, and each residual structure and attention mechanism combination module is densely connected. Optionally, each residual structure and attention mechanism combination module includes a residual module, a spatial attention module and a channel attention module. Optionally, the number of residual modules included in each parameter structure and attention mechanism combination module can be set according to actual needs. Usually, the number of residual modules is two. Optionally, the feature learning subnetwork can be a DRA (Dense Residual Attention, a combination of dense connections, residual structure and attention mechanism) subnetwork. It should be noted here that the dense connections between the residual structure and attention mechanism combination modules included in the feature learning subnetwork can enhance information flow and accelerate network convergence. Exemplarily, the structure of each residual structure and attention mechanism combination module can be as follows Figure 3cAs shown in FIG. 1 , the first part is two residual modules, each of which contains two convolutional layers. Optionally, the expansion coefficient of the second convolutional layer can be 1, 2, 3, etc.; the second part is the SA (spatial attention) module, whose structure can be as follows Figure 3d As shown in the figure, the spatial attention module first performs global mean pooling on the input feature map in the channel dimension, and then uses a 3×3 convolution layer with a step size of 2 to downsample the features of the input first medical image, and then uses a 3×3 deconvolution layer with a step size of 2 to upsample, and the activation function after the deconvolution is the sigmoid function; the third part is the CA (channel attention) module, whose structure can be shown as follows Figure 3e As shown in the figure, the channel attention module first performs global mean pooling on the features of the first input medical image in the spatial dimension, and then uses a 1×1 convolution layer to reduce the dimension of the channel of the feature map, and the dimension becomes 1 / n of the original, generally taking n=16, and then uses a 1×1 convolution to increase the dimension, and the dimension is increased to the original number of channels, and the activation function of the second convolution layer must be the sigmoid function.

[0062] S303, inputting the feature map into an image reconstruction network, and obtaining a second medical image through the image reconstruction network.

[0063] Specifically, the computer device inputs the obtained feature map containing the relationship between each pixel of the first medical image into the image reconstruction network of the processing model, and obtains the above-mentioned second medical image through the image reconstruction network. Optionally, the reconstruction network is also composed of a 3x3 convolution layer, and a zero padding method is used for all 3x3 convolution layers, and the number of convolution kernels is 64. It should be noted here that the activation functions used in the network structure of the processing model are all Relu functions, and the network does not contain a batch normalization layer.

[0064] In this embodiment, the processing model is an LRDNet convolutional neural network model, which is a modular neural network structure. The number of cascaded modules can be increased according to actual applications. The channel attention structure of the feature learning network can obtain the relationship between channels of the feature map, and the spatial attention structure can obtain the connection between pixels at different spatial positions on the feature map. The dilated convolution is used to expand the receptive field, and a dense connection structure is added to the network to promote information flow and long-distance dependencies between different levels, so that the processing model can fully learn the long-distance dependency information of the medical image, and then the image reconstruction network can reconstruct the feature map containing the relationship between pixels of the first medical image to obtain a second medical image with a contrast greater than that of the first medical image.

[0065] In the above scenario of inputting the first medical image into a preset processing model, the preset processing model is a pre-trained processing model. In one embodiment, Figure 4 As shown, the training method of the above preset processing model includes:

[0066] S401, obtaining a first sample medical image and a gold standard image.

[0067] Specifically, the computer device obtains a first sample medical image and a gold standard image corresponding to the first sample medical image. The contrast of the gold standard image is greater than the contrast of the first sample medical image. Optionally, the first sample medical image may be obtained by scanning the scanned object with a low dose of contrast agent, and the gold standard image may be obtained by scanning the scanned object with a standard dose of contrast agent. Exemplarily, a standard dose of contrast agent may be injected into the scanned object, and after waiting for several minutes, a high-contrast gold standard image may be obtained by magnetic resonance scanning; after waiting for several hours, due to the metabolism of the human body, the contrast agent concentration in the scanned object will gradually decrease. When the contrast agent concentration drops to half of the standard dose (a low-dose threshold may be set according to the specific situation), a low-contrast medical image may be obtained again by magnetic resonance scanning, and the above operation steps are repeated to obtain the first sample medical image in this embodiment and the gold standard image corresponding to the first sample medical image. Optionally, after the computer device acquires the first sample medical image and the gold standard image corresponding to the first sample medical image, it may perform data enhancement operations on the acquired first sample medical image and the gold standard image corresponding to the first sample medical image, so as to prevent overfitting of the processing model to a certain extent, wherein the data enhancement operations performed on the acquired first sample medical image and the gold standard image corresponding to the first sample medical image may include horizontal flipping, upside down flipping, 90-degree rotation, etc.

