Mammary gland X-ray photography method and device, control host, medium and program product

Abstract

The embodiment of the invention discloses a mammography method and device, a control host, a medium and a program product. The method comprises the following steps: acquiring sensing data of a breast positioned between a compression plate and a bearing table; the sensing data are input into a trained artificial intelligence model, the artificial intelligence model comprises a first sub-model, the first sub-model is suitable for determining a positioning state based on the sensing data, and the positioning state represents whether the positioning of the breast is abnormal or not; based on the positioning state, an execution state of performing X-ray photography on the breast is determined. Before exposure, whether breast positioning is abnormal or not can be determined based on the positioning state output by the artificial intelligence model, and dependence on experience of technicians is reduced. Moreover, the execution state of the X-ray photography is determined based on the positioning state, so that unnecessary radiation exposure caused by blind exposure is avoided.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical equipment, in particular to a mammography method, a mammography device, a control host, a medium and a program product. BACKGROUND

[0002] In mammography, a visualized image of internal tissue structure of a breast is presented by applying a certain dose of X-rays to the breast. A mammography system usually includes an X-ray generating assembly, a bearing table, a compression plate and a control host, etc.

[0003] Image quality plays a crucial role in diagnostic accuracy. Improper breast positioning not only increases the risk of radiation exposure for patients, but also may affect accurate diagnosis.

[0004] Currently, breast positioning in mammography mainly relies on the experience of technicians. SUMMARY

[0005] Embodiments of the present application propose a mammography method, a mammography device, a control host, a medium and a program product, which are beneficial to reduce the dependence on the experience of technicians.

[0006] A mammography method includes:

[0007] Obtaining sensing data of a breast positioned between a compression plate and a bearing table;

[0008] Inputting the sensing data into a trained artificial intelligence model, the artificial intelligence model including a first sub-model, the first sub-model being adapted to determine a positioning state based on the sensing data, the positioning state representing whether positioning of the breast is abnormal;

[0009] Determining an execution state of performing X-ray photography on the breast based on the positioning state.

[0010] It can be seen that, before exposure, whether the breast positioning is abnormal can be known based on the positioning state, which is beneficial to reduce the dependence on the experience of technicians. Moreover, the execution state of X-ray photography is determined based on the positioning state, which avoids unnecessary radiation exposure caused by blind exposure.

[0011] In an embodiment, the method includes a first training process of the first sub-model;

[0012] The first training process includes:

[0013] Inputting a first training sample including training sensing data and a first label into the first sub-model, wherein the first label is used to annotate the positioning state of the training sensing data;

[0014] receiving, from the first sub-model, a predicted positioning state based on the training sensing data;

[0015] determining a loss function value of the first sub-model based on a first difference between the predicted positioning state and the first label;

[0016] configuring model parameters of the first sub-model to make the loss function value lower than a first preset threshold.

[0017] Therefore, the fast training for the first sub-model is realized.

[0018] In an embodiment, the artificial intelligence model further comprises a second sub-model, the second sub-model being adapted to determine positioning information based on the sensing data, the positioning information representing a position and / or a posture of the breast.

[0019] It can be seen that the artificial intelligence model can also output the positioning information of the breast, which helps to guide positioning based on the positioning information, thereby improving the positioning accuracy.

[0020] In an embodiment, the method comprises a second training process of the second sub-model;

[0021] The second training process comprises:

[0022] inputting a second training sample comprising the training sensing data and a second label into the second sub-model, wherein the second label is used to annotate the positioning information of the training sensing data;

[0023] receiving, from the second sub-model, predicted positioning information based on the training sensing data;

[0024] determining a loss function value of the second sub-model based on a second difference between the predicted positioning information and the second label;

[0025] configuring model parameters of the second sub-model to make the loss function value lower than a second preset threshold.

[0026] Therefore, the fast training for the second sub-model is realized.

[0027] In an embodiment, the determining the execution state of performing X-ray photography on the breast based on the positioning state comprises:

[0028] when the positioning state represents that the positioning is normal, performing the X-ray photography on the breast;

[0029] when the positioning state represents that the positioning is abnormal, issuing a prompt message of repositioning.

