Medical equipment operation status monitoring method and system based on artificial intelligence
Through artificial intelligence, the metal artifacts in MRI devices are identified, and the problem of the impact of MRI devices is solved, and the accurate judgment and automated adjustment of the impact of metal artifacts is achieved, which improves the accuracy of diagnosis.
Patent Information
- Application Number
- CN202410096847.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-01-23
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2044-01-23
AI Technical Summary
During operation, MRI equipment may produce metal artifacts, affecting imaging quality, and it is difficult for the prior art to accurately judge the impact of metal artifacts on the lesion area, resulting in difficulty in diagnosis.
Using an artificial intelligence-based approach, metal artifacts in MRI devices are extracted through U-Net models, their location and the degree of impact on the imaging area are determined, and scanning parameters or patient positions are adjusted according to preset thresholds to reduce artifacts Influence.
It realizes automated judgment of the impact of metal artifacts on the imaging area, avoids the inaccuracy of naked-eye observation, and improves the operating status monitoring accuracy and diagnostic accuracy of MRI equipment.
Smart Images

Figure CN118071686B_ABST
Abstract
Description
Technical Field
[0001] The present application belongs to the field of medical equipment monitoring technology, and more specifically, to a method and system for monitoring the operating status of medical equipment based on artificial intelligence. Background Art
[0002] Monitoring the operating status of medical equipment plays a vital role in the modern medical environment. With the continuous development of medical technology, medical equipment plays an increasingly important role in the diagnosis and treatment process. However, various problems may occur during the use of these devices, such as equipment failure, performance degradation, etc. These problems may affect the normal operation of medical equipment. Therefore, real-time monitoring of the operating status of medical equipment is of great significance to ensure the normal operation of the equipment and improve the quality of medical services.
[0003] Among many medical devices, MRI (magnetic resonance imaging) equipment is a very important device. It can provide high-quality images and provide doctors with accurate diagnosis basis. At present, the operation status monitoring system of MRI equipment is very complete, but it ignores the monitoring of images obtained by MRI equipment, that is, metal artifacts may be generated during the operation of MRI equipment, thereby affecting the imaging quality. The current MRI equipment is committed to reducing the impact of metal artifacts. However, when metal artifacts appear, it is difficult for operators to judge whether the impact of metal artifacts on the image is serious by naked eyes. Therefore, how to effectively detect the impact of metal artifacts on the imaging of the lesion area is an important issue currently facing the operation status monitoring of MRI equipment. Summary of the invention
[0004] The present invention provides a medical equipment operation status monitoring method and system based on artificial intelligence, which intends to recognize the image of MRI equipment, analyze the impact caused by metal artifacts, and thus remind operators to take corresponding measures.
[0005] The method for monitoring the operation status of medical equipment based on artificial intelligence includes the following steps:
[0006] S1. Acquire an image in an MRI device, and segment metal artifacts in the image based on the acquired image;
[0007] S2. Determine the position of the metal artifact based on the extracted metal artifact;
[0008] S3. Based on the position of the metal artifact and the area to be imaged of the patient, determining whether the metal artifact affects the area to be imaged;
[0009] S4. If the metal artifact affects the imaging area, the degree of impact is determined; the determined degree of impact is compared with the impact threshold; if the degree of impact is lower than the impact threshold, it can be ignored; if the degree of impact exceeds the impact threshold, an alarm is issued;
[0010] S5. Adjust the patient's position or scan parameters based on the alarm information to reduce the impact of metal artifacts.
[0011] The present invention extracts the position of the metal artifact, and determines whether the metal artifact will affect the area to be imaged based on the position of the metal artifact and the position between the area to be imaged of the patient (i.e., the patient lesion area to be acquired); if the impact exceeds the set threshold, an alarm is issued to remind the operator to reduce the impact of the metal artifact by changing the patient's position or adjusting the scanning parameters; based on the above technical means, the impact caused by the metal artifact can be automatically judged, avoiding the problem of inaccurate judgment by naked eye observation.