[0068] S402, obtaining a second sample medical image according to the first sample medical image and a preset initial processing model; the contrast of the second sample medical image is greater than the contrast of the first sample medical image.

[0069] Specifically, the computer device obtains a second sample medical image according to the above-mentioned first sample medical image and the preset initial processing model; wherein the contrast of the second sample medical image is greater than the contrast of the first sample medical image. Optionally, the computer device may obtain the amplitude map corresponding to the first sample medical image, input the amplitude map corresponding to the first sample medical image into the initial processing model, and obtain the above-mentioned second sample medical image through the initial processing model. Optionally, the computer device may perform a preprocessing operation on the first sample medical image in the complex domain, and calculate the amplitude map corresponding to the first sample medical image using its real part and imaginary part. It should be noted here that the computer device may also directly input the first sample medical image into the preset initial processing model to obtain the second sample medical image. It can be understood that the preset initial processing model has the same network structure as the above-mentioned processing model, that is, the preset initial processing model also includes a feature extraction network, a feature learning network and a reconstruction network. For the network structure and principle of the initial processing model, please refer to the above-mentioned description of the processing model, which will not be repeated in this embodiment.

[0070] S403: Obtain a value of a loss function of the initial processing model according to the second sample medical image and the gold standard image.

[0071] Specifically, the computer device obtains the value of the loss function of the processing model based on the obtained second sample medical image and the gold standard image obtained above. Optionally, the computer device can adopt a coarse-to-fine training strategy to train the initial processing model in the process of training the preset initial processing model, obtain the intermediate result of the network included in the processing model, calculate the intermediate loss of the network, and obtain the value of the loss function of the initial processing model based on the intermediate loss of the network. Optionally, the computer device can also obtain the value of the loss function of the processing model based on the pixel value of each pixel of the obtained second sample medical image and the pixel value of each pixel of the gold standard image. Optionally, the loss function of the processing model may include L1loss and Perceptual loss.

[0072] S404, adjusting the parameters of the initial processing model according to the value of the loss function of the initial processing model to obtain the processing model.

[0073] Specifically, the computer device adjusts the parameters of the initial processing model according to the obtained value of the loss function of the initial processing model to obtain the above processing model. Optionally, the computer device can adjust the parameters of the initial processing model according to the value of the loss function of the initial processing model until the value of the loss function of the initial processing model reaches a stable value or a minimum value, and the initial processing model when the value of the loss function of the initial processing model reaches a stable value or a minimum value is determined as the above processing model.

[0074] In this embodiment, the computer device obtains a large number of first sample medical images and gold standard images corresponding to the first sample medical images, and inputs the first sample medical images into a preset initial processing model. The second sample medical image can be obtained through the initial processing model, so that the value of the loss function of the initial processing model can be obtained according to the second sample medical image and the gold standard image corresponding to the first sample medical image, and then the parameters of the initial processing model can be adjusted according to the value of the loss function of the initial processing model. Since the computer device trains the initial processing model through a large number of first sample medical images, the number of first sample medical images is large, so that the initial processing model can be trained more accurately, thereby improving the accuracy of the obtained processing model.

[0075] In the above scenario where the value of the loss function of the initial processing model is obtained based on the obtained second sample medical image and the gold standard image corresponding to the first sample medical image, the second sample medical image includes a plurality of images. In one embodiment, Figure 5 As shown, the above S403 includes:

[0076] S501, obtaining the value of the loss function of each second sample medical image and the gold standard image.