[0030] It can be seen that whether to perform X-ray photography on the breast is determined based on whether the positioning state is normal, thereby avoiding blind exposure caused by abnormal positioning.

[0031] In one embodiment, comprising:

[0032] When the positioning state represents that the positioning is not abnormal, mapping positioning information of an opposite breast of the breast based on the positioning information;

[0033] Displaying the positioning information of the opposite breast after performing X-ray photography on the breast and before performing X-ray photography on the opposite breast;

[0034] Issuing a prompt message for prompting positioning of the opposite breast based on the positioning information of the opposite breast.

[0035] Therefore, based on the positioning information of the breast on which X-ray photography is currently performed and which is normally positioned, the positioning information of the opposite breast is determined and displayed to assist positioning of the opposite breast, thereby facilitating rapid and accurate positioning of the opposite breast.

[0036] In one embodiment, comprising:

[0037] When the positioning state represents that the positioning is not abnormal, storing the positioning information;

[0038] Performing X-ray photography on the breast;

[0039] Displaying the stored positioning information before performing X-ray photography on the breast again;

[0040] Issuing a prompt message for prompting positioning of the breast based on the positioning information;

[0041] Performing X-ray photography on the breast again.

[0042] Therefore, by storing the positioning information of the breast on which X-ray photography is currently performed and which is normally positioned, when X-ray photography of the breast is performed again later (such as a second examination), the stored positioning information is displayed to assist positioning, thereby facilitating rapid and accurate positioning of the subsequent X-ray photography.

[0043] A breast X-ray photography device, comprising:

[0044] An acquisition module configured to acquire sensing data of a breast positioned between a compression plate and a bearing table;

[0045] An input module configured to input the sensing data into a trained artificial intelligence model, the artificial intelligence model comprising a first sub-model adapted to determine a positioning state based on the sensing data, the positioning state representing whether positioning of the breast is abnormal;

[0046] an execution module configured to determine an execution state of performing X-ray photography on the breast based on the positioning state.

[0047] It can be seen that, before exposure, whether the breast positioning is abnormal can be known based on the positioning state, which is beneficial to reduce the dependence on the experience of technicians. Moreover, the execution state of X-ray photography is determined based on the positioning state, which avoids unnecessary radiation exposure caused by blind exposure.

[0048] In an embodiment, the artificial intelligence model further comprises a second sub-model adapted to determine positioning information based on the sensing data, the positioning information representing a position and / or a posture of the breast.

[0049] It can be seen that the artificial intelligence model can also output the positioning information of the breast, which is helpful to guide positioning based on the positioning information, thereby improving the positioning accuracy.

[0050] In an embodiment, the execution module is configured to, when the positioning state represents that the positioning is not abnormal, map out positioning information of an opposite breast of the breast based on the positioning information; display the positioning information of the opposite breast after performing X-ray photography on the breast and before performing X-ray photography on the opposite breast; and issue a prompt message for prompting positioning of the opposite breast based on the positioning information of the opposite breast.

[0051] Therefore, based on the positioning information of the breast on which X-ray photography is currently performed and which has normal positioning, the positioning information of the opposite breast is determined and displayed to assist positioning of the opposite breast, which facilitates rapid and accurate positioning of the opposite breast.

[0052] In an embodiment, the execution module is configured to, when the positioning state represents that the positioning is not abnormal, store the positioning information; perform X-ray photography on the breast; display the stored positioning information before performing X-ray photography on the breast again; issue a prompt message for prompting positioning of the breast based on the positioning information; and perform X-ray photography on the breast again.

[0053] Therefore, by storing the positioning information of the breast on which X-ray photography is currently performed and which has normal positioning, when X-ray photography of the breast is performed again subsequently (for example, a second examination), the stored positioning information is displayed to assist positioning, which facilitates rapid and accurate positioning of the subsequent X-ray photography.

[0054] A control host of a breast X-ray photography system, comprising:

[0055] a processor;

[0056] a memory configured to store executable instructions of the processor;

[0057] The processor is configured to read the executable instructions from the memory and execute the executable instructions to implement the mammography method as described above.

[0058] A computer-readable storage medium having stored thereon computer instructions that, when executed by a processor, perform the mammography method as described above.