[0012] Preferably, the U-Net model is used in S1 to extract metal artifacts;
[0013] The U-Net model includes an encoder, a decoder, a final layer, and a skip connection;
[0014] The encoder includes a convolution layer and a pooling layer; the convolution layer includes two 3*3 convolutions, and a ReLU activation function is set after each convolution; the pooling layer uses a 2*2 maximum pooling operation to reduce the spatial dimension of the feature map. After each pooling, the dimension of the feature map is halved; and the number of channels is doubled after each pooling layer;
[0015] The decoder includes an upsampling layer, a splicing layer, and a convolution layer. The upsampling layer uses transposed convolution or upsampling to increase the spatial dimension of the feature map; the splicing layer splices the upsampled feature map with the feature map of the corresponding encoder part; the convolution layer uses two 3*3 convolutions, and a ReLU activation function is set after each convolution to restore the features;
[0016] The final layer includes a 1*1 convolution to reduce the number of channels to the number of target categories; and the Sigmoid function is used as the activation function in the final layer;
[0017] The jump connection is a direct connection, and the feature map before each pooling layer in the encoder is directly concatenated with the output of the corresponding upsampling layer of the decoder.
[0018] Preferably, the calculation formula in S2 is as follows:
[0019]
[0020]
[0021] Where: n is the number of pixels of metal artifacts; x i and i are the horizontal and vertical coordinates of the ith metal artifact pixel; v i is the value of the i-th pixel; C x and C y Let the abscissa and ordinate be the centroid.
[0022] Preferably, S3 comprises the following steps:
[0023] S3.1: Determine the area to be imaged and identify the center point of the area to be imaged;
[0024] S3.2: Calculate the distance between the center point of the imaged area and the centroid of the metal artifact:
[0025]
[0026] Where: D is the distance between the centroid of the metal artifact and the center of the imaged area; C x and C y is the horizontal and vertical coordinates of the center of mass; T x and T y is the coordinate of the center point of the area to be imaged;
[0027] S3.3: Compare the calculated distance between the area to be imaged and the centroid of the metal artifact with a preset first distance threshold. If D is less than or equal to the first distance threshold, the metal artifact will affect the area to be imaged; if D is greater than the first distance threshold, the influence of the metal artifact on the area to be imaged is negligible.
[0028] Preferably, the S3.1 comprises the following steps:
[0029] Obtain the lesion area that needs to be examined according to the doctor's order information;
[0030] Based on the lesion area to be examined, extracting the area to be imaged by using the U-Net model, wherein the original image used to extract the area to be imaged by using the U-Net model is the same as the original image used to segment the metal artifacts;
[0031] Based on the extracted area to be imaged, determine the center point of the area to be imaged:
[0032]
[0033]
[0034] Where: n is the number of pixels in the area to be imaged; x i ′ and yi ′ is the horizontal and vertical coordinates of the i-th area to be imaged; v i is the value of the i-th pixel; T x and T y Make the horizontal and vertical coordinates of the center point.
[0035] Preferably, in S3, the shortest distance between the metal artifact area and all points in the area to be imaged is used to determine whether the metal artifact affects the local area to be imaged, and the specific steps are as follows:
[0036] Define area: determine the boundary between the metal artifact and the area to be imaged, and extract all two-dimensional coordinate points in the metal artifact and the area to be imaged;
[0037] Calculate the distance: For each point A in the metal artifact area i (x i ,y i ), calculate its relationship with each point B in the area to be imaged j (x j ,y j ) is:
[0038]
[0039] Where: D ij Indicates point A in the metal artifact area i and the midpoint B of the area to be imaged j The Euclidean distance of
[0040] Find the shortest distance: traverse all distances D ij , find the shortest distance D min ;
[0041] If D min If D is less than or equal to the second distance threshold, the metal artifact affects the imaging area; min If the distance is greater than the second distance threshold, the influence of the metal artifact on the imaging area is negligible.
