Medical image processing method and device, computer device and storage medium

Abstract

The application relates to a medical image processing method and device, computer equipment and a storage medium, and is suitable for the technical field of computers. The method comprises the following steps: inputting a to-be-processed medical image into a target artifact recognition model to obtain target artifact attribute information output by the target artifact recognition model, wherein the target artifact attribute information is used to indicate the attribute characteristics of artifacts in the to-be-processed medical image; inputting the to-be-processed medical image and the target artifact attribute information into a target artifact degree recognition model to obtain artifact degree indication information output by the target artifact degree recognition model, wherein the artifact degree indication information is used to indicate the influence degree of the artifacts in the to-be-processed medical image on the image quality of the to-be-processed medical image; and if the influence degree of the artifacts in the to-be-processed medical image on the image quality of the to-be-processed medical image is greater than or equal to a preset artifact influence degree threshold, outputting prompt information. The method can improve the image quality.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a medical image processing method, apparatus, computer device, and storage medium. Background Technology

[0002] With the rapid development of science and technology, nuclear magnetic resonance imaging (NMR) technology has become increasingly mature. NMR technology primarily infers the distribution of water molecules within a scanned object by identifying the distribution of hydrogen atom signals in water molecules, thereby probing the internal structure of the scanned object. However, during NMR imaging, factors such as the state of the scanned object, the state of the scanning equipment, and the external environment can cause motion artifacts in the NMR images. Motion artifacts negatively impact the image quality of NMR images; therefore, it is necessary to minimize the presence of artifacts in NMR images.

[0003] In traditional techniques, after an MRI image is generated, the scanning operator usually examines the generated MRI image to determine the impact of artifacts on the image quality, and determines whether the scanned area needs to be re-scanned or re-scanned based on the impact of artifacts on the image quality.

[0004] However, since the scanning operator needs to examine and judge the MRI images, it is highly subjective and there is a possibility of inaccurate judgment, making it difficult to guarantee image quality. Summary of the Invention

[0005] Therefore, it is necessary to provide a medical image processing method, apparatus, computer equipment, and storage medium that can improve image quality in response to the above-mentioned technical problems.

[0006] In a first aspect, a medical image processing method is provided, comprising: inputting a medical image to be processed into a target artifact recognition model to obtain target artifact attribute information output by the target artifact recognition model, wherein the target artifact attribute information is used to indicate the attribute characteristics of artifacts in the medical image to be processed; inputting the medical image to be processed and the target artifact attribute information into a target artifact degree recognition model to obtain artifact degree indication information output by the target artifact degree recognition model, wherein the artifact degree indication information is used to indicate the degree of influence of artifacts in the medical image to be processed on the image quality of the medical image to be processed; and outputting a prompt message if the degree of influence of artifacts in the medical image to be processed on the image quality of the medical image to be processed is greater than or equal to a preset artifact influence degree threshold.

[0007] In one embodiment, before or after inputting the medical image to be processed into the target artifact recognition model, the method further includes: identifying the scanned area and scanning orientation in the medical image to be processed.

[0008] In one embodiment, before inputting the medical image to be processed into the target artifact recognition model, the scanned area and scanning orientation are identified in the medical image to be processed, including: obtaining the field strength information of the medical device corresponding to the medical image to be processed; obtaining the area model corresponding to the field strength information of the medical device; inputting the medical image to be processed into the area model to obtain the scanned area and scanning orientation included in the medical image to be processed.

[0009] In one embodiment, after inputting the medical image to be processed into the target artifact recognition model, the scanned area and scanning orientation are identified in the medical image to be processed, including: obtaining the field strength information of the medical device corresponding to the medical image to be processed; determining the area model based on the field strength information of the medical device and the target artifact attribute information; inputting the medical image to be processed into the area model to obtain the scanned area and scanning orientation included in the medical image to be processed.

[0010] In one embodiment, the target artifact degree recognition model is determined as follows: the target artifact degree recognition model is determined based on at least one of the following: medical device field strength information, target artifact attribute information, scanned area and scanning orientation.

[0011] In one embodiment, the training process of the target artifact recognition model is as follows: a first training sample set is obtained, which includes multiple first training samples, each of which includes a first training sample image and training artifact attribute information corresponding to the first training sample image; the image brightness of each first training sample image in the first training sample set is normalized based on the Z score; and the artifact recognition network is trained based on the normalized first training sample set to obtain the target artifact recognition model.

[0012] In one embodiment, the training process of the target artifact recognition model is as follows: a second training sample set is obtained, which includes multiple second training samples, each of which includes a second training sample image and artifact degree indication information corresponding to the second training sample image; the image brightness of each second training sample image in the second training sample set is normalized based on the Z score; and the artifact recognition network is trained based on the normalized second training sample set to obtain the target artifact recognition model.

[0013] In one embodiment, the training process of the part model is as follows: a third training sample set is obtained, which includes multiple third training samples. Each third training sample includes a third training sample image and the scanned part and scanning direction corresponding to the third training sample image; the image brightness of each third training sample image in the third training sample set is normalized based on the Z score; and the part recognition network is trained based on the normalized third training sample set to obtain the part model.

[0014] Secondly, a medical image processing apparatus is provided, the apparatus comprising:

[0015] The first input module is used to input the medical image to be processed into the target artifact recognition model to obtain the target artifact attribute information output by the target artifact recognition model. The target artifact attribute information is used to indicate the attribute characteristics of artifacts in the medical image to be processed.

[0016] The second input module is used to input the medical image to be processed and the target artifact attribute information into the target artifact degree recognition model to obtain the artifact degree indication information output by the target artifact degree recognition model. The artifact degree indication information is used to indicate the degree of influence of artifacts in the medical image to be processed on the image quality of the medical image to be processed.

[0017] The output module is used to output a prompt message when the impact of artifacts in the medical image being processed on the image quality of the medical image being processed is greater than or equal to a preset artifact impact threshold.

[0018] In one embodiment, the above-mentioned medical image processing apparatus further includes:

[0019] The recognition module is used to identify the scanned area and scanning orientation in the medical image to be processed.

[0020] In one embodiment, the aforementioned identification module is specifically used to: acquire the field strength information of the medical device corresponding to the medical image to be processed; acquire the part model corresponding to the field strength information of the medical device; input the medical image to be processed into the part model to obtain the scanned part and scanning orientation included in the medical image to be processed.

[0021] In one embodiment, the aforementioned identification module is specifically used to: acquire the field strength information of the medical device corresponding to the medical image to be processed; determine the location model based on the field strength information of the medical device and the target artifact attribute information; input the medical image to be processed into the location model to obtain the scanned location and scanning orientation included in the medical image to be processed.

[0022] In one embodiment, the above-mentioned medical image processing apparatus further includes:

[0023] The determination module is used to determine the target artifact degree recognition model based on at least one of the following: medical device field strength information, target artifact attribute information, scanned area, and scanning orientation.

[0024] In one embodiment, the above-mentioned medical image processing apparatus further includes:

[0025] The first acquisition module is used to acquire a first training sample set, which includes multiple first training samples. Each first training sample includes a first training sample image and training artifact attribute information corresponding to the first training sample image.

[0026] The first processing module is used to normalize the image brightness of each first training sample image in the first training sample set based on the Z score.

[0027] The first training module is used to train the artifact recognition network based on the normalized first training sample set to obtain the target artifact recognition model.

[0028] In one embodiment, the above-mentioned medical image processing apparatus further includes:

[0029] The second acquisition module is used to acquire a second training sample set, which includes multiple second training samples. Each second training sample includes a second training sample image and artifact degree indication information corresponding to the second training sample image.

[0030] The second processing module is used to normalize the image brightness of each second training sample image in the second training sample set based on the Z score.

[0031] The second training module is used to train the artifact recognition network based on the normalized second training sample set to obtain the target artifact recognition model.

[0032] In one embodiment, the above-mentioned medical image processing apparatus further includes:

[0033] The third acquisition module is used to acquire the third training sample set, which includes multiple third training samples. Each third training sample includes a third training sample image and the scanned part and scanning direction corresponding to the third training sample image.

[0034] The third processing module is used to normalize the image brightness of each third training sample image in the third training sample set based on the Z score.

[0035] Based on the Z-score, the image brightness of each third training sample image in the third training sample set is normalized.

[0036] The third training module is used to train the part recognition network based on the normalized third training sample set to obtain the part model.

[0037] Thirdly, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method described in any of the first aspects above.

[0038] Fourthly, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the method described in any of the first aspects above.

[0039] The aforementioned medical image processing method, apparatus, computer equipment, and storage medium input the medical image to be processed into a target artifact recognition model to obtain target artifact attribute information output by the model. The medical image to be processed and the target artifact attribute information are then input into a target artifact severity recognition model to obtain artifact severity indication information output by the model. If the impact of artifacts in the medical image to be processed on the image quality is greater than or equal to a preset artifact impact severity threshold, a prompt message is output. This method accurately obtains the attribute information of the target artifacts after inputting the medical image to be processed into the target artifact recognition model. Then, after inputting the attribute information of the medical image to be processed and the target artifacts into the target artifact severity recognition model, the model can accurately and effectively determine the impact of artifacts in the medical image to be processed on the image quality based on the target artifact attribute information. Furthermore, the computer equipment can compare the impact of artifacts in the medical image to be processed on the image quality with the preset artifact impact severity threshold, thereby accurately determining whether the medical image to be processed needs to be rescanned or re-scanned, further ensuring image quality. Attached Figure Description

[0040] Figure 1 This is a flowchart illustrating a medical image processing method in one embodiment;

[0041] Figure 2 This is a schematic diagram of the target artifact recognition model structure in a medical image processing method in one embodiment.

[0042] Figure 3 This is a schematic diagram of the interface in a medical image processing method in one embodiment;

[0043] Figure 4 This is a flowchart illustrating a medical image processing method in another embodiment;

[0044] Figure 5 This is a flowchart illustrating a medical image processing method in another embodiment;

[0045] Figure 6 This is a flowchart illustrating a medical image processing method in another embodiment;

[0046] Figure 7 This is a flowchart illustrating a medical image processing method in another embodiment;

[0047] Figure 8 This is a flowchart illustrating a medical image processing method in another embodiment;

[0048] Figure 9 This is a flowchart illustrating a medical image processing method in another embodiment;

[0049] Figure 10 This is a flowchart illustrating a medical image processing method in another embodiment;

[0050] Figure 11 This is a structural block diagram of a medical image processing device in one embodiment;

[0051] Figure 12 This is a structural block diagram of a medical image processing device in one embodiment;

[0052] Figure 13 This is a structural block diagram of a medical image processing device in one embodiment;

[0053] Figure 14 This is a structural block diagram of a medical image processing device in one embodiment;

[0054] Figure 15 This is a structural block diagram of a medical image processing device in one embodiment;

[0055] Figure 16 This is a structural block diagram of a medical image processing device in one embodiment;

[0056] Figure 17 This is an internal structure diagram of a computer device that is a server in one embodiment.