[0059] A computer program product includes a computer program that, when executed by a processor, implements the mammography method as described above. Attached Figure Description

[0060] Preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings, which will make the above and other features and advantages of the present invention more apparent to those skilled in the art. In the drawings:

[0061] Figure 1 This is a schematic diagram of the gantry structure in a mammography system.

[0062] Figure 2 This is an exemplary flowchart of a mammography method according to an embodiment of the present invention.

[0063] Figure 3 This is an exemplary schematic diagram of training an artificial intelligence model according to an embodiment of the present invention.

[0064] Figure 4 This is an exemplary schematic diagram of a mammography process according to an embodiment of the present invention.

[0065] Figure 5 This is an exemplary structural diagram of a mammography apparatus according to an embodiment of the present invention.

[0066] Figure 6 This is an exemplary structural diagram of the control host of a mammography system according to an embodiment of the present invention.

[0067] The accompanying figure is labeled as follows:

[0068]

[0069] Detailed Implementation

[0070] To make the objectives, technical solutions, and advantages of this invention clearer, the following embodiments are provided to further illustrate the invention in detail. The nouns and pronouns referring to "person" in this patent application are not limited to specific genders.

[0071] For the sake of brevity and intuitiveness, the following description uses several representative embodiments to illustrate the solution of the present invention. Numerous details in the embodiments are only used to aid in understanding the solution of the present invention. However, it is obvious that the technical solution of the present invention can be implemented without being limited to these details. To avoid unnecessarily obscuring the solution of the present invention, some embodiments are not described in detail, but only a framework is given. In the following text, "comprising" means "including but not limited to," and "according to..." means "at least according to..., but not limited to only according to...". Due to Chinese language habits, unless the quantity of a component is specifically indicated below, it means that the component can be one or more, or can be understood as at least one.

[0072] A mammography system typically includes an X-ray generating assembly, a stage, a compression plate, and a control unit. The X-ray generating assembly emits X-rays using high voltage from a high-voltage generator. The breast is fixed between the compression plate and the stage. The compression plate compresses the breast to provide a clearer view of the breast tissue during X-ray imaging. A flat panel detector is usually housed within the stage. Medical images of the breast are formed on the flat panel detector. The flat panel detector then transmits these medical images to the control unit. The X-ray generating assembly, stage, and compression plate are usually housed within a gantry.

[0073] Figure 1 This is a schematic diagram of the gantry structure in a mammography system. (Example:) Figure 1 As shown, the gantry includes an X-ray generating assembly 10, a pressure plate 11, and a support stage 12, wherein a flat panel detector 13 is arranged in the support stage 12. Before performing mammography, the breast to be imaged needs to be accurately positioned between the pressure plate 11 and the support stage 12.

[0074] Currently, breast localization in mammography relies heavily on the technician's experience. Furthermore, there is a lack of guidance during the localization process. Additionally, it is typically determined after the X-ray is performed, based on observation of the X-ray image, whether errors occurred during localization. However, identifying localization abnormalities after exposure increases the risk of radiation exposure.

[0075] In this embodiment of the invention, the detection of positioning anomalies before X-ray imaging based on an artificial intelligence model reduces reliance on technician experience and lowers unnecessary radiation risks.

[0076] The above disclosure details the technical defects existing in the relevant technology, the causes of these defects, and the analytical process for overcoming them. In fact, the understanding of these technical defects is not common knowledge in the field, but rather a novel discovery made by the applicant during their research. Furthermore, the tracing of the causes of these technical defects and the analytical process for overcoming them are also the results of the applicant's gradual analysis during the actual research process, and are not common knowledge in the field.

[0077] Figure 2 This is an exemplary flowchart of a mammography method according to an embodiment of the present invention. Preferably, it can be executed by a controller. Figure 2 The method is illustrated. The controller can be implemented as a control host integrated into a mammography system, or as a control unit independent of the control host. For example... Figure 2 As shown, the method includes:

[0078] Step 101: Acquire sensor data of the breast positioned between the compression plate and the support platform.