[0077] Specifically, the computer device obtains the value of the loss function of each second sample medical image and the gold standard image. It can be understood that the preset initial processing model can have multiple outputs in the middle, and each second sample medical image is the intermediate output of the preset initial processing model. Optionally, the computer device can use the following formula Obtain the value of the loss function of each second sample medical image and the gold standard image, where, represents the value of the loss function between each first sample medical image and the gold standard image, α, β are parameters, and illustratively, in this embodiment, α=1, β=1, Represents the i-th output result of the preset processing model, I gt represents the gold standard image corresponding to the first sample medical image, Φ l (I gt ) represents the feature map of the first layer corresponding to the gold standard image as input of the processing model. It represents the feature map of the lth output corresponding to the initial processing model with I as input. It should be noted here that when the initial processing model does not include a batch normalization layer, l can be 34. It should be noted here that the initial processing model is the VGG-19 model.

[0078] S502, summing the loss function values ​​of each second sample medical image and the gold standard image to obtain the loss function value of the initial processing model.

[0079] Specifically, the computer device sums the loss function values ​​of each second sample medical image and the gold standard image obtained above to obtain the loss function value of the initial processing model. Optionally, the computer device can be based on the formula: Get the value of the loss function of the initial processing model, where L represents the value of the loss function of the initial processing model, and W i represents the weight coefficient, represents the value of the loss function between each second sample medical image and the gold standard image, and n represents the total number of values ​​of the loss function between the second sample medical images and the gold standard image.

[0080] In this embodiment, the loss function values ​​of each second sample medical image and the gold standard image obtained by the computer device are the loss function values ​​of each intermediate output of the initial processing model and the gold standard image, so that the loss function values ​​of each second sample medical image and the gold standard image can be summed to accurately obtain the loss function value of the initial processing model, and then the initial processing model can be accurately trained according to the loss function value of the initial processing model, thereby improving the training accuracy of the initial processing model.

[0081] It should be understood that although Figure 2-5 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 2-5 At least part of the steps may include multiple steps or multiple stages. These steps or stages are not necessarily performed at the same time, but can be performed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed in turn or alternately with other steps or at least part of the steps or stages in other steps.

[0082] In one embodiment, Figure 6 As shown, an image processing device is provided, comprising: a first acquisition module and a first processing module, wherein:

[0083] The first acquisition module is used to acquire a first medical image.

[0084] The first processing module is used to input the first medical image into a preset processing model, and obtain a second medical image through the processing model; wherein the contrast of the second medical image is greater than the contrast of the first medical image; the processing model is obtained by training based on the first sample medical image and the gold standard image corresponding to the first sample medical image, and the contrast of the gold standard image is greater than the contrast of the first sample medical image.

[0085] The image processing device provided in this embodiment can execute the above method embodiment, and its implementation principle and technical effect are similar, which will not be repeated here.

[0086] On the basis of the above embodiment, optionally, the above processing model includes a feature extraction network, a feature learning network and an image reconstruction network; the above first processing module includes: a feature extraction unit, a first acquisition unit and a reconstruction unit, wherein:

[0087] The feature extraction unit is used to input the first medical image into the feature extraction network, extract features of the first medical image through the feature extraction network, and obtain features of the first medical image.

[0088] The first acquisition unit is used to input the features of the first medical image into the feature learning network, learn the features of the first medical image through the feature learning network, and obtain a feature map; the feature map is a feature map containing the relationship between each pixel of the first medical image.

[0089] The reconstruction unit is used to input the feature map into the image reconstruction network, and obtain the second medical image through the image reconstruction network.

[0090] Optionally, the feature learning network includes multiple feature learning sub-networks, each feature learning sub-network includes at least three residual structure and attention mechanism combination modules, and each residual structure and attention mechanism combination module is densely connected.

[0091] Optionally, each residual structure and attention mechanism combination module includes a residual module, a spatial attention module and a channel attention module.