[0042] Preferably, S4 comprises the following steps:
[0043] S4.1: Select multiple ROIs in the area to be imaged, and for each ROI, calculate the average signal intensity within the ROI:
[0044]
[0045] Where: MU i Represents the signal mean of each ROI; N i Indicates the total number of pixels in the ROI; S p represents the signal intensity of the pth pixel in the ROI;
[0046] S4.2: Calculate the mean of the overall signal:
[0047]
[0048] Where: N represents the number of ROIs; MU i ′ represents the signal mean of all ROIs;
[0049] S4.3: Calculate the standard deviation of the overall signal:
[0050]
[0051] S4.4: Compare the standard deviation of the overall signal with a preset threshold T. If sigma is greater than T, it is considered that the metal artifact has a significant impact on the imaging area.
[0052] Preferably, S4 comprises the following steps:
[0053] Select the ROI closest to the metal artifact in the imaging area as the reference area;
[0054] Calculate the signal mean of the reference ROI:
[0055]
[0056] Where: MU c represents the signal mean of the reference ROI; S p Represents the signal value of each pixel; N represents the number of pixels in the reference ROI;
[0057] Calculate the signal mean of other ROIs: Calculate the signal mean MU for other ROIs in other areas to be imaged i ;
[0058] Calculate signal non-uniformity: Use the signal mean of the reference ROI to calculate the signal non-uniformity of all ROIs in the area to be imaged:
[0059]
[0060] Where: sigma represents the signal non-uniformity of all ROIs in the area to be imaged, that is, the standard deviation;
[0061] Determine the degree of influence: If the calculated sigma′ is greater than a preset threshold value T′, it is considered that the metal artifact has a significant influence on the imaging area.
[0062] An artificial intelligence-based medical device operation status monitoring system includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the artificial intelligence-based medical device operation status monitoring method described in the present invention is implemented.
[0063] The beneficial effects of the present invention include:
[0064] The present invention extracts the position of the metal artifact, and determines whether the metal artifact will affect the area to be imaged based on the position of the metal artifact and the position between the area to be imaged of the patient (i.e., the patient lesion area to be acquired); if the impact exceeds the set threshold, an alarm is issued to remind the operator to reduce the impact of the metal artifact by changing the patient's position or adjusting the scanning parameters; based on the above technical means, the impact caused by the metal artifact can be automatically judged, avoiding the problem of inaccurate judgment by naked eye observation. BRIEF DESCRIPTION OF THE DRAWINGS
[0065] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0066] Figure 1 This is a flowchart of the overall steps of Embodiments 1 and 2 of the present invention.
[0067] Figure 2 This is a specific step block diagram of step 3 provided in Example 1 of the present invention.
[0068] Figure 3 This is a specific step block diagram of step 3 provided in Example 2 of the present invention.
[0069] Figure 4 This is a specific step block diagram of step 4 provided in Example 1 of the present invention.
[0070] Figure 5 This is a specific step flowchart of step 4 provided in Example 2 of the present invention. DETAILED DESCRIPTION
[0071] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.
[0072] It should be noted that when an element is referred to as being "fixed to" or "disposed on" another element, it can be directly on the other element or indirectly on the other element. When an element is referred to as being "connected to" another element, it can be directly connected to the other element or indirectly connected to the other element.
[0073] It should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating the orientation or position relationship, are based on the orientation or position relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present application.
[0074] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.
[0075] In the following embodiments, ROI means region of interest.