[0057] Figure 18 This is an internal structure diagram of a computer device that is a terminal in one embodiment. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0059] With the rapid development of technology, medical imaging technology is becoming increasingly sophisticated. During the medical imaging process, artifacts may appear in the final reconstructed medical images due to factors such as the state of the scanned object, the state of the scanning equipment, and the external environment. Artifacts correspond to the influence of tissues or lesions that are not actually present in the detected object, often manifesting as image distortion, overlap, missing parts, or blurring. Artifacts can degrade medical image quality, mask lesions, and create false lesions, thus significantly increasing the likelihood of misdiagnosis by clinicians. Therefore, it is necessary to minimize artifacts that may occur during medical imaging to ensure that medical imaging equipment operates at its optimal state.

[0060] In traditional techniques, after medical images are generated, the scanning operator usually examines them to determine the impact of artifacts on image quality and decides whether to rescan or re-scan the scanned area based on the impact of artifacts on image quality.

[0061] However, since the scanning operator needs to examine and judge the medical images, it is highly subjective and there is a possibility of inaccurate judgment, making it difficult to guarantee image quality.

[0062] This application proposes a medical image processing method to address the aforementioned technical problems. The method mainly includes the following steps: inputting the medical image to be processed into a target artifact recognition model to obtain target artifact attribute information output by the model. This target artifact attribute information indicates the attribute characteristics of artifacts in the medical image to be processed. Then, inputting the medical image to be processed and the target artifact attribute information into a target artifact degree recognition model to obtain artifact degree indication information output by the model. This artifact degree indication information indicates the degree of influence of artifacts in the medical image to be processed on the image quality. If the degree of influence of artifacts in the medical image to be processed on the image quality is greater than or equal to a preset artifact influence degree threshold, a prompt message is output. This prompt message prompts the user to confirm whether to accept the artifacts in the medical image to be processed and whether a rescan of the corresponding scanned area is required. This application provides a medical image processing method in which the attribute information of the target artifacts can be accurately obtained after inputting the medical image to be processed into the target artifact recognition model. Then, after inputting the attribute information of the medical image to be processed and the target artifact into the target artifact degree recognition model, the model can accurately and effectively determine the degree of influence of artifacts in the medical image on the image quality of the medical image to be processed based on the attribute information of the target artifacts. Furthermore, the computer device can compare the degree of influence of artifacts in the medical image on the image quality of the medical image to be processed with a preset artifact influence threshold. This allows for timely notification of information to the user interface when artifacts have a significant impact on the image quality of the medical image, further ensuring image quality.

[0063] It should be noted that the medical image processing method provided in this application embodiment can be executed by a medical image processing device. This device can be implemented as part or all of a computer device through software, hardware, or a combination of both. The computer device can be a server or a terminal. In this application embodiment, the server can be a single server or a server cluster composed of multiple servers. The terminal in this application embodiment can be a smartphone, personal computer, tablet computer, wearable device, or other intelligent hardware device such as an intelligent robot. The processed medical images can be magnetic resonance (MR) images, computed tomography (CT) images, positron emission tomography (PET) images, digital radiography (DR) images, ultrasound (US) images, and fused images of two of the above modalities. In the following method embodiments, the execution subject is always described using a computer device as an example.

[0064] In one embodiment of this application, such as Figure 1As shown, a medical image processing method is provided, and the method is illustrated using an application to a computer device as an example. The method includes the following steps:

[0065] Step 101: The computer device inputs the medical image to be processed into the target artifact recognition model to obtain the target artifact attribute information output by the target artifact recognition model.

[0066] The target artifact attribute information is used to indicate the attribute characteristics of artifacts in the medical image to be processed. The medical image to be processed can be a magnetic resonance (MR) image, a computed tomography (CT) image, a positron emission tomography (PET) image, a digital radiography (DR) image, an ultrasound (US) image, or a fusion image of two or more of the above modalities.

[0067] Specifically, the computer device can send scanning instructions to the medical device. After receiving the scanning instructions from the computer device, the medical device can scan the area to be scanned and send the scanned data back to the computer device. The computer device receives the scan data sent by the medical device and generates a medical image corresponding to the scanned area based on the scan data.

[0068] For example, taking an MRI image as the medical image to be processed, the computer device can send a scanning command to the MRI machine. After receiving the scanning command from the computer device, the MRI machine can scan the area to be scanned and send the scanned data back to the computer device. The computer device receives the scan data sent by the MRI machine and generates a medical image corresponding to the scanned area based on the scan data.

[0069] The computer device can input the medical image to be processed into the target artifact recognition model. Optionally, the target artifact recognition model can be a machine learning network model, or other network models. The machine learning network model can be DNN (Deep Neural Networks), CNN (Convolutional Neural Networks), RNN (Recurrent Neural Networks), etc. When the target artifact recognition model is a CNN, it can be a V-Net model, U-Net model, Generative Adversarial Nets model, etc. This application does not specifically limit the type of target artifact recognition model.

[0070] Optionally, when the target artifact recognition model is a CNN, the target artifact recognition model may include a 50-layer deep convolutional neural network, comprising 4 residual blocks, 49 convolutions, and one fully connected layer. The structure of the target artifact recognition model is as follows: Figure 2 As shown, the activation function used is the ReLU function, and the formula is:

[0071] F n =R(W n *F n-1 +B n )

[0072] Where R represents the nonlinear activation function ReLU, W n B n F represents the weights and biases of the convolutional layer in the feature extraction stage, respectively. n-1 F represents the feature map output by the previous convolution. n This represents the output feature map obtained in the current feature extraction stage.

[0073] After processing the medical image to be processed, the target artifact recognition model can output target artifact attribute information, which indicates the attribute characteristics of artifacts in the medical image. The target artifact attribute information may include at least one of the following: artifact size, artifact location, artifact quantity, and artifact type.

[0074] Artifacts can be categorized into several types, including zipper artifacts, spark artifacts, involuntary motion artifacts, breathing artifacts, and vascular pulsation artifacts. They can also be classified by their source into equipment artifacts and human artifacts. Equipment artifacts include, for example, measurement error artifacts in the imaging system, X-ray beam hardening artifacts, high voltage fluctuation artifacts in the imaging system, temperature drift artifacts in electronic circuits, and detector drift artifacts. Human artifacts include, for example, artifacts caused by the movement of the object being detected, artifacts caused by the peristalsis of internal organs, and artifacts caused by internal metal implants.

[0075] Step 102: The computer device inputs the medical image to be processed and the target artifact attribute information into the target artifact degree recognition model to obtain the artifact degree indication information output by the target artifact degree recognition model.

[0076] Among them, the artifact degree indicator information is used to indicate the degree to which artifacts in the medical image to be processed affect the image quality of the medical image to be processed.

[0077] Specifically, after inputting the medical image to be processed into the target artifact recognition model and obtaining the target artifact attribute information output by the model, the computer device can input the medical image to be processed and the target artifact attribute information into the target artifact severity recognition model. The target artifact severity recognition model can determine the degree of influence of artifacts in the medical image to be processed on the image quality of the medical image based on the target artifact attribute information.

[0078] Optionally, the target artifact recognition model can identify the medical image to be processed and divide it into regions of interest (ROI) and non-ROI. The ROI can be the scanned area within the medical image. For example, when the scanned area is the brain, the image to be processed includes both image information corresponding to the brain and a small portion of the neck. The target artifact recognition model classifies the neck in the medical image as a non-ROI and the brain as a ROI.

[0079] After identifying the region of interest (ROI) in the medical image to be processed, the target artifact identification model can determine the degree of impact of artifacts on the image quality of the medical image to be processed based on the location information of the ROI, the attribute information of the ROI, and the attribute information of the target artifacts.

[0080] For example, the scanned area in the medical image to be processed is the brain. The target artifact severity identification model identifies brain tissues such as white matter and gray matter in the image as regions of interest (ROIs), and the neck as a non-ROI. Based on the target artifact attribute information, the model determines that the target artifact in the image is a neck motion artifact. Since neck motion artifacts have minimal impact on brain tissue, the target artifact severity identification model determines that the artifacts in the image have a relatively small impact on the image quality.

[0081] Optionally, the target artifact severity identification model can determine the degree of influence of artifacts in the medical image to be processed on the image quality of the medical image to be processed based on the difference between the signal-to-noise ratio, contrast, etc. of the medical image to be processed containing target artifacts and the set image quality thresholds (e.g., signal-to-noise ratio threshold, contrast threshold). If the difference between the signal-to-noise ratio, contrast, etc. of the medical image to be processed containing target artifacts and the set image quality thresholds is larger, the target artifact severity identification model determines that the artifacts in the medical image to be processed have a greater influence on the image quality of the medical image to be processed; conversely, if the difference between the signal-to-noise ratio, contrast, etc. of the medical image to be processed containing target artifacts and the set image quality thresholds is smaller, the target artifact severity identification model determines that the artifacts in the medical image to be processed have a smaller influence on the image quality of the medical image to be processed.

[0082] Optionally, the target artifact severity identification model can determine the degree of influence of artifacts in the medical image being processed on the image quality of the medical image based on the positional relationship between the target artifact and the scanned area. If the distance between the target artifact and the scanned area is less than a preset distance threshold, the target artifact severity identification model determines that the artifacts in the medical image being processed have a significant impact on the image quality of the medical image being processed; if the distance between the target artifact and the scanned area is greater than or equal to the preset distance threshold, the target artifact severity identification model determines that the artifacts in the medical image being processed have a relatively small impact on the image quality of the medical image being processed.