[0092] The image processing device provided in this embodiment can execute the above method embodiment, and its implementation principle and technical effect are similar, which will not be repeated here.

[0093] On the basis of the above embodiment, optionally, the above device further includes: a second acquisition module, a second processing module, a third acquisition module and a training module, wherein:

[0094] The second acquisition module is used to acquire the first sample medical image and the gold standard image corresponding to the first sample medical image.

[0095] The second processing module is used to obtain a second sample medical image according to the first sample medical image and a preset initial processing model; the contrast of the second sample medical image is greater than the contrast of the first sample medical image.

[0096] The third acquisition module is used to obtain the value of the loss function of the initial processing model according to the second sample medical image and the gold standard image.

[0097] The training module is used to adjust the parameters of the initial processing model according to the value of the loss function of the initial processing model to obtain the processing model.

[0098] The image processing device provided in this embodiment can execute the above method embodiment, and its implementation principle and technical effect are similar, which will not be repeated here.

[0099] On the basis of the above embodiment, optionally, the second sample medical image includes a plurality of images, and the third acquisition module includes: a second acquisition unit and a third acquisition unit, wherein:

[0100] The second acquisition unit is used to acquire the value of the loss function of each second sample medical image and the gold standard image.

[0101] The third acquisition unit is used to sum the values ​​of the loss function of each second sample medical image and the gold standard image to obtain the value of the loss function of the initial processing model.

[0102] The image processing device provided in this embodiment can execute the above method embodiment, and its implementation principle and technical effect are similar, which will not be repeated here.

[0103] On the basis of the above embodiment, optionally, the above second processing module includes: a fourth acquisition unit and a processing unit, wherein:

[0104] The fourth acquisition unit is used to acquire the amplitude map corresponding to the first sample medical image.

[0105] The processing unit is used to input the amplitude map corresponding to the first sample medical image into the initial processing model, and obtain the second sample medical image through the initial processing model.

[0106] The image processing device provided in this embodiment can execute the above method embodiment, and its implementation principle and technical effect are similar, which will not be repeated here.

[0107] For the specific definition of the image processing device, please refer to the definition of the image processing method above, which will not be repeated here. Each module in the above image processing device can be implemented in whole or in part by software, hardware and a combination thereof. The above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0108] In one 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, the following steps are implemented:

[0109] acquiring a first medical image;

[0110] The first medical image is input into a preset processing model, and a second medical image is obtained through the processing model; wherein the contrast of the second medical image is greater than the contrast of the first medical image; the processing model is obtained by training based on the first sample medical image and a gold standard image corresponding to the first sample medical image, and the contrast of the gold standard image is greater than the contrast of the first sample medical image.

[0111] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0112] acquiring a first medical image;

[0113] The first medical image is input into a preset processing model, and a second medical image is obtained through the processing model; wherein the contrast of the second medical image is greater than the contrast of the first medical image; the processing model is obtained by training based on the first sample medical image and a gold standard image corresponding to the first sample medical image, and the contrast of the gold standard image is greater than the contrast of the first sample medical image.

[0114] Those of ordinary skill 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 memory, storage, database or other media 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 or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. 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).

[0115] The technical features of the above embodiments may be combined arbitrarily. 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.

[0116] The above-mentioned 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 invention patent. 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 patent of the present application shall be subject to the attached claims.

Claims

1. An image processing method, characterized in that: The method comprises: acquiring a first medical image; The first medical image is input into a preset processing model, and a second medical image is obtained through the processing model; wherein the contrast of the second medical image is greater than the contrast of the first medical image; the processing model is obtained by training according to the first sample medical image and the gold standard image corresponding to the first sample medical image, and the contrast of the gold standard image is greater than the contrast of the first sample medical image; the first sample medical image is obtained by scanning the scanned object with a low dose of contrast agent; the gold standard image is obtained by scanning the scanned object with a standard dose of contrast agent; the processing model includes a feature learning network, the feature learning network includes multiple feature learning sub-networks, each of the feature learning sub-networks includes at least three residual structure and attention mechanism combination modules, and each of the residual structure and attention mechanism combination modules is densely connected; the first part of each of the residual structure and attention mechanism combination modules is two residual modules, the second part is a spatial attention module, and the third part is a channel attention module; each of the residual modules includes a dilated convolution.