[0079] Here, various types of sensors can be used to acquire sensory data of the breast. For example, sensor data can include: images (e.g., two-dimensional or three-dimensional images), depth information, and point cloud data, etc. Correspondingly, sensors can be implemented as cameras (two-dimensional or three-dimensional cameras), depth sensors, and point cloud scanning devices, etc. Specifically, a point cloud scanning device is a device that uses three-dimensional scanning technology to acquire a set of three-dimensional coordinate points on the surface of a scanned object; this set of three-dimensional coordinate points is called a point cloud. For example, a point cloud scanning device can capture the set of three-dimensional coordinate points on the surface of an object through methods such as laser scanning and structured light scanning.

[0080] Step 102: Input the sensor data into the trained artificial intelligence model. The artificial intelligence model includes a first sub-model, which is adapted to determine the positioning status based on the sensor data. The positioning status characterizes whether the positioning of the breast is abnormal.

[0081] Sensor data is input into a trained artificial intelligence model, which then determines the localization status, representing whether the breast's location is abnormal, based on the sensor data. Localization abnormalities typically have multiple types, and therefore, localization statuses also have multiple types. For example, localization statuses can include:

[0082] (1) The location status is determined based on whether all breast tissue is included. Wherein: when all breast tissue is included, the location status is not abnormal (i.e., normal); otherwise, the location status is abnormal.

[0083] (2) The positioning status is determined based on whether the nipple is displayed tangentially and does not overlap with the breast tissue. Wherein: when the nipple is displayed tangentially and does not overlap with the breast tissue, the positioning status is not abnormal (i.e., normal); otherwise, the positioning status is abnormal.

[0084] (3) The positioning status is determined based on whether the left and right breasts are symmetrical. Wherein: when the left and right breasts are symmetrical, the positioning status is not abnormal (i.e., normal); otherwise, the positioning status is abnormal.

[0085] (4) The positioning status is determined based on whether there is motion. Wherein: when there is no motion, the positioning status is not abnormal (i.e., normal); otherwise, the positioning status is abnormal.

[0086] (5) The positioning status is determined based on whether the inframammary folds are dispersed and unfolded. Wherein: when they are dispersed and unfolded, the positioning status is not abnormal (i.e., normal); otherwise, the positioning status is abnormal.

[0087] (6) Location status determined based on whether the inframammary fold can be distinguished. Wherein: when it can be distinguished, the location status is not abnormal (i.e., normal); otherwise, the location status is abnormal.

[0088] (7) The positioning status is determined based on the presence or absence of skin folds. Wherein: when there are no skin folds, the positioning status is not abnormal (i.e., normal); otherwise, the positioning status is abnormal.

[0089] (8) The location status is determined based on whether there is body occlusion. Wherein: when there is no body occlusion, the location status is not abnormal (i.e., normal); otherwise, the location status is abnormal.

[0090] The above description provides a typical example of the positioning state. Those skilled in the art will recognize that this description is merely exemplary and is not intended to limit the scope of protection of the embodiments of the present invention.

[0091] Step 103: Based on the positioning status, determine the execution status of performing X-ray imaging on the breast.

[0092] In one implementation, step 103 includes: performing X-ray imaging on the breast when the positioning status indicates normal positioning; and issuing a repositioning prompt message when the positioning status indicates abnormal positioning. Various prompting methods, such as audio, optical, and electrical prompts (e.g., text, images, diagrams), can be used to issue prompt messages to the patient and / or the technician assisting with positioning, indicating that the current breast positioning is abnormal and requires repositioning. Examples include: using an audio prompt through a loudspeaker in the examination room, using light prompts in the examination room, displaying text prompts on the display interface of the support platform, or displaying image prompts on a monitor in the examination room, etc.

[0093] Therefore, understanding the breast's positioning status before exposure helps reduce reliance on the technician's experience. Furthermore, determining the execution of X-ray imaging based on the positioning status avoids unnecessary radiation exposure caused by blind exposure.

[0094] In one implementation, the method includes a first training process for a first sub-model. The first training process includes: inputting a first training sample containing training sensor data and a first label into the first sub-model, wherein the first label is used to label the location state of the training sensor data; receiving a location state predicted based on the training sensor data from the first sub-model; determining a loss function value for the first sub-model based on a first difference between the predicted location state and the first label; and configuring model parameters for the first sub-model such that the loss function value is lower than a first preset threshold.