[0083] Optionally, the target artifact severity identification model can determine the degree of influence of artifacts in the medical image being processed on the image quality based on the area size of the target artifacts. If the area of ​​the target artifact exceeds a preset area threshold, the target artifact severity identification model determines that the artifacts in the medical image being processed have a significant impact on the image quality; if the area of ​​the target artifact is less than the preset area threshold, the target artifact severity identification model determines that the artifacts in the medical image being processed have a relatively small impact on the image quality.

[0084] Optionally, the target artifact severity identification model can determine the degree of impact of artifacts in the medical image being processed on the image quality based on the number of target artifacts. If the number of target artifacts exceeds a preset threshold, the target artifact severity identification model determines that the artifacts in the medical image being processed have a significant impact on the image quality; if the number of target artifacts is less than the preset threshold, the target artifact severity identification model determines that the artifacts in the medical image being processed have a relatively small impact on the image quality.

[0085] Optionally, the target artifact identification model can determine the degree of impact of artifacts in the medical image being processed on the image quality based on the type of target artifact. If the type of target artifact is one that is unavoidable during the scanning process of the scanned area, the target artifact identification model determines that the artifacts in the medical image being processed have a relatively small impact on the image quality; if the type of target artifact is one that is avoidable during the scanning process of the scanned area, the target artifact identification model determines that the artifacts in the medical image being processed have a relatively large impact on the image quality.

[0086] Step 103: If the impact of artifacts in the medical image to be processed on the image quality of the medical image to be processed is greater than or equal to the preset artifact impact threshold, the computer device outputs a prompt message.

[0087] The prompt information can be a prompt icon to prompt the user to confirm whether to accept the artifacts in the medical image to be processed, and whether to rescan the scanned area corresponding to the medical image to be processed; or, the prompt information can be just a warning icon to indicate that there are artifacts in the medical image to be processed that affect the image quality; the prompt information can be a specific sequence, which corresponds to the medical image affected by the artifacts, and the sequence is a time sequence in the entire medical imaging scan.

[0088] In one embodiment, to facilitate determining the degree of impact of artifacts in the medical image being processed on its quality, the target artifact degree recognition model can classify the artifact degree indication information when outputting it. Optionally, the computer device can classify the artifact degree indication information into four levels: Level 1, Level 2, Level 3, and Level 4. Level 1 indicates that the medical image being processed is normal and unaffected by artifacts; Level 2 indicates that artifacts have a slight impact on the medical image being processed; Level 3 indicates that artifacts have a moderate impact on the medical image being processed; and Level 4 indicates that artifacts have a severe impact on the medical image being processed.

[0089] Optionally, the artifact severity level classification can be determined by multiple researchers studying multiple scanned images containing artifacts, or it can be obtained by multiple researchers labeling the impact of multiple scanned images containing artifacts and using the labeled scanned images to train a machine learning model.

[0090] Optionally, assuming a preset artifact severity threshold of level two, the computer device outputs a prompt message when the artifact severity indication information output by the target artifact severity recognition model is level three or four. The computer device may output the prompt message by emitting a sound, emitting a red light, displaying a prompt for rescanning on the screen, or displaying a scan sequence segment corresponding to the medical image containing the artifact. This embodiment does not specifically limit the method by which the computer device outputs the prompt message.

[0091] For example, after completing a scan, the medical device can obtain multiple medical images based on the scan data, each corresponding to a different scan sequence. After the computer device identifies the artifact levels of the multiple medical images using a target artifact level identification model, it can determine the target medical image from the multiple medical images based on the artifact level indication information output by the target artifact level identification model and a preset artifact level threshold. The target medical image is defined as one where the artifacts in the target medical image have an impact on the image quality greater than or equal to the preset artifact impact threshold. The computer device then outputs the scan sequence corresponding to the target medical image, prompting the user to confirm whether to accept the artifacts in the target medical image and whether a rescan of the corresponding scanned area is necessary.

[0092] For example, such as Figure 3 The diagram shows a scanning interface. The first row displays preview images of the scanned object from three angles. The second column, from left to right, shows the scanning protocol execution area, the site display area, and the recommended protocol area. The scanning protocol execution area contains multiple categories of scanning protocols required for this scan (arranged from top to bottom, set by the physician or automatically by the system). The site display area shows multiple scanned sites of the object, and each scanned site in this area is linked to the recommended protocol area. Clicking on a scanned site displays recommended scanning protocols that can be executed. The selected recommended scanning protocol is then added to the scanning protocol execution area. In this embodiment... Figure 3 The scanning protocol execution area in the computer device has already executed two scan sequences, numbered 1 and 2. Figure 1 The method shown processes medical images obtained from scan sequences and outputs a prompt message on the interface. This prompt message corresponds to scan sequence number 2, which is associated with the medical image affected by artifacts. Furthermore, the operator can select to rescan scan sequence number 2 based on the prompt message.

[0093] In another embodiment, the prompt information remains the scan sequence numbered 2 corresponding to the medical image affected by artifacts (the scan sequence in this time period is affected by artifacts). A recommended scan protocol is generated in the recommendation protocol area based on the order and functionality of multiple categories of scan protocols set in the scan protocol execution area. The recommended scan protocol may have a different category or time sequence than the multiple categories of scan protocols set in the scan protocol execution area, but it can achieve the same image display effect. Optionally, the recommended scan protocol can be automatically generated using a protocol recommendation model trained based on big data. This protocol recommendation model can be obtained by training a neural network using multiple sets of scan protocols. For example, the multiple sets of scan protocols include a first set of sample scan protocols and a second set of sample scan protocols with equivalent or similar image display effects. The first set of sample scan protocols may have different types of sub-scan protocols than the second set of sample scan protocols, or the first set of sample scan protocols may have sub-scan protocols with different time sequences than the second set of sample scan protocols.

[0094] In the aforementioned medical image processing method, the computer device inputs the medical image to be processed, obtained from scanning by a medical device, into a target artifact recognition model to obtain target artifact attribute information output by the model. The computer device then inputs the medical image to be processed and the target artifact attribute information into a target artifact severity recognition model to obtain artifact severity indication information output by the model. Furthermore, if the impact of artifacts in the medical image to be processed on the image quality is greater than or equal to a preset artifact impact threshold, the computer device outputs a prompt. This method, by inputting the medical image to be processed into the target artifact recognition model, can accurately obtain the attribute information of the target artifacts. Then, by inputting the attribute information of the medical image to be processed and the target artifacts into the target artifact severity recognition model, the model can accurately and effectively determine the impact of artifacts in the medical image to be processed on the image quality based on the target artifact attribute information. In addition, the computer equipment can compare the degree of impact of artifacts in the medical images to be processed on the image quality of the medical images to be processed with a preset artifact impact threshold, so as to accurately determine whether the medical images to be processed need to be scanned again or re-scanned, thereby further ensuring the image quality.

[0095] In an optional embodiment of this application, before or after step 101 "the computer device inputs the medical image to be processed into the target artifact recognition model" described above, the following may also be included:

[0096] Computer equipment identifies the scanned area and scanning orientation in medical images to be processed.

[0097] Optionally, the computer device can identify the scanned area in the medical image to be processed using a preset first image recognition algorithm, and determine the orientation of the scanned area. The orientation of the scanned area can be any one of coronal, sagittal, and transverse views. The preset first image recognition algorithm can identify both the scanned area in the medical image to be processed and the corresponding scanning orientation.

[0098] Specifically, the computer device can extract features from the medical image to be processed using a preset first image recognition algorithm, and determine the scanned part and the corresponding scanning orientation in the medical image to be processed based on the extracted features.

[0099] For example, the computer device extracts features from the medical image to be processed using a preset first image recognition algorithm. Based on the extracted features, it determines that the scanned part in the medical image to be processed is the head, and determines that the orientation corresponding to the scanned part is the coronal plane.

[0100] Optionally, the computer device can read the orientation label information in the medical image to be processed, thereby determining the scanning orientation of the corresponding scanned part in the medical image. The orientation label information can include any one of coronal, sagittal, and transverse views. After reading the scanning orientation of the part to be scanned, the computer device can input the medical image to be processed into a preset second image recognition algorithm to identify the scanned part in the medical image.

[0101] For example, the computer device determines that the scanning orientation of the scanned area corresponding to the medical image to be processed is coronal by reading the orientation label information in the medical image to be processed. The computer device inputs the medical image to be processed into a preset second image recognition algorithm corresponding to the coronal position, and extracts features from the medical image to be processed by the preset scanned area image recognition algorithm corresponding to the coronal position, thereby determining that the scanned area corresponding to the medical image to be processed is the brain.

[0102] Optionally, the computer device can first input the medical image to be processed into a preset third image recognition algorithm to identify the scanned area in the medical image. After the computer device determines the scanned area corresponding to the medical image, it can read the orientation label information in the medical image to determine the scanning orientation of the scanned area. The orientation label information can include any one of coronal, sagittal, and transverse views.

[0103] In this embodiment, the computer device identifies the scanned area and scanning direction in the medical image to be processed, enabling the computer device to determine a target artifact recognition model corresponding to the scanned area and scanning direction. Since different scanned areas and scanning directions correspond to different types of artifacts, the above method allows each scanning direction of the scanned area in the medical image to correspond to the target artifact recognition model, thereby improving the accuracy of artifact recognition in the medical image. Furthermore, identifying the scanned area and scanning direction in the medical image also enables the computer device to determine a target artifact severity recognition model corresponding to the scanned area and scanning direction. Since different scanned areas and scanning directions correspond to different types of artifacts, and the motion sensitivity also varies, the above method allows each scanning direction of the scanned area in the medical image to correspond to the target artifact severity recognition model, thereby improving the accuracy of artifact severity recognition in the medical image.

[0104] In one optional implementation of this application, before inputting the medical image to be processed into the target artifact recognition model, it is necessary to determine the target artifact recognition model corresponding to the medical image to be processed based on the scanned area and scanning direction in the medical image to be processed. Therefore, it is necessary to identify the scanned area and scanning orientation in the medical image to be processed. Figure 4 As shown, identifying the scanned area and scanning orientation in the medical image to be processed may include the following steps:

[0105] First, it should be noted that different scanned areas exhibit different types of artifacts, and these areas also have varying degrees of motion sensitivity. Therefore, before inputting the medical image to be processed into the target artifact recognition model, it is necessary to determine the target artifact recognition model corresponding to the scanned area and scanning direction in the medical image.

[0106] Step 401: The computer device acquires the field strength information of the medical device corresponding to the medical image to be processed.