[0076] Example 1
[0077] See also Figure 1 , a medical equipment operation status monitoring method based on artificial intelligence, comprising the following steps:
[0078] S1. Acquire an image in an MRI device, and segment metal artifacts in the image based on the acquired image;
[0079] In S1, a U-Net model is used to extract metal artifacts;
[0080] The U-Net model includes an encoder, a decoder, a final layer, and a skip connection;
[0081] The encoder includes a convolution layer and a pooling layer; the convolution layer includes two 3*3 convolutions, and a ReLU activation function is set after each convolution; the pooling layer uses a 2*2 maximum pooling operation to reduce the spatial dimension of the feature map. After each pooling, the dimension of the feature map is halved; and the number of channels is doubled after each pooling layer;
[0082] The decoder includes an upsampling layer, a splicing layer, and a convolution layer. The upsampling layer uses transposed convolution or upsampling to increase the spatial dimension of the feature map; the splicing layer splices the upsampled feature map with the feature map of the corresponding encoder part; the convolution layer uses two 3*3 convolutions, and a ReLU activation function is set after each convolution to restore the features;
[0083] The final layer includes a 1*1 convolution to reduce the number of channels to the number of target categories; and the Sigmoid function is used as the activation function in the final layer;
[0084] The jump connection is a direct connection, and the feature map before each pooling layer in the encoder is directly concatenated with the output of the corresponding upsampling layer of the decoder.
[0085] The specific structure of the U-Net model has been given in the above steps, and how to train the U-Net model and how to tune the model are conventional technical means in this field, so no relevant details are given in this embodiment; it should be noted that the use of the U-Net model to extract metal artifacts in this embodiment is not a limitation of the present invention; in the current prior art, the technical means of image recognition are already quite mature; therefore, using the corresponding model in the prior art to realize the extraction of metal artifacts described in the present invention is within the scope of protection of the present invention; currently, image recognition technology has penetrated into various industries, and the application of the model has only changed to different fields.
[0086] S2. Determine the position of the metal artifact based on the extracted metal artifact;
[0087] The calculation formula in S2 is as follows:
[0088]
[0089]
[0090] Where: n is the number of pixels of metal artifacts; x i and i are the horizontal and vertical coordinates of the ith metal artifact pixel; v i is the value of the i-th pixel; C x and C y Let the abscissa and ordinate be the centroid.
[0091] See also Figure 2 , S3. Based on the position of the metal artifact and the area to be imaged of the patient, determine whether the metal artifact affects the area to be imaged;
[0092] The S3 comprises the following steps:
[0093] S3.1: Determine the area to be imaged and identify the center point of the area to be imaged;
[0094] The S3.1 comprises the following steps:
[0095] Obtain the lesion area that needs to be examined according to the doctor's order information;
[0096] Based on the lesion area to be examined, extracting the area to be imaged by using the U-Net model, wherein the original image used to extract the area to be imaged by using the U-Net model is the same as the original image used to segment the metal artifacts;
[0097] Based on the extracted area to be imaged, determine the center point of the area to be imaged:
[0098]
[0099]
[0100] Where: n is the number of pixels in the area to be imaged; x i ′ and y i ′ is the horizontal and vertical coordinates of the i-th area to be imaged; v i is the value of the i-th pixel; T x and T y Make the horizontal and vertical coordinates of the center point.
[0101] S3.2: Calculate the distance between the center point of the imaged area and the centroid of the metal artifact:
[0102]
[0103] Where: D is the distance between the centroid of the metal artifact and the center of the imaged area; C x and C y is the horizontal and vertical coordinates of the center of mass; T x and T y is the coordinate of the center point of the area to be imaged;
[0104] S3.3: Compare the calculated distance between the area to be imaged and the centroid of the metal artifact with a preset first distance threshold. If D is less than or equal to the first distance threshold, the metal artifact will affect the area to be imaged; if D is greater than the first distance threshold, the influence of the metal artifact on the area to be imaged is negligible.