[0095] In one embodiment, the first label is used to mark at least one of the following positioning states: (1) a positioning state indicating whether all breast tissue is included; (2) a positioning state indicating whether the nipple is displayed tangentially and does not overlap with breast tissue; (3) a positioning state indicating whether the left and right breast images are symmetrical; (4) a positioning state indicating whether there is movement; (5) a positioning state indicating whether the inframammary folds are spread out; (6) a positioning state indicating whether the inframammary folds are distinguishable; (7) a positioning state indicating whether there are skin folds; and (8) a positioning state indicating whether there is body obstruction.

[0096] Therefore, by designing multiple types of first labels, the first sub-model can be trained to detect various localization anomalies. Furthermore, by configuring the model parameters of the first sub-model based on the loss function value, rapid training of the first sub-model is achieved.

[0097] In one implementation, the artificial intelligence model further includes a second sub-model adapted to determine positioning information based on sensor data, the positioning information representing the position and / or posture of the breast.

[0098] It is evident that the artificial intelligence model can also output breast location information, which helps guide the positioning based on the location information, thereby improving the positioning accuracy.

[0099] In one implementation, the method includes a second training process for a second sub-model. The second training process includes: inputting a second training sample containing training sensor data and a second label into the second sub-model, wherein the second label is used to annotate the location information of the training sensor data; receiving location information predicted based on the training sensor data from the second sub-model; determining a loss function value for the second sub-model based on a second difference between the predicted location information and the second label; and configuring model parameters of the second sub-model such that the loss function value is lower than a second preset threshold.

[0100] In one embodiment, the second label is used to mark at least one of the following positioning information: (1) skin contour line; (2) nipple position; (3) skin fold position.

[0101] Therefore, by designing various types of second labels, the second sub-model can be trained to output a variety of localization information. Furthermore, by configuring the model parameters of the second sub-model based on the loss function value, rapid training of the second sub-model is achieved.

[0102] The network architecture of the first and second sub-models can be implemented as an encoder-decoder architecture, a convolutional neural network (CNN), a recurrent neural network (RNN), or a long short-term memory network (LSTM), etc. In fact, the embodiments of this invention do not limit the network architecture of the first and second sub-models.

[0103] In one embodiment, the method further includes: mapping the positioning information of the opposite breast based on the positioning information when the positioning status indicates no abnormality; displaying the positioning information of the opposite breast after X-ray imaging of the breast and before X-ray imaging of the opposite breast; and issuing a prompt message to guide the positioning of the opposite breast based on the positioning information of the opposite breast. For example, the positioning information of the opposite breast can be displayed on the display interface of the support platform to assist in the positioning of the opposite breast. Various prompting methods, such as sound prompts, optical prompts, and electrical prompts, can be used to issue prompt messages to the patient and / or the technician assisting the patient in positioning, to guide the positioning of the opposite breast based on the positioning information of the opposite breast. Examples include: issuing sound prompts through a speaker in the examination room, issuing text prompts on the display interface of the support platform, or displaying image prompts on a monitor in the examination room, etc.

[0104] Therefore, based on the positioning information of the normally positioned breast obtained from current X-ray imaging, the positioning information of the opposite breast is determined and displayed to assist in the positioning of the opposite breast, facilitating rapid and accurate positioning of the opposite breast.

[0105] Example: Suppose we are preparing to perform an X-ray on patient A's left breast. Before performing the X-ray, we acquire sensor data of patient A's left breast positioned between the compression plate and the support platform. This sensor data is input into a trained artificial intelligence model to obtain the positioning status and positioning information of the left breast. The positioning status indicates that there are no abnormalities in the positioning. The positioning information includes the skin contour line. Next, we perform an X-ray on patient A's left breast. Then, before performing an X-ray on patient A's right breast, based on the skin contour line from the left breast positioning process, we calculate (e.g., using a mirroring algorithm) the skin contour line of the right breast. This skin contour line is displayed as a reference point during the right breast positioning process, thereby improving the positioning accuracy of the right breast. For example, the skin contour line of the right breast can be displayed on the support platform's display interface. Patient A positions her right breast based on the skin contour line displayed on the interface (e.g., matching the right breast with the skin contour line on the display interface).