[0107] Specifically, the clarity of images obtained after scanning by medical devices varies due to differences in their field strength. Therefore, in order to improve the accuracy of identifying the scanned areas and scanning directions in medical images, it is necessary to obtain the field strength information of the medical devices.

[0108] Optionally, the computer device can display an input interface to the user, who then inputs the field strength information of the medical device into the interface, allowing the computer device to obtain the field strength information of the medical device corresponding to the medical image to be processed. Alternatively, if the medical image to be processed contains label information, the computer device can automatically read this label information and directly obtain the field strength information through the label information.

[0109] Optionally, the computer device can use a preset resolution recognition algorithm to perform resolution recognition on the medical image to be processed, thereby determining the resolution of the medical image to be processed, and determining the field strength information of the medical device corresponding to the medical image to be processed based on the resolution of the medical image to be processed.

[0110] For example, taking an MRI scanner as an example, when the MRI field strength is 3.0T, the image obtained after scanning is clearer; when the MRI field strength is 1.5T, the image obtained after scanning is slightly blurry. The computer uses a preset resolution recognition algorithm to identify the resolution of the medical image to be processed, thereby determining the resolution of the medical image. When the resolution of the medical image to be processed is greater than a preset resolution threshold, the corresponding MRI field strength is determined to be 3.0T; when the resolution of the medical image to be processed is less than or equal to the preset resolution, the corresponding MRI field strength is determined to be 1.5T. The preset resolution threshold can be determined by comparing the resolution of multiple images. These multiple images are obtained by scanning the same area with MRI scanners of different field strengths.

[0111] Step 402: The computer device acquires the part model corresponding to the field strength information of the medical device.

[0112] Among them, the site model can identify both the scanned site in the medical image to be processed and the scanning direction corresponding to the scanned site.

[0113] Specifically, the computer equipment's database stores different site models for medical devices with different field strengths, and also stores the correspondence between the medical device's field strength information and the site models. After determining the MRI field strength information corresponding to the medical image to be processed, the computer equipment can search the database for the site model corresponding to the medical device's field strength information for the medical image. Based on the search results, the site model corresponding to the medical device's field strength information is determined.

[0114] For example, after determining that the field strength of the medical device corresponding to the medical image to be processed is 3.0T, the computer device searches the database for the part model corresponding to the field strength of 3.0T. After searching, the computer device determines a second part recognition model corresponding to the field strength of 3.0T. The computer device then calls this part model from the database to identify the scanned part and scanning direction in the medical image to be processed.

[0115] Step 403: The computer device inputs the medical image to be processed into the site model to obtain the scanned site and scanning orientation included in the medical image to be processed.

[0116] Specifically, after the computer equipment determines the site model based on the medical field strength information of the medical image to be processed, the medical image to be processed is input into the site model. The site model extracts features from the medical image to be processed and determines the scanned site and scanning direction included in the medical image to be processed based on the extracted features.

[0117] The part model can be a machine learning network model, or other network models. Machine learning network models can be DNN (Deep Neural Networks), CNN (Convolutional Neural Networks), RNN (Recurrent Neural Networks), etc. When the part model is a CNN, it can be a V-Net model, a U-Net model, a Generative Adversarial Network (GAN) model, etc. This application does not specifically limit the type of part model.

[0118] Optionally, when the part model is a CNN, the part model can include a 50-layer deep convolutional neural network, which includes 4 residual blocks, 49 convolutions and one fully connected layer, and uses the ReLU activation function, with the formula as follows:

[0119] F n =R(W n *F n-1 +B n )

[0120] Where R represents the nonlinear activation function ReLU, W n B n F represents the weights and biases of the convolutional layer in the feature extraction stage, respectively. n-1 F represents the feature map output by the previous convolution. n This represents the output feature map obtained in the current feature extraction stage.

[0121] In this embodiment, due to the different field strengths of medical devices, the clarity of the images obtained after scanning by the medical devices also varies. In the above method, the computer device acquires the field strength information of the medical device corresponding to the medical image to be processed, and acquires the part model corresponding to the field strength information of the medical device, ensuring that the field strength information of the medical device corresponding to the medical image to be processed corresponds to the part model. This ensures the accuracy of the scanned parts and scanning directions included in the medical image to be processed identified by the part model. In addition, the computer device inputs the medical image to be processed into the part model to obtain the scanned parts and scanning directions included in the medical image to be processed. This allows artifacts in the medical image to be identified based on the scanned parts and scanning directions included in the medical image to be processed, thereby improving the accuracy of medical image processing.

[0122] In one alternative implementation of this application, the computer device can determine a target artifact recognition model based on at least one of the following: medical device field strength information, the scanned area, and the scanning orientation.

[0123] Optionally, the computer equipment can determine the target artifact recognition model corresponding to the field strength information of the medical equipment.

[0124] Specifically, the computer equipment's database stores different target artifact recognition models for medical devices with different field strengths, and also stores the correspondence between the medical device's field strength information and the target artifact recognition models. After determining the field strength information of the MRI corresponding to the medical image to be processed, the computer equipment can search the database for the target artifact recognition model corresponding to the medical device's field strength information for the medical image to be processed. Based on the search results, the target artifact recognition model corresponding to the medical device's field strength information is determined.

[0125] Optionally, after determining the scanned area and scanning orientation included in the medical image to be processed, the computer device can determine a target artifact recognition model corresponding to the scanned area and scanning orientation based on the scanned area and scanning orientation included in the medical image to be processed.

[0126] For example, when the scanned area is the chest and the scan direction is coronal, the chest image in the medical image to be processed is easily affected by breathing artifacts due to lung respiration. Optionally, the recognition of breathing artifacts in a coronal chest scan image may be relatively coarse. Therefore, when inputting the medical image to be processed into the target artifact recognition model, it is first necessary to determine the scanned area and scan direction in the medical image to be processed. After determining that the scanned area is the chest and the scan direction is coronal, the medical image to be processed is input into the target artifact recognition model corresponding to the coronal chest scan image.

[0127] Optionally, the computer device database stores various artifact recognition models and the correspondence between the scanned area and scanning direction and the artifact recognition model. Different artifact recognition models focus on identifying different types of artifacts corresponding to different scanned areas and scanning directions. After the computer device determines the scanned area and scanning direction in the medical image to be processed, it can search the database for the correspondence between the scanned area / direction and the artifact recognition model. Based on the found correspondence, the target artifact recognition model corresponding to the scanned area / direction in the medical image to be processed is determined.

[0128] For example, after determining that the scanned area in the medical image to be processed is the chest and the scanning direction is coronal, the computer device can look up the correspondence between the scanned area, scanning direction, and artifact recognition models in the database. Based on the correspondence between the scanned area, scanning direction, and artifact recognition models, the computer device determines the artifact recognition model corresponding to the coronal chest medical image as the third artifact recognition model. The computer device then identifies this third artifact recognition model as the target artifact recognition model corresponding to the scanned area in the medical image to be processed.

[0129] Optionally, the computer device first selects candidate target artifact recognition models that match the field strength information of the medical device. Then, based on the scanned area and scanning direction, the computer device determines the target artifact recognition model that matches the target artifact attribute information from the candidate area models.

[0130] Optionally, the computer device first selects candidate target artifact recognition models that match the scanned area and scanning direction. Then, based on the field strength information of the medical device, the computer device determines the target artifact recognition model that matches the field strength information of the medical device from the candidate area models.

[0131] In this embodiment, the computer device determines the target artifact recognition model based on at least one of the medical device field strength information, the scanned part, and the scanning orientation, ensuring the matching between the target artifact recognition model and the medical device field strength information, as well as the scanned part and the scanning orientation, thereby ensuring the accuracy of the artifact attribute information in the medical image to be processed identified by the target artifact recognition model.

[0132] In one optional implementation of this application, after inputting the medical image to be processed into the target artifact recognition model, it is necessary to determine the target artifact degree recognition model corresponding to the medical image to be processed based on the scanned area and scanning direction in the medical image to be processed. Therefore, it is necessary to identify the scanned area and scanning direction in the medical image to be processed. Figure 5 As shown, identifying the scanned area and scanning orientation in the medical image to be processed may include the following steps:

[0133] Step 501: The computer device acquires the field strength information of the medical device corresponding to the medical image to be processed.

[0134] Specifically, the clarity of images obtained after scanning by medical devices varies due to differences in their field strength. Therefore, in order to improve the accuracy of identifying the scanned areas and scanning directions in medical images, it is necessary to obtain the field strength information of the medical devices.

[0135] Optionally, the computer device can display an input interface to the user, who then inputs the field strength information of the medical device into the interface, allowing the computer device to obtain the field strength information of the medical device corresponding to the medical image to be processed. Alternatively, if the medical image to be processed contains label information, the computer device can automatically read this label information and directly obtain the field strength information through the label information.

[0136] Optionally, the computer device can use a preset resolution recognition algorithm to perform resolution recognition on the medical image to be processed, thereby determining the resolution of the medical image to be processed, and determining the field strength information of the medical device corresponding to the medical image to be processed based on the resolution of the medical image to be processed.

[0137] For example, taking an MRI scanner as an example, when the MRI field strength is 3.0T, the image obtained after scanning is clearer; when the MRI field strength is 1.5T, the image obtained after scanning is slightly blurry. The computer uses a preset resolution recognition algorithm to identify the resolution of the medical image to be processed, thereby determining the resolution of the medical image. When the resolution of the medical image to be processed is greater than a preset resolution threshold, the corresponding MRI field strength is determined to be 3.0T; when the resolution of the medical image to be processed is less than or equal to the preset resolution, the corresponding MRI field strength is determined to be 1.5T. The preset resolution threshold can be determined by comparing the resolution of multiple images. These multiple images are obtained by scanning the same area with MRI scanners of different field strengths.

[0138] Step 502: The computer device determines the part model based on the field strength information of the medical device and the target artifact attribute information.

[0139] Among them, the site model can identify both the scanned site in the medical image to be processed and the scanning direction corresponding to the scanned site.

[0140] Specifically, because medical images generated by medical devices with different field strengths have varying clarity, the corresponding body part models differ depending on the field strength information of the medical device. Furthermore, because different target artifact attribute information corresponds to different scanned areas and scanning directions, the corresponding body part models also differ. Therefore, computer equipment needs to determine the body part model based on the medical device's field strength information and the target artifact attribute information.