[0105] See also Figure 4 , S4. If the metal artifact affects the imaging area, the degree of impact is determined; the determined degree of impact is compared with the impact threshold; if the degree of impact is lower than the impact threshold, it can be ignored; if the degree of impact exceeds the impact threshold, an alarm is issued;
[0106] The S4 comprises the following steps:
[0107] S4.1: Select multiple ROIs in the area to be imaged, and for each ROI, calculate the average signal intensity within the ROI:
[0108]
[0109] Where: MU i Represents the signal mean of each ROI; N i Indicates the total number of pixels in the ROI; S p represents the signal intensity of the pth pixel in the ROI;
[0110] S4.2: Calculate the mean of the overall signal:
[0111]
[0112] Where: N represents the number of ROIs; MU i ′ represents the signal mean of all ROIs;
[0113] S4.3: Calculate the standard deviation of the overall signal:
[0114]
[0115] S4.4: Compare the standard deviation of the overall signal with a preset threshold T. If sigma is greater than T, it is considered that the metal artifact has a significant impact on the imaging area.
[0116] S5. Adjust the patient's position or scan parameters based on the alarm information to reduce the impact of metal artifacts.
[0117] Since the distance method used in S3 is used to determine whether there is an impact, that is, when the distance between the metal artifact and the area to be imaged is too close, we believe that it has an impact, but we do not know the actual impact on the area to be imaged, that is, whether it affects the image quality of the area to be imaged, resulting in unclear imaging of the area to be imaged, etc. Therefore, the impact is further determined in S4. Secondly, the distance is determined by the center point method in S3, and such determination ignores the impact of the minimum distance between the metal artifact and the area to be imaged. Therefore, we use the standard deviation of the global signal to further compensate for the above problem. Thus, from a global perspective, the impact of the metal artifact on the global situation is considered. Of course, in this embodiment, S4 can also be determined, that is, S4 is directly performed to calculate the impact after acquiring the MRI image, so as to identify the impact of the metal artifact on the area to be imaged. However, this step also requires the steps of S1 and S2, that is, it is also necessary to extract the metal artifact. The step in S3 is simpler than step S4; the calculation method is more convenient, so in this embodiment, it is actually the data that is unnecessary to be processed in S4 that is filtered through S3, thereby increasing the processing speed of the entire data.
[0118] Example 2
[0119] The difference between this embodiment 2 and embodiment 1 is that the specific steps of S3 and S4 are different:
[0120] See also Figure 3 In S3, the shortest distance between the metal artifact area and all points in the area to be imaged is used to determine whether the metal artifact affects the local area to be imaged. The specific steps are as follows:
[0121] Define area: determine the boundary between the metal artifact and the area to be imaged, and extract all two-dimensional coordinate points in the metal artifact and the area to be imaged;
[0122] Calculate the distance: For each point A in the metal artifact area i (x i ,y i ), calculate its relationship with each point B in the area to be imaged j (x j ,y j ) is:
[0123]
[0124] Where: D ij Indicates point A in the metal artifact area i and the midpoint B of the area to be imaged j The Euclidean distance of
[0125] Find the shortest distance: traverse all distances Dij , find the shortest distance D min ;
[0126] If D min If D is less than or equal to the second distance threshold, the metal artifact affects the imaging area; min If the distance is greater than the second distance threshold, the influence of the metal artifact on the imaging area is negligible.
[0127] See also Figure 5 , said S4 comprises the following steps:
[0128] Select the ROI closest to the metal artifact in the imaging area as the reference area;
[0129] Calculate the signal mean of the reference ROI:
[0130]
[0131] Where: MU c represents the signal mean of the reference ROI; S p Represents the signal value of each pixel; N represents the number of pixels in the reference ROI;
[0132] Calculate the signal mean of other ROIs: Calculate the signal mean MU for other ROIs in other areas to be imaged i ;
[0133] Calculate signal non-uniformity: Use the signal mean of the reference ROI to calculate the signal non-uniformity of all ROIs in the area to be imaged:
[0134]
[0135] Where: sigma represents the signal non-uniformity of all ROIs in the area to be imaged, that is, the standard deviation;
[0136] Determine the degree of influence: If the calculated sigma′ is greater than a preset threshold value T′, it is considered that the metal artifact has a significant influence on the imaging area.