[0106] In one implementation, the method includes: storing location information when the location status indicates that the location is not abnormal; performing an X-ray on the breast; displaying the stored location information before performing another X-ray on the breast; issuing a prompt message to indicate that the breast is located based on the location information; and performing another X-ray on the breast.

[0107] Therefore, by storing the positioning information of a normally positioned breast during the current X-ray imaging, the stored positioning information can be displayed to assist in positioning before subsequent X-ray imaging of the same breast (e.g., a second examination), facilitating rapid and accurate positioning in subsequent X-ray imaging.

[0108] For example: Suppose that during the first examination, an X-ray is to be performed on patient A's left breast. Before performing the X-ray, sensor data of patient A's left breast positioned between the compression plate and the support platform is acquired. This sensor data is input into a trained artificial intelligence model to obtain the positioning status and positioning information of the left breast. The positioning status indicates that the positioning is normal. The positioning information includes the skin contour and nipple position. Next, the X-ray of patient A's left breast is performed during the first examination, and the positioning information from the first examination is stored in memory. Then, at a certain time later (e.g., during the second examination one month later), before performing another X-ray on patient A's left breast, the positioning information from the first examination is retrieved from memory. The skin contour and nipple position from the first examination are displayed as reference points during the positioning process of the left breast in the second examination, thereby improving the positioning accuracy of the breast in the second examination and facilitating the accurate identification of differences between the X-ray images from the first and second examinations.

[0109] Figure 3This is an exemplary schematic diagram illustrating the training of an artificial intelligence model according to an embodiment of the present invention. Figure 3 As shown, the first training sample 31 includes training sensor data 20 and a first label 21 annotating the positioning status of the training sensor data 20, wherein the positioning status characterizes whether the positioning is abnormal. The second training sample 32 includes training sensor data 20 and a second label 22 annotating the positioning information of the training sensor data 20.

[0110] The first training sample 31 is input into the first sub-model 41 of the artificial intelligence model 40. The first sub-model 41 predicts the location state 43 based on the training sensor data 20. The difference between the predicted location state 43 and the first label 21 is calculated to obtain the loss function value 45 of the first sub-model 41. Then, based on the loss function value of the first sub-model 41, backpropagation is performed on the first sub-model 41 to update the model parameters of the first sub-model 41 until the loss function value 45 of the first sub-model 41 is less than a preset value, thus completing the training process of the first sub-model 41. The trained first sub-model 41 has the ability to predict location anomalies based on the input sensor data.

[0111] The second training sample 32 is input into the second sub-model 42 of the artificial intelligence model 40. The second sub-model 42 predicts the location information 44 based on the training sensor data 20. The difference between the predicted location information 44 and the second label 22 is calculated to obtain the loss function value 46 of the second sub-model 42. Then, based on the loss function value 46 of the second sub-model 42, backpropagation is performed on the second sub-model 42 to update the model parameters of the second sub-model 42 until the loss function value 46 of the second sub-model 42 is less than a preset value, thus completing the training process of the second sub-model 42. The trained second sub-model 42 has the ability to predict location information based on the input sensor data.

[0112] Figure 4 This is an exemplary schematic diagram of a mammography process according to an embodiment of the present invention. Figure 4 In this process, sensor 40 collects data from the target 51 (i.e., the breast positioned between the compression plate and the support platform) to obtain sensing data 52. The sensing data 52 is input into a trained artificial intelligence model 53 to obtain a predicted positioning state 54 and predicted positioning information 55. Positioning information 55 includes skin contour lines 56. When the positioning state 54 is normal, the positioning information 55 and the X-ray image 56 are stored in memory.

[0113] Then, in the subsequent repositioning of the target 51, the skin contour line 56 and the X-ray image 57 are read from memory as a reference for repositioning. A prompt message (e.g., various types of audio-visual prompts) containing the skin contour line 56 and the X-ray image 57 is issued to prompt the user to perform positioning based on the skin contour line 56. For example, the skin contour line 56 is displayed on the stage's display interface, and an audio prompt is issued through a speaker in the examination room to prompt the user to position the target 51 based on the skin contour line 56 displayed on the interface (e.g., aligning the target 51 with the skin contour line on the display interface to match the target 51 with the skin contour line). Here, during the subsequent re-photographing of the target 51, the X-ray image 57 is further provided as auxiliary information, which helps the technician identify differences between the X-ray image from the first examination and the X-ray image from the second examination.