[0141] Since the impact of target artifact attribute information varies on different scanned areas and scanned areas in different scanning directions, it is necessary to use a location model to identify the scanned areas and scanning directions in the medical image to be processed before determining the target artifact degree recognition model.

[0142] Optionally, the computer device first selects candidate site models that match the field strength information of the medical device. Then, based on the target artifact attribute information, the computer device determines the site model that matches the target artifact attribute information from the candidate site models.

[0143] Specifically, the computer equipment's database stores different site models for medical devices with different field strengths, and also stores the correspondence between the medical device's field strength information and the site models. After determining the MRI field strength information corresponding to the medical image to be processed, the computer equipment can search the database for the site model corresponding to the medical device's field strength information. Based on the search results, candidate site models corresponding to the medical device's field strength information are determined. After determining the candidate site models, the computer equipment can determine the site model that matches the target artifact attribute information from the candidate site models based on the target artifact attribute information.

[0144] For example, after determining that the field strength of the medical device corresponding to the medical image to be processed is 3.0T, the computer device searches the database for the body part model corresponding to the field strength of 3.0T. After the search, the computer device selects 5 candidate body part models corresponding to the field strength of 3.0T. Based on the target artifact attribute information, the computer device determines the body part model that matches the target artifact attribute information from the 5 candidate body part models.

[0145] Optionally, the computer device first selects candidate site models that match the target artifact attribute information based on the target artifact attribute information. Then, based on the medical device field strength information, the computer device determines the site model that matches the medical device field strength information from the candidate site models.

[0146] Specifically, the computer device's database stores different site models for different target artifact attribute information, and also stores the correspondence between the target artifact attribute information and the site models. After determining the target artifact attribute information in the medical image to be processed, the computer device can search the database for the site model corresponding to the target artifact attribute information. Based on the search results, candidate site models corresponding to the target artifact attribute information are determined. After determining the candidate site models, the computer device can determine the site model that matches the target artifact attribute information from the candidate site models based on the medical device's field strength information.

[0147] Step 503: The computer device inputs the medical image to be processed into the site model to obtain the scanned site and scanning orientation included in the medical image to be processed.

[0148] Specifically, after determining the site model based on the medical equipment's field strength information and target artifact attribute information, the computer equipment inputs the medical image to be processed into the site model. The site model extracts features from the medical image to be processed and determines the scanned site and scanning direction included in the medical image to be processed based on the extracted features.

[0149] The part model can be a machine learning network model, or other network models. Machine learning network models can be DNN (Deep Neural Networks), CNN (Convolutional Neural Networks), RNN (Recurrent Neural Networks), etc. When the part model is a CNN, it can be a V-Net model, a U-Net model, a Generative Adversarial Network (GAN) model, etc. This application does not specifically limit the type of part model.

[0150] Optionally, when the part model is a CNN, the part model can include a 50-layer deep convolutional neural network, which includes 4 residual blocks, 49 convolutions and one fully connected layer, and uses the ReLU activation function, with the formula as follows:

[0151] F n =R(W n *F n-1 +B n )

[0152] Where R represents the nonlinear activation function ReLU, W n B n F represents the weights and biases of the convolutional layer in the feature extraction stage, respectively. n-1 F represents the feature map output by the previous convolution.n This represents the output feature map obtained in the current feature extraction stage.

[0153] In this embodiment, the clarity of the image obtained after scanning by the medical device varies due to differences in the field strength of the medical device. Furthermore, the impact of target artifact attribute information varies depending on the scanned area and the scanning direction. In the above method, the computer device acquires the field strength information of the medical device corresponding to the medical image to be processed. Based on the medical device field strength information and target artifact attribute information, a location model is determined. The medical image to be processed is input into the location model to obtain the scanned area and scanning orientation included in the medical image. This ensures that the medical device field strength information and target artifact attribute information corresponding to the medical image to be processed correspond to the location model, thereby guaranteeing the accuracy of the location model in identifying the scanned area and scanning direction included in the medical image to be processed. Furthermore, by inputting the medical image to be processed into the location model to obtain the scanned area and scanning direction included in the medical image, the computer device can determine the target artifact degree recognition model based on the scanned area and scanning direction included in the medical image to be processed, thereby improving the accuracy of medical image processing.

[0154] In one optional implementation of this application, the computer device can determine a target artifact degree recognition model based on at least one of the following: medical device field strength information, target artifact attribute information, scanned area, and scanning orientation.

[0155] Optionally, since the field strength information of the medical device affects the sharpness of the medical image to be processed, and the sharpness of the medical image to be processed may affect the artifact severity indication information output by the target artifact recognition model, for example, when the field strength information of the medical device is large, the sharpness of the medical image to be processed will be high; when the field strength information of the medical device is small, the sharpness of the medical image to be processed will be low. The same artifact attribute information may have a lower impact on the sharper medical image to be processed, but a higher impact on the sharper medical image to be processed. Therefore, when determining the target artifact severity recognition model, it may be necessary to consider the influence of the medical device field strength information on the target artifact severity recognition model. For example, when the field strength information of the medical device is small, the determined target artifact recognition model is more refined.

[0156] Optionally, since artifacts with different attribute information have varying degrees of impact on the medical image being processed, the target artifact attribute information may also affect the artifact severity indication information output by the target artifact recognition model. For example, due to lung respiration, lung respiration is unavoidable when the medical image being processed is a lung scan image. Therefore, respiration artifacts have a relatively small impact on lung scan images. Thus, when determining the target artifact severity recognition model, it may be necessary to consider the influence of the target artifact attribute information on the target artifact severity recognition model.

[0157] Optionally, since different scanning sites and directions are affected by artifacts to varying degrees, these factors may also influence the artifact severity indication output by the target artifact recognition model. For example, when the scanned area is the brain, due to the brain's intricate structure, even small motion artifacts can affect the brain scan results. However, since respiratory artifacts are unavoidable in the abdomen, their impact on the abdominal scan image is relatively weak. Therefore, when determining the target artifact severity recognition model, it may be necessary to consider the influence of different scanning sites and directions. For instance, when the scanned area is the brain, the corresponding target artifact severity recognition model is more refined.

[0158] Based on the above, computer equipment can determine the target artifact degree recognition model based on any one of the following: medical equipment field strength information, target artifact attribute information, scanned area, and scanning orientation; it can also determine the target artifact degree recognition model based on any two of the following: medical equipment field strength information, target artifact attribute information, scanned area, and scanning orientation; or it can comprehensively consider the influencing factors of medical equipment field strength information, target artifact attribute information, scanned area, and scanning orientation to determine the target artifact degree recognition model.

[0159] In this embodiment, the computer device can determine a target artifact degree recognition model based on at least one of the following: medical device field strength information, target artifact attribute information, scanned area, and scanning orientation. This ensures the target artifact degree recognition model is compatible with the medical device field strength information, target artifact attribute information, scanned area, and scanning orientation, thereby guaranteeing more accurate artifact degree indication information output by the model and improving the accuracy of medical image processing.

[0160] In an optional implementation of this application, the medical image to be processed does not contain label information, and the "computer device acquires the medical device field strength information corresponding to the medical image to be processed" in steps 401 and 501 above may include the following:

[0161] The computer device inputs the medical image to be processed into the field strength recognition model to obtain the field strength information of the medical device.

[0162] Specifically, after acquiring the medical image to be processed, the computer device can input the medical image to be processed into the field strength recognition model in order to determine the field strength information of the medical device corresponding to the medical image.

[0163] Optionally, the field strength recognition model can extract features from the medical image to be processed, and based on the extracted features, identify the sharpness of the medical image to be processed, thereby determining the field strength information of the medical device corresponding to the medical image to be processed.

[0164] The field strength recognition model can be a machine learning network model, or other network models. Machine learning network models can be DNN (Deep Neural Networks), CNN (Convolutional Neural Networks), RNN (Recurrent Neural Networks), etc. When the field strength recognition model is a CNN, it can be a V-Net model, a U-Net model, a Generative Adversarial Network (GAN) model, etc. This application does not specifically limit the type of target artifact recognition model in its embodiments.

[0165] Optionally, when the field strength recognition model is a CNN, the field strength recognition model may include a 50-layer deep convolutional neural network, which includes 4 residual blocks, 49 convolutions and one fully connected layer, and the activation function is the ReLU activation function, with the formula as follows:

[0166] F n =R(W n *F n-1 +B n )

[0167] Where R represents the non-linear activation function ReLU, Wn and Bn represent the weights and biases of the convolutional layer in the feature extraction stage, respectively, Fn-1 represents the feature map output by the previous convolution, and Fn represents the output feature map obtained in the current feature extraction stage.

[0168] In this embodiment, the computer device inputs the medical image to be processed into the field strength recognition model to obtain the field strength information of the medical device. This ensures the accuracy of the obtained field strength information.

[0169] In an optional embodiment of this application, such as Figure 6 As shown, the training process of the above-mentioned target artifact recognition model may include the following steps:

[0170] Step 601: The computer device acquires the first training sample set.

[0171] Specifically, the computer device acquires a first training sample set. The first training sample set includes multiple first training samples, each of which includes a first training sample image and training artifact attribute information corresponding to the first training sample image.

[0172] It should be noted that the types of artifacts vary depending on the scanned area. Therefore, the training sample sets for artifact recognition models differ depending on the scanned area. For example, the training sample set for the chest artifact recognition model only includes various types of chest scan images.

[0173] Optionally, the computer device may obtain the first training sample set from a PACS (Picture Archiving and Communication Systems) server or from the medical imaging device in real time.

[0174] Optionally, to facilitate the recognition of each first training sample image in the first training sample set, after obtaining the first training sample set, the computer device can split each first training sample image along the x-axis, y-axis, and z-axis to obtain two-dimensional cross-sectional images of each first training sample image from each viewpoint, where the x-axis, y-axis, and z-axis correspond to the coronal, sagittal, and transverse views, respectively. The computer device then performs separate training on the coronal, sagittal, and transverse views corresponding to the first training sample images in the first training sample set.

[0175] Step 602: The computer device normalizes the image brightness of each first training sample image in the first training sample set based on the Z score.

[0176] Specifically, in order to ensure the accuracy of the target artifact recognition model trained, and to prevent the computer device from identifying the scanned area in the first training sample image as an artifact or processing the medical image in the first training sample image as the scanned area, the computer device can normalize the image brightness of each first training sample image in the first training sample set based on the Z score.