2. The method according to claim 1, characterized in that The processing model further includes a feature extraction network and an image reconstruction network. The first medical image is input into a preset processing model, and a second medical image is obtained through the processing model, including: Inputting the first medical image into the feature extraction network, performing feature extraction on the first medical image through the feature extraction network, and obtaining features of the first medical image; Inputting the features of the first medical image into the feature learning network, and learning the features of the first medical image through the feature learning network to obtain a feature map; the feature map is a feature map containing the relationship between each pixel of the first medical image; The feature map is input into the image reconstruction network, and the second medical image is obtained through the image reconstruction network.

3. The method according to claim 1, characterized in that The training process of the processing model includes: Acquire a first sample medical image and a gold standard image corresponding to the first sample medical image; Obtaining a second sample medical image according to the first sample medical image and a preset initial processing model; the contrast of the second sample medical image is greater than the contrast of the first sample medical image; Obtaining a value of a loss function of the initial processing model according to the second sample medical image and the gold standard image; According to the value of the loss function of the initial processing model, the parameters of the initial processing model are adjusted to obtain the processing model.

4. The method according to claim 3, characterized in that The second sample medical image includes a plurality of images, and obtaining the value of the loss function of the initial processing model according to the second sample medical image and the gold standard image includes: Obtaining a value of a loss function between each of the second sample medical images and the gold standard image; The loss function values ​​of each of the second sample medical images and the gold standard image are summed to obtain the loss function value of the initial processing model.

5. The method according to claim 4, characterized in that The step of obtaining a second sample medical image according to the first sample medical image and a preset initial processing model includes: Obtaining an amplitude map corresponding to the first sample medical image; The amplitude map corresponding to the first sample medical image is input into the initial processing model, and the second sample medical image is obtained through the initial processing model.

6. The method according to any one of claims 1 to 5, characterized in that: The method further comprises: A data enhancement operation is performed on the first sample medical image and the gold standard image; the data enhancement operation includes horizontal flipping, upside down flipping, and 90-degree rotation.

7. An image processing device, characterized in that: The device comprises: A first acquisition module, used for acquiring a first medical image; A first processing module is used to input the first medical image into a preset processing model, and obtain a second medical image through the processing model; wherein the contrast of the second medical image is greater than the contrast of the first medical image; the processing model is obtained by training based on the first sample medical image and the gold standard image corresponding to the first sample medical image, and the contrast of the gold standard image is greater than the contrast of the first sample medical image; the first sample medical image is obtained by scanning the scanned object with a low dose of contrast agent; the gold standard image is obtained by scanning the scanned object with a standard dose of contrast agent; the processing model includes a feature learning network, the feature learning network includes multiple feature learning sub-networks, each of the feature learning sub-networks includes at least three residual structure and attention mechanism combination modules, and each of the residual structure and attention mechanism combination modules is densely connected; the first part of each of the residual structure and attention mechanism combination modules is two residual modules, the second part is a spatial attention module, and the third part is a channel attention module; each of the residual modules includes a dilated convolution.

8. The device according to claim 7, characterized in that The processing model also includes a feature extraction network and an image reconstruction network, and the first processing module includes: a feature extraction unit, configured to input the first medical image into the feature extraction network, perform feature extraction on the first medical image through the feature extraction network, and obtain features of the first medical image; A first acquisition unit is used to input the features of the first medical image into the feature learning network, and learn the features of the first medical image through the feature learning network to obtain a feature map; the feature map is a feature map containing the relationship between each pixel of the first medical image; A reconstruction unit is used to input the feature map into the image reconstruction network, and obtain the second medical image through the image reconstruction network.

9. 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.

10. 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.

Images