[0114] In various embodiments, the display interface of the support stage can be implemented as a display screen covering the surface of the support stage (such as the surface in contact with the detection target), and the embodiments of the present invention are not limited thereto.

[0115] Figure 5 This is an exemplary structural diagram of a mammography apparatus according to an embodiment of the present invention. Figure 5 As shown, the mammography device 500 includes: an acquisition module 501 for acquiring sensor data of the breast positioned between a compression plate and a support platform; an input module 502 for inputting the sensor data into a trained artificial intelligence model 504, the artificial intelligence model 504 including a first sub-model 505, the first sub-model 505 being adapted to determine the positioning state based on the sensor data, the positioning state representing whether the positioning of the breast is abnormal; and an execution module 503 for determining the execution state of performing X-ray imaging on the breast based on the positioning state.

[0116] In one implementation, the artificial intelligence model 504 further includes a second sub-model 506, which is adapted to determine positioning information based on sensor data, the positioning information representing the position and / or posture of the breast.

[0117] In one implementation, the execution module 503 is configured to map the positioning information of the opposite breast based on the positioning information when the positioning status indicates that the positioning is not abnormal; display the positioning information of the opposite breast after performing X-ray imaging on the breast and before performing X-ray imaging on the opposite breast; and issue a prompt message to prompt the positioning of the opposite breast based on the positioning information of the opposite breast.

[0118] In one implementation, the execution module 503 is configured to: store positioning information when the positioning status indicates that the positioning is not abnormal; perform X-ray imaging on the breast; display the stored positioning information before performing X-ray imaging on the breast again; issue a prompt message to prompt the positioning of the breast based on the positioning information; and perform X-ray imaging on the breast again.

[0119] The present invention also proposes a control host for a mammography system with a processor-memory architecture. Figure 6 This is a structural diagram of the control host according to an embodiment of the present invention. Figure 6 As shown, the control host 600 includes a processor 601, a memory 602, and a computer program stored in the memory 602 and executable on the processor 601. When executed by the processor 601, the computer program implements any of the above-described mammography methods. Specifically, the memory 602 can be implemented as an electrically erasable programmable read-only memory (EEPROM), flash memory, programmable programmable read-only memory (PROM), or other storage media. The processor 601 can be implemented as including one or more central processing units (CPUs) or one or more field-programmable gate arrays (FPGAs), wherein the FPGA integrates one or more CPU cores. Specifically, the CPU or CPU core can be implemented as a CPU, MCU, or DSP, etc.

[0120] It should be noted that not all steps and modules in the above processes and structural diagrams are mandatory; some steps or modules can be omitted as needed. The execution order of the steps is not fixed and can be adjusted as required. The division of modules is merely for the convenience of description and functional division. In actual implementation, a module can be implemented by multiple modules, and the functions of multiple modules can also be implemented by the same module. These modules can be located in the same device or in different devices.

[0121] The hardware modules in each embodiment can be implemented mechanically or electronically. For example, a hardware module may include specially designed permanent circuitry or logic devices (such as dedicated processors, such as FPGAs or ASICs) to perform specific operations. A hardware module may also include programmable logic devices or circuitry (such as general-purpose processors or other programmable processors) temporarily configured by software to perform specific operations. The choice between mechanical implementation, dedicated permanent circuitry, or temporarily configured circuitry (such as software-configured circuitry) can be made based on cost and time considerations.