[0177] The Z-score, also called the standard score, is the difference between a number and the mean, divided by the standard deviation. In statistics, the standard score is the sign of the standard deviation of an observed or measured value from the mean.

[0178] Specifically, the computer device can calculate the image brightness of each first training sample image separately, and calculate the average and standard deviation of the image brightness of the first training sample set based on the image brightness of each first training sample image. The computer device can obtain the normalized image brightness of each first training sample image by dividing the difference between the image brightness of each first training sample image and the average image brightness of the first training sample set by the standard deviation of the image brightness of the first training sample set.

[0179] The computer device normalizes the image brightness of each first training sample image in the first training sample set based on the Z-score, which can ensure that the difference in image brightness between each first training sample image is small. Using the normalized first training sample images to train the target artifact recognition model is beneficial to ensuring the accuracy of the target artifact recognition model.

[0180] Step 603: The computer device trains the artifact recognition network based on the normalized first training sample set to obtain the target artifact recognition model.

[0181] Specifically, after obtaining the normalized first training sample set, the computer device can input the normalized first training sample set into the untrained artifact recognition model to train the artifact recognition model, thereby obtaining the target artifact recognition model.

[0182] Optionally, the artifact recognition model can identify artifacts in each of the first training sample images. First, it identifies each pixel in the first training sample image to determine the scanned area. After determining the scanned area, it identifies other pixels in the first training sample image besides the scanned area to identify artifacts and determine their locations. Then, based on the distribution of artifacts and their corresponding pixel values, it determines the artifact's depth, size, and type. Finally, it outputs the artifact's attribute information.

[0183] Furthermore, during training, the Adam optimizer can be selected to optimize the target artifact recognition model, thereby enabling the target artifact recognition model to converge quickly and have good generalization ability.

[0184] When optimizing the target artifact recognition model using the Adam optimizer, a learning rate can be set for the optimizer. Here, the Learning Rate Range Test (LR Range Test) technique can be used to select the optimal learning rate and set it for the optimizer. The learning rate selection process for this test technique is as follows: First, set the learning rate to a very small value. Then, iterate the target artifact recognition model and the first training sample image data a few times. After each iteration, increase the learning rate and record the training loss for each iteration. Then, plot the LR Range Test graph. An ideal LR Range Test graph typically contains three regions: the first region has a small learning rate and the loss remains relatively constant; the second region shows a decreasing loss and rapid convergence; and the last region has a large learning rate that causes the loss to diverge. Therefore, the learning rate corresponding to the lowest point in the LR Range Test graph can be taken as the optimal learning rate, and this optimal learning rate can be used as the initial learning rate for the Adam optimizer and set for the optimizer.

[0185] In this embodiment, a computer device acquires a first training sample set and normalizes the image brightness of each first training sample image in the first training sample set based on the Z-score. The computer device then trains an artifact recognition network based on the normalized first training sample set to obtain a target artifact recognition model. In this embodiment, the target artifact recognition model is trained using the first training sample set; therefore, the obtained target artifact recognition model is more accurate. Consequently, the attribute information of artifacts identified using the target artifact recognition model is more accurate.

[0186] In an optional embodiment of this application, such as Figure 7 As shown, the training process of the above-mentioned target artifact recognition model may include the following steps:

[0187] Step 701: The computer device acquires the second training sample set.

[0188] Specifically, the computer device acquires a second training sample set. This second training sample set includes multiple second training samples, each of which includes a second training sample image and artifact level indication information corresponding to that image.

[0189] Optionally, the computer device can acquire the second training sample images from a PACS (Picture Archiving and Communication Systems) server, and annotate the artifact level indication information corresponding to the second training sample images using an expert group or machine learning algorithm to obtain the second training sample set; alternatively, it can acquire the second training sample images in real time from medical imaging equipment, and annotate the artifact level indication information corresponding to the second training sample images using an expert group or machine learning algorithm to obtain the second training sample set.

[0190] Step 702: The computer device normalizes the image brightness of each second training sample image in the second training sample set based on the Z score.

[0191] Specifically, in order to ensure the accuracy of the target artifact recognition model trained, and to avoid the computer device misjudging the influence of artifacts on the second training sample image due to the different brightness of each second training sample image, the computer device can normalize the image brightness of each second training sample image in the second training sample set based on the Z score.

[0192] The Z-score, also called the standard score, is the difference between a number and the mean, divided by the standard deviation. In statistics, the standard score is the sign of the standard deviation of an observed or measured value from the mean.

[0193] Specifically, the computer device can calculate the image brightness of each second training sample image separately, and calculate the average and standard deviation of the image brightness of the second training sample set based on the image brightness of each second training sample image. The computer device can obtain the normalized image brightness of each second training sample image by dividing the difference between the image brightness of each second training sample image and the average image brightness of the second training sample set by the standard deviation of the image brightness of the second training sample set.

[0194] The computer device normalizes the image brightness of each first training sample image in the first training sample set based on the Z-score, which can ensure that the difference in image brightness between each second training sample image is small. Using the normalized first training sample images to train the target artifact recognition model is beneficial to ensuring the accuracy of the target artifact recognition model.

[0195] Step 703: The computer device trains the artifact recognition network based on the normalized second training sample set to obtain the target artifact recognition model.

[0196] Specifically, after obtaining the normalized second training sample set, the computer device can input the normalized second training sample set into the untrained artifact recognition model to train the artifact recognition model, thereby obtaining the target artifact recognition model.

[0197] Optionally, the artifact severity recognition model can identify the attribute information of artifacts, and then, based on the attribute information of artifacts and the image information of the scanned part in the second training sample image, determine the degree of influence of artifacts in the second training image on the second training sample image, and output the artifact severity indication information corresponding to the second sample image.

[0198] Computer equipment can classify the artifact severity level based on the degree of influence of artifacts on the second training sample image. The loss function of the artifact severity recognition model can be the cross-entropy loss function.

[0199]

[0200] Where, x i Let i = 1, 2, 3, 4, and let p(x) be the true probability distribution and q(x) be the predicted probability distribution.

[0201] Furthermore, during training, the Adam optimizer can be selected to optimize the target artifact recognition model, thereby enabling the target artifact recognition model to converge quickly and have good generalization ability.

[0202] In this embodiment, the computer device acquires a second training sample set and normalizes the image brightness of each second training sample image in the second training sample set based on the Z-score. The computer device then trains the artifact recognition network based on the normalized second training sample set to obtain a target artifact recognition model. In this embodiment, the target artifact recognition model is trained using the second training sample images and the corresponding artifact indication information, thus ensuring the accuracy of the obtained target artifact recognition model. Consequently, the target artifact indication information obtained using the target artifact recognition model is more accurate.

[0203] In an optional embodiment of this application, such as Figure 8 As shown, the training process for the above-mentioned part model may include the following steps:

[0204] Step 801: The computer device acquires the third training sample set.

[0205] Specifically, the computer device acquires a third training sample set. This third training sample set includes multiple third training samples, each of which includes a third training sample image and the corresponding scanned region and scanning direction.

[0206] Optionally, the computer device can obtain a third training sample set from a PACS (Picture Archiving and Communication Systems) server or from the medical imaging device in real time.

[0207] Step 802: The computer device normalizes the image brightness of each third training sample image in the third training sample set based on the Z score.

[0208] Specifically, to ensure the accuracy of the trained part models and avoid errors in the computer's recognition of the sharpness of each third training sample image due to differences in brightness, which could lead to incorrect identification of the scanned part and scanning direction, the computer can normalize the image brightness of each third training sample image in the third training sample set based on the Z-score.

[0209] Specifically, the computer device can calculate the image brightness of each third training sample image separately, and calculate the average and standard deviation of the image brightness of the third training sample set based on the image brightness of each third training sample image. The computer device can obtain the normalized image brightness of each third training sample image by dividing the difference between the image brightness of each third training sample image and the average image brightness of the third training sample set by the standard deviation of the image brightness of the third training sample set.

[0210] By normalizing the image brightness of each third training sample image in the third training sample set based on the Z-score, the computer equipment can ensure that the difference in image brightness between each third training sample image is small. Using the normalized third training sample images to train the part model helps to ensure the accuracy of the part model.

[0211] Step 803: The computer device trains the part recognition network based on the normalized third training sample set to obtain the part model.

[0212] Specifically, after obtaining the normalized third training sample set, the computer device can input the normalized third training sample set into the part recognition network to train the part recognition network, thereby obtaining the target part model.

[0213] Optionally, the part model can extract the feature information of each third training sample image and determine the scanning direction of the scanned part in each third training sample image based on the collected feature information.

[0214] Furthermore, during training, the Adam optimizer can be selected to optimize the part model, thereby enabling the part model to converge quickly and have good generalization ability.

[0215] When optimizing the part model using the Adam optimizer, a learning rate can be set for the optimizer. Here, the Learning Rate Range Test (LR Range Test) technique can be used to select the optimal learning rate and set it for the optimizer. The learning rate selection process for this test technique is as follows: First, set the learning rate to a very small value. Then, iterate the part model and the third training sample image data a few times. After each iteration, increase the learning rate and record the training loss for each iteration. Then, plot the LR Range Test graph. An ideal LR Range Test graph typically contains three regions: the first region has a small learning rate and the loss remains relatively constant; the second region shows a decreasing loss and rapid convergence; and the last region has a large learning rate that causes the loss to diverge. Therefore, the learning rate corresponding to the lowest point in the LR Range Test graph can be taken as the optimal learning rate, and this optimal learning rate can be used as the initial learning rate for the Adam optimizer and set for the optimizer.

[0216] In this embodiment, the computer device acquires a third training sample set and normalizes the image brightness of each third training sample image in the third training sample set based on the Z-score. The computer device then trains an untrained part recognition network based on the normalized third training sample set to obtain a part model. In this embodiment, the part model is obtained through training on the normalized third training sample set, ensuring the accuracy of the obtained part model. Therefore, only with an accurate part model can the accuracy of the artifact attribute information identified by the target artifact recognition model be guaranteed.

[0217] It should be noted that the first training sample set, the second training sample set, and the third training sample set in the above embodiments can be generated from the same medical images with different annotations, or they can be generated from different medical images with different annotations. This application does not specifically limit the first training sample set, the second training sample set, and the third training sample set.

[0218] To better illustrate the medical image processing method provided in this application, this application provides an embodiment that explains the overall process of the medical image processing method, such as... Figure 9 As shown, the method includes:

[0219] Step 901: The computer device acquires the first training sample set.