[0137] In this embodiment, the shortest distance method is used to filter out the data that does not need to be processed in S4; that is, the shortest distance between the metal artifact and the area to be imaged is judged by the set second distance threshold, and S4 is executed if it is greater than the second distance threshold; here, since the shortest distance between the area to be imaged of the metal artifact and the metal artifact area is known, we can calculate the influence based on the shortest distance, and directly compare the influence of the metal artifact on the shortest distance with the set influence threshold; if it exceeds the threshold, it is considered that the influence of the metal artifact on the area to be imaged cannot be ignored and needs to be adjusted.
[0138] Example 3
[0139] An artificial intelligence-based medical device operation status monitoring system includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the artificial intelligence-based medical device operation status monitoring method as described in Example 1 or Example 2 is implemented.
[0140] The above are only preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application.
Claims
1. A method for monitoring the operation status of medical equipment based on artificial intelligence, characterized in that: The following steps are involved: S1. Acquire an image in an MRI device, and segment metal artifacts in the image based on the acquired image; In S1, a U-Net model is used to extract metal artifacts; The U-Net model includes an encoder, a decoder, a final layer, and a skip connection; The encoder includes a convolution layer and a pooling layer; the convolution layer includes two 3*3 convolutions, and a ReLU activation function is set after each convolution; the pooling layer uses a 2*2 maximum pooling operation to reduce the spatial dimension of the feature map. After each pooling, the dimension of the feature map is halved; and the number of channels after each pooling layer is doubled; The decoder includes an upsampling layer, a splicing layer, and a convolution layer. The upsampling layer uses transposed convolution or upsampling to increase the spatial dimension of the feature map; the splicing layer splices the upsampled feature map with the feature map of the corresponding encoder part; the convolution layer uses two 3*3 convolutions, and a ReLU activation function is set after each convolution to restore the features; The final layer includes a 1*1 convolution to reduce the number of channels to the number of target categories; And the Sigmoid function is used as the activation function in the final layer; The jump connection is a direct connection, and the feature map before each pooling layer in the encoder is directly spliced with the output of the corresponding upsampling layer of the decoder; S2. Determine the position of the metal artifact based on the extracted metal artifact; S3. Based on the position of the metal artifact and the area to be imaged of the patient, determining whether the metal artifact affects the area to be imaged; S4. If the metal artifact affects the imaging area, the degree of impact is determined; the determined degree of impact is compared with the impact threshold; if the degree of impact is lower than the impact threshold, it can be ignored; if the degree of impact exceeds the impact threshold, an alarm is issued; S5. Adjust the patient's position or scan parameters based on the alarm information to reduce the impact of metal artifacts.
2. The method for monitoring the operation status of medical equipment based on artificial intelligence according to claim 1, characterized in that: The calculation formula in S2 is as follows: Where: n is the number of pixels of metal artifacts; x i and i are the horizontal and vertical coordinates of the ith metal artifact pixel; v i is the value of the i-th pixel; C x and C y Let the abscissa and ordinate be the centroid.
3. The method for monitoring the operation status of medical equipment based on artificial intelligence according to claim 1, characterized in that: The S3 comprises the following steps: S3.1: Determine the area to be imaged and identify the center point of the area to be imaged; S3.2: Calculate the distance between the center point of the imaged area and the centroid of the metal artifact: Where: D is the distance between the centroid of the metal artifact and the center of the imaged area; C x and C y is the horizontal and vertical coordinates of the center of mass; T x and T y is the coordinate of the center point of the area to be imaged; S3.3: Compare the calculated distance between the area to be imaged and the centroid of the metal artifact with a preset first distance threshold. If D is less than or equal to the first distance threshold, the metal artifact will affect the area to be imaged; if D is greater than the first distance threshold, the influence of the metal artifact on the area to be imaged is negligible.