[0122] The present invention also provides a machine-readable storage medium storing instructions for causing a machine to perform the methods described in this application. Specifically, a system or apparatus equipped with a storage medium storing software program code that implements the functions of any of the embodiments described above, and causing a computer (e.g., CPU, MCU, or MPU) of the system or apparatus to read and execute the program code stored in the storage medium. Furthermore, an operating system or similar device operating on a computer can perform some or all of the actual operations through instructions based on the program code. The program code read from the storage medium can also be written to a memory located in an expansion board inserted into a computer or to a memory located in an expansion unit connected to the computer. Subsequently, a control unit or similar device installed on the expansion board or expansion unit can perform some or all of the actual operations based on the instructions in the program code, thereby implementing the functions of any of the embodiments described above. Storage medium embodiments for providing program code include floppy disks, hard disks, magneto-optical disks, optical disks (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD+RW), magnetic tapes, non-volatile memory cards, and ROMs. Alternatively, program code can be downloaded from a server computer or the cloud via a communication network.

[0123] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A mammography method, characterized by, The method comprises: obtaining sensing data of a breast positioned between a compression paddle and a support table (101); inputting the sensing data into a trained artificial intelligence model, the artificial intelligence model comprising a first sub-model adapted to determine a positioning state based on the sensing data, the positioning state representing whether positioning of the breast is abnormal (102); determining an execution state of performing X-ray photography on the breast based on the positioning state (103).

2. The method of claim 1, wherein, The method comprises a first training process of the first sub-model; wherein the first training process comprises: inputting a first training sample comprising training sensing data and a first label into the first sub-model, wherein the first label is used to label the positioning state of the training sensing data; receiving a predicted positioning state from the first sub-model based on the training sensing data; determining a loss function value of the first sub-model based on a first difference between the predicted positioning state and the first label; configuring model parameters of the first sub-model to make the loss function value lower than a first preset threshold.

3. The method of claim 2, wherein, The artificial intelligence model further comprises a second sub-model adapted to determine positioning information based on the sensing data, the positioning information representing a position and / or posture of the breast.

4. The method of claim 3, wherein, The method comprises a second training process of the second sub-model; wherein the second training process comprises: inputting a second training sample comprising the training sensing data and a second label into the second sub-model, wherein the second label is used to label the positioning information of the training sensing data; receiving predicted positioning information from the second sub-model based on the training sensing data; determining a loss function value of the second sub-model based on a second difference between the predicted positioning information and the second label; configuring model parameters of the second sub-model to make the loss function value lower than a second preset threshold.

5. The method according to any one of claims 1-4, characterized in that, The determination of the execution state of performing X-ray photography on the breast based on the positioning state (103) comprises: when the positioning state represents that positioning is not abnormal, performing the X-ray photography on the breast; when the positioning state represents that positioning is abnormal, issuing a prompt message for repositioning.

6. The method of claim 3, wherein, The method comprises: when the positioning state represents that positioning is not abnormal, mapping positioning information of an opposite breast of the breast based on the positioning information; after performing X-ray photography on the breast and before performing X-ray photography on the opposite breast, displaying the positioning information of the opposite breast; issuing a prompt message for prompting positioning of the opposite breast based on the positioning information of the opposite breast.

7. The method of claim 3, wherein, The method comprises: when the positioning state represents that positioning is not abnormal, storing the positioning information; performing X-ray photography on the breast; before performing X-ray photography on the breast again, displaying the stored positioning information; issuing a prompt message for prompting positioning of the breast based on the positioning information; performing X-ray photography on the breast again.

8. A mammography apparatus characterized by comprising: The method comprises: an obtaining module (501) configured to obtain sensing data of a breast positioned between a compression paddle and a support table; The input module (502) is configured to input the sensing data into a trained artificial intelligence model (504), wherein the artificial intelligence model (504) comprises a first sub-model (505) adapted to determine a positioning state based on the sensing data, wherein the positioning state represents whether the positioning of the breast is abnormal. The execution module (503) is configured to determine an execution state of performing X-ray photography on the breast based on the positioning state.

9. The apparatus of claim 8, wherein, The artificial intelligence model (504) further comprises a second sub-model (506) adapted to determine positioning information based on the sensing data, wherein the positioning information represents a position and / or a posture of the breast.

10. The apparatus of claim 9, wherein The execution module (503) is configured to, when the positioning state represents that the positioning is not abnormal, map positioning information of an opposite breast of the breast based on the positioning information; display the positioning information of the opposite breast after performing X-ray photography on the breast and before performing X-ray photography on the opposite breast; and send a prompt message for prompting positioning of the opposite breast based on the positioning information of the opposite breast.

11. The apparatus of