[0220] Step 902: The computer device normalizes the image brightness of each first training sample image in the first training sample set based on the Z score.

[0221] Step 903: The computer device trains the artifact recognition network based on the normalized first training sample set to obtain the target artifact recognition model.

[0222] Step 904: The computer device acquires the second training sample set.

[0223] Step 905: The computer device normalizes the image brightness of each second training sample image in the second training sample set based on the Z score.

[0224] Step 906: The computer device trains the artifact recognition network based on the normalized second training sample set to obtain the target artifact recognition model.

[0225] Step 907: The computer device acquires the third training sample set.

[0226] Step 908: The computer device normalizes the image brightness of each third training sample image in the third training sample set based on the Z score.

[0227] Step 909: The computer device trains the part recognition network based on the normalized third training sample set to obtain the part model.

[0228] Step 910: The computer device reads the tag information of the medical image to be processed, thereby obtaining the field strength information of the medical device corresponding to the medical image to be processed.

[0229] Step 911: The computer device acquires the part model corresponding to the field strength information of the medical device.

[0230] Step 912: The computer device inputs the medical image to be processed into the site model to obtain the scanned site and scanning direction included in the medical image to be processed.

[0231] Step 913: The computer device determines the target artifact recognition model based on the field strength information of the medical device, the scanned area, and the scanning direction.

[0232] Step 914: The computer device inputs the medical image to be processed into the target artifact recognition model to obtain the target artifact attribute information output by the target artifact recognition model.

[0233] Step 915: Determine the target artifact degree recognition model based on at least one of the following: medical device field strength information, target artifact attribute information, scanned area, and scanning orientation.

[0234] Step 916: The computer device inputs the medical image to be processed and the target artifact attribute information into the target artifact degree recognition model to obtain the artifact degree indication information output by the target artifact degree recognition model.

[0235] Step 917: If the impact of artifacts in the medical image to be processed on the image quality of the medical image to be processed is greater than the preset artifact impact threshold, the computer device outputs a prompt message.

[0236] To better illustrate the medical image processing method provided in this application, this application provides an embodiment of another medical image processing method to explain its overall process, such as... Figure 10 As shown, the method includes:

[0237] Step 1001: The computer device acquires the first training sample set.

[0238] Step 1002: The computer device normalizes the image brightness of each first training sample image in the first training sample set based on the Z score.

[0239] Step 1003: The computer device trains the artifact recognition network based on the normalized first training sample set to obtain the target artifact recognition model.

[0240] Step 1004: The computer device acquires the second training sample set.

[0241] Step 1005: The computer device normalizes the image brightness of each second training sample image in the second training sample set based on the Z score.

[0242] Step 1006: The computer device trains the artifact recognition network based on the normalized second training sample set to obtain the target artifact recognition model.

[0243] Step 1007: The computer device acquires the third training sample set.

[0244] Step 1008: The computer device normalizes the image brightness of each third training sample image in the third training sample set based on the Z score.

[0245] Step 1009: The computer device trains the part recognition network based on the normalized third training sample set to obtain the part model.

[0246] Step 1010: The computer device reads the tag information of the medical image to be processed, thereby obtaining the field strength information of the medical device corresponding to the medical image to be processed.

[0247] Step 1011: The computer device acquires the target artifact recognition model corresponding to the field strength information of the medical device.

[0248] Step 1012: The computer device inputs the medical image to be processed into the target artifact recognition model to obtain the target artifact attribute information output by the target artifact recognition model.

[0249] Step 1013: The computer device determines the part model based on the field strength information of the medical device and the target artifact attribute information.

[0250] Step 1014: The computer device inputs the medical image to be processed into the site model to obtain the scanned site and scanning direction included in the medical image to be processed.

[0251] Step 1015: Determine the target artifact degree recognition model based on at least one of the following: medical device field strength information, target artifact attribute information, scanned area, and scanning orientation.

[0252] Step 1016: The computer device inputs the medical image to be processed and the target artifact attribute information into the target artifact degree recognition model to obtain the artifact degree indication information output by the target artifact degree recognition model.

[0253] Step 1017: If the impact of artifacts in the medical image to be processed on the image quality of the medical image to be processed is greater than the preset artifact impact threshold, the computer device outputs a prompt message.

[0254] It should be understood that, although Figure 1 as well as Figure 3-10 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 1-8 At least some of the steps in the process may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but may be executed at different times. The execution order of these steps or stages is not necessarily sequential, but may be executed in turn or alternately with other steps or at least some of the steps or stages in other steps.

[0255] In one embodiment of this application, such as Figure 11 As shown, a medical image processing device 1100 is provided, including: a first input module 1101, a second input module 1102, and an output module 1103, wherein:

[0256] The first input module 1101 is used to input the medical image to be processed obtained by scanning with medical equipment into the target artifact recognition model, and obtain the target artifact attribute information output by the target artifact recognition model. The target artifact attribute information is used to indicate the attribute characteristics of artifacts in the medical image to be processed.

[0257] The second input module 1102 is used to input the medical image to be processed and the target artifact attribute information into the target artifact degree recognition model to obtain the artifact degree indication information output by the target artifact degree recognition model. The artifact degree indication information is used to indicate the degree of influence of artifacts in the medical image to be processed on the image quality of the medical image to be processed.

[0258] The output module 1103 is used to output a prompt message when the influence of artifacts in the medical image to be processed on the image quality of the medical image to be processed is greater than or equal to a preset artifact influence threshold. The prompt message is used to prompt the user to confirm whether to accept the artifacts in the medical image to be processed and whether to rescan the corresponding scanned area of ​​the medical image to be processed.

[0259] In one embodiment of this application, such as Figure 12 As shown, the aforementioned medical image processing device 1100 further includes: a recognition module 1104, wherein:

[0260] The recognition module 1104 is used to identify the scanned area and scanning orientation in the medical image to be processed.

[0261] In one embodiment of this application, the aforementioned identification module 1104 is specifically used for: acquiring the field strength information of the medical device corresponding to the medical image to be processed; acquiring the site model corresponding to the field strength information of the medical device; and inputting the medical image to be processed into the site model to obtain the scanned site and scanning orientation included in the medical image to be processed.

[0262] In one embodiment of this application, the above-mentioned identification module 1104 is specifically used for: acquiring the field strength information of the medical device corresponding to the medical image to be processed; determining the part model based on the field strength information of the medical device and the target artifact attribute information; inputting the medical image to be processed into the part model to obtain the scanned part and scanning orientation included in the medical image to be processed.

[0263] In one embodiment of this application, such as Figure 13 As shown, the aforementioned medical image processing device 1100 further includes: a determining module 1105, wherein:

[0264] The determination module 1105 is used to determine the target artifact degree recognition model based on at least one of the following: medical device field strength information, target artifact attribute information, scanned area and scanning orientation.

[0265] In one embodiment of this application, such as Figure 14 As shown, the aforementioned medical image processing device 1100 further includes:

[0266] The third acquisition module 1106, the first processing module 1107, and the first training module 1108, wherein:

[0267] The first acquisition module 1106 is used to acquire a first training sample set, which includes multiple first training samples. Each first training sample includes a first training sample image and training artifact attribute information corresponding to the first training sample image.

[0268] The first processing module 1107 is used to normalize the image brightness of each first training sample image in the first training sample set based on the Z score.

[0269] The first training module 1108 is used to train the artifact recognition network based on the normalized first training sample set to obtain the target artifact recognition model.

[0270] In one embodiment of this application, such as Figure 15 As shown, the aforementioned medical image processing device 1100 further includes:

[0271] The fourth acquisition module 1109, the second processing module 1110, and the second training module 1111, wherein:

[0272] The second acquisition module 1109 is used to acquire a second training sample set, which includes multiple second training samples. Each second training sample includes a second training sample image and artifact degree indication information corresponding to the second training sample image.

[0273] The second processing module 1110 is used to normalize the image brightness of each second training sample image in the second training sample set based on the Z score.

[0274] The second training module 1111 is used to train the artifact recognition network based on the normalized second training sample set to obtain the target artifact recognition model.

[0275] In one embodiment of this application, such as Figure 16 As shown, the aforementioned medical image processing device 1100 further includes: a third acquisition module 1112, a third processing module 1113, and a third training module 1114, wherein:

[0276] The third acquisition module 1112 is used to acquire a third training sample set, which includes multiple third training samples. Each third training sample includes a third training sample image and a scanned part and scanning direction corresponding to the third training sample image.

[0277] The third processing module 1113 is used to normalize the image brightness of each third training sample image in the third training sample set based on the Z score.

[0278] Based on the Z-score, the image brightness of each third training sample image in the third training sample set is normalized.

[0279] The third training module 1114 is used to train the part recognition network based on the normalized third training sample set to obtain the part model.

[0280] Specific limitations regarding the medical image processing device can be found in the limitations of the medical image processing method described above, and will not be repeated here. Each module in the aforementioned medical image processing device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0281] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 17 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores medical image processing data. The network interface communicates with external terminals via a network connection. When executed by the processor, the computer program implements a medical image processing method.

[0282] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 18As shown, the computer device includes a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a medical image processing method. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0283] Those skilled in the art will understand that Figure 17 and Figure 18 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0284] In one embodiment of this application, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to perform the following steps: inputting a medical image to be processed into a target artifact recognition model to obtain target artifact attribute information output by the target artifact recognition model, the target artifact attribute information being used to indicate the attribute characteristics of artifacts in the medical image to be processed; inputting the medical image to be processed and the target artifact attribute information into a target artifact degree recognition model to obtain artifact degree indication information output by the target artifact degree recognition model, the artifact degree indication information being used to indicate the degree of influence of artifacts in the medical image to be processed on the image quality of the medical image to be processed; if the degree of influence of artifacts in the medical image to be processed on the image quality of the medical image to be processed is greater than or equal to a preset artifact influence degree threshold, then outputting a prompt message, the prompt message being used to prompt the user to confirm whether to accept the artifacts in the medical image to be processed, and whether it is necessary to rescan the scanned area corresponding to the medical image to be processed.

[0285] In one embodiment of this application, when the processor executes the computer program, it further performs the following steps: identifying the scanned area and scanning orientation in the medical image to be processed.