4. The method for monitoring the operation status of medical equipment based on artificial intelligence according to claim 3 is characterized in that: The S3.1 comprises the following steps: Obtain the lesion area that needs to be examined according to the doctor's order information; Based on the lesion area to be examined, extracting the area to be imaged by using the U-Net model, wherein the original image used to extract the area to be imaged by using the U-Net model is the same as the original image used to segment the metal artifacts; Based on the extracted area to be imaged, determine the center point of the area to be imaged: Where: n is the number of pixels in the area to be imaged; x i ' and y i ' is the horizontal and vertical coordinates of the i-th area to be imaged; v i is the value of the i-th pixel; T x and T y Make the horizontal and vertical coordinates of the center point.
5. The method for monitoring the operation status of medical equipment based on artificial intelligence according to claim 1, characterized in that: In S3, the shortest distance between the metal artifact area and all points in the area to be imaged is used to determine whether the metal artifact affects the local area to be imaged. The specific steps are as follows: Define area: determine the boundary between the metal artifact and the area to be imaged, and extract all two-dimensional coordinate points in the metal artifact and the area to be imaged; Calculate the distance: For each point A in the metal artifact area i (x i ,y i ), calculate its relationship with each point B in the area to be imaged j (x j ,y j ) is: Where: D ij Indicates point A in the metal artifact area i and the midpoint B of the area to be imaged j The Euclidean distance of Find the shortest distance: traverse all distances D ij , find the shortest distance D min ; If D min If D is less than or equal to the second distance threshold, the metal artifact affects the imaging area; min If the distance is greater than the second distance threshold, the influence of the metal artifact on the imaging area is negligible.
6. The method for monitoring the operation status of medical equipment based on artificial intelligence according to claim 1, characterized in that: The S4 comprises the following steps: S4.1: Select multiple ROIs in the area to be imaged, and for each ROI, calculate the average signal intensity within the ROI: Where: MU i Represents the signal mean of each ROI; N i Indicates the total number of pixels in the ROI; S p represents the signal intensity of the pth pixel in the ROI; S4.2: Calculate the mean of the overall signal: Where: N represents the number of ROIs; MU i ' represents the signal mean of all ROIs; S4.3: Calculate the standard deviation of the overall signal: S4.4: Compare the standard deviation of the overall signal with a preset threshold T. If sigma is greater than T, it is considered that the metal artifact has a significant impact on the imaging area.
7. The method for monitoring the operation status of medical equipment based on artificial intelligence according to claim 1, characterized in that: The S4 comprises the following steps: Select the ROI closest to the metal artifact in the imaging area as the reference area; Calculate the signal mean of the reference ROI: Where: MU c represents the signal mean of the reference ROI; S p Represents the signal value of each pixel; N represents the number of pixels in the reference ROI; Calculate the signal mean of other ROIs: Calculate the signal mean MU for other ROIs in other areas to be imaged i ; Calculate signal non-uniformity: Use the signal mean of the reference ROI to calculate the signal non-uniformity of all ROIs in the area to be imaged: Where: Simga represents the signal non-uniformity of all ROIs in the area to be imaged, that is, the standard deviation; Determine the degree of influence: If the calculated sigma' is greater than a preset threshold value T', it is considered that the metal artifact has a significant influence on the imaging area.
8. The medical equipment operation status monitoring system based on artificial intelligence is characterized by: It includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the artificial intelligence-based medical device operating status monitoring method as described in any one of claims 1 to 7.
Citation Information
Patent Citations
Metal artifact removal segmented image correction method and system for CT image and computer readable storage medium
CN112767265A
Eyeball segmentation method and device based on convolutional neural network and mixed loss function
CN113706542A
Medical image processing method and device, computer equipment and storage medium
CN115249279A
Lightweight CT system calibration method
CN117030756A
CT apparatus and an image processing method used by the same
US20140211911A1