[0286] In one embodiment of this application, when the processor executes the computer program, it further performs the following steps: acquiring the field strength information of the medical device corresponding to the medical image to be processed; acquiring the part model corresponding to the field strength information of the medical device; inputting the medical image to be processed into the part model to obtain the scanned part and scanning orientation included in the medical image to be processed.

[0287] In one embodiment of this application, when the processor executes the computer program, it further performs the following steps: acquiring the field strength information of the medical device corresponding to the medical image to be processed; determining the part model based on the field strength information of the medical device and the target artifact attribute information; inputting the medical image to be processed into the part model to obtain the scanned part and scanning orientation included in the medical image to be processed.

[0288] In one embodiment of this application, when the processor executes the computer program, it further implements the following steps: determining a target artifact degree recognition model based on at least one of the medical device field strength information, target artifact attribute information, scanned area and scanning orientation.

[0289] In one embodiment of this application, when the processor executes the computer program, it further implements the following steps: obtaining a first training sample set, the first training sample set including multiple first training samples, each first training sample including a first training sample image and training artifact attribute information corresponding to the first training sample image; normalizing the image brightness of each first training sample image in the first training sample set based on the Z score; and training the artifact recognition network based on the normalized first training sample set to obtain a target artifact recognition model.

[0290] In one embodiment of this application, when the processor executes the computer program, it further implements the following steps: obtaining a second training sample set, the second training sample set including multiple second training samples, each second training sample including a second training sample image and artifact degree indication information corresponding to the second training sample image; normalizing the image brightness of each second training sample image in the second training sample set based on the Z score; and training the artifact degree recognition network based on the normalized second training sample set to obtain a target artifact degree recognition model.

[0291] In one embodiment of this application, when the processor executes the computer program, it further implements the following steps: obtaining a third training sample set, the third training sample set including multiple third training samples, each third training sample including a third training sample image and a scanned part and scanning direction corresponding to the third training sample image; normalizing the image brightness of each third training sample image in the third training sample set based on the Z score; and training the part recognition network based on the normalized third training sample set to obtain a part model.

[0292] In one embodiment of this application, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, it performs the following steps:

[0293] The medical image to be processed is input into the target artifact recognition model to obtain the target artifact attribute information output by the target artifact recognition model. The target artifact attribute information is used to indicate the attribute characteristics of artifacts in the medical image to be processed. The medical image to be processed and the target artifact attribute information are input into the target artifact degree recognition model to obtain the artifact degree indication information output by the target artifact degree recognition model. The artifact degree indication information is used to indicate the degree of influence of artifacts in the medical image to be processed on the image quality of the medical image to be processed. If the degree of influence of artifacts in the medical image to be processed on the image quality of the medical image to be processed is greater than or equal to the preset artifact influence degree threshold, a prompt message is output. The prompt message is used to prompt the user to confirm whether to accept the artifacts in the medical image to be processed and whether to rescan the corresponding scanned area of ​​the medical image to be processed.

[0294] In one embodiment of this application, when the computer program is executed by the processor, it further performs the following steps: identifying the scanned area and scanning orientation in the medical image to be processed.

[0295] In one embodiment of this application, when the computer program is executed by the processor, it further performs the following steps: obtaining the field strength information of the medical device corresponding to the medical image to be processed; obtaining the part model corresponding to the field strength information of the medical device; inputting the medical image to be processed into the part model to obtain the scanned part and scanning orientation included in the medical image to be processed.

[0296] In one embodiment of this application, when the computer program is executed by the processor, it further performs the following steps: obtaining the field strength information of the medical device corresponding to the medical image to be processed; determining the part model based on the field strength information of the medical device and the target artifact attribute information; inputting the medical image to be processed into the part model to obtain the scanned part and scanning orientation included in the medical image to be processed.

[0297] In one embodiment of this application, when the computer program is executed by the processor, it further implements the following steps: determining a target artifact degree recognition model based on at least one of the medical device field strength information, target artifact attribute information, scanned area and scanning orientation.

[0298] In one embodiment of this application, when the computer program is executed by the processor, it further performs the following steps: obtaining a first training sample set, the first training sample set including multiple first training samples, each first training sample including a first training sample image and training artifact attribute information corresponding to the first training sample image; normalizing the image brightness of each first training sample image in the first training sample set based on the Z score; and training the artifact recognition network based on the normalized first training sample set to obtain a target artifact recognition model.

[0299] In one embodiment of this application, when the computer program is executed by the processor, it further implements the following steps: obtaining a second training sample set, the second training sample set including multiple second training samples, each second training sample including a second training sample image and artifact degree indication information corresponding to the second training sample image; normalizing the image brightness of each second training sample image in the second training sample set based on the Z score; and training the artifact degree recognition network based on the normalized second training sample set to obtain a target artifact degree recognition model.

[0300] In one embodiment of this application, when the computer program is executed by the processor, it further implements the following steps: obtaining a third training sample set, the third training sample set including multiple third training samples, each third training sample including a third training sample image and a scanned part and scanning direction corresponding to the third training sample image; normalizing the image brightness of each third training sample image in the third training sample set based on the Z score; and training the part recognition network based on the normalized third training sample set to obtain a part model.

[0301] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, 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 storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0302] The technical features of the above embodiments can be combined in any way. For the sake of brevity, 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.

[0303] The above embodiments merely illustrate several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A medical image processing method, characterized in that, The method includes: The medical image to be processed is input into the target artifact recognition model to obtain the target artifact attribute information output by the target artifact recognition model. The medical image to be processed is obtained by sequentially executing multiple categories of scanning protocols. The target artifact attribute information is used to indicate the attribute characteristics of artifacts in the medical image to be processed. The target artifact attribute information includes at least one of the following: artifact size information, artifact position information, artifact quantity information, and artifact type information. The medical image to be processed and the target artifact attribute information are input into the target artifact degree recognition model to obtain the artifact degree indication information output by the target artifact degree recognition model. The artifact degree indication information is used to indicate the degree of influence of the artifacts in the medical image to be processed on the image quality of the medical image to be processed. The artifact degree indication information is divided into levels representing normal, slight influence, moderate influence and severe influence, respectively. If the impact of artifacts in the medical image to be processed on the image quality of the medical image is greater than or equal to a preset artifact impact threshold, a prompt message is output on the scanning interface, and a recommended scanning protocol is generated in the recommended protocol area of ​​the scanning interface according to the order and function of multiple categories of scanning protocols set in the scanning protocol execution area of ​​the scanning interface. The prompt message includes the scanning protocol corresponding to the medical image affected by artifacts, and the prompt message is used to prompt the user to confirm whether to accept the artifacts in the medical image to be processed and whether to rescan the scanned area corresponding to the medical image to be processed. The recommended scanning protocol is different in category or timing from the multiple categories of scanning protocols set in the scanning protocol execution area, but the image display effect is the same.

2. The method according to claim 1, characterized in that, Before or after inputting the medical image to be processed into the target artifact recognition model, the method further includes: Identify the scanned area and scanning orientation in the medical image to be processed.

3. The method according to claim 2, characterized in that, Before inputting the medical image to be processed into the target artifact recognition model, the step of identifying the scanned area and scanning orientation in the medical image to be processed includes: Obtain the field strength information of the medical device corresponding to the medical image to be processed; Obtain the location model corresponding to the field strength information of the medical device; The medical image to be processed is input into the body part model to obtain the scanned body part and the scanning orientation included in the medical image to be processed.

4. The method according to claim 2, characterized in that, After inputting the medical image to be processed into the target artifact recognition model, the step of identifying the scanned area and scanning orientation in the medical image to be processed includes: Obtain the field strength information of the medical device corresponding to the medical image to be processed; The location model is determined based on the field strength information of the medical device and the target artifact attribute information; The medical image to be processed is input into the body part model to obtain the scanned body part and the scanning orientation included in the medical image to be processed.

5. The method according to claim 3 or 4, characterized in that, The target artifact recognition model is determined in the following way: The target artifact degree recognition model is determined based on at least one of the following: the field strength information of the medical device, the target artifact attribute information, the scanned area, and the scanning orientation.

6. The method according to claim 1, characterized in that, The training process of the target artifact recognition model is as follows: Obtain a first training sample set, which includes multiple first training samples, each of which includes a first training sample image and training artifact attribute information corresponding to the first training sample image; Based on the Z-score, the image brightness of each of the first training sample images in the first training sample set is normalized. The artifact recognition network is trained based on the first training sample set after normalization to obtain the target artifact recognition model.

7. The method according to claim 1, characterized in that, The training process of the target artifact recognition model is as follows: Obtain a second training sample set, which includes multiple second training samples, each of which includes a second training sample image and artifact degree indication information corresponding to the second training sample image; Based on the Z-score, the image brightness of each second training sample image in the second training sample set is normalized. The artifact recognition network is trained based on the normalized second training sample set to obtain the target artifact recognition model.

8. A medical image processing device, characterized in that, The device includes: The first input module is used to input the medical image to be processed into the target artifact recognition model to obtain the target artifact attribute information output by the target artifact recognition model. The medical image to be processed is obtained by sequentially executing multiple categories of scanning protocols. The target artifact attribute information is used to indicate the attribute characteristics of artifacts in the medical image to be processed. The target artifact attribute information includes at least one of the following: artifact size information, artifact position information, artifact quantity information, and artifact type information. The second input module is used to input the medical image to be processed and the target artifact attribute information into the target artifact degree recognition model to obtain the artifact degree indication information output by the target artifact degree recognition model. The artifact degree indication information is used to indicate the degree of influence of the artifacts in the medical image to be processed on the image quality of the medical image to be processed. The artifact degree indication information is divided into levels representing normal, slight influence, moderate influence and severe influence, respectively. The output module is configured to output a prompt message on the scanning interface when the impact of artifacts in the medical image to be processed on the image quality of the medical image to be processed is greater than or equal to a preset artifact impact threshold. It also generates a recommended scanning protocol in the recommended protocol area of ​​the scanning interface based on the order and functionality of multiple categories of scanning protocols set in the scanning protocol execution area of ​​the scanning interface. The prompt message includes the scanning protocol corresponding to the medical image affected by artifacts. The prompt message is used to prompt the user to confirm whether to accept the artifacts in the medical image to be processed and whether to rescan the scanned area corresponding to the medical image to be processed. The recommended scanning protocol differs in category or timing from the multiple categories of scanning protocols set in the scanning protocol execution area, but achieves the same image display effect.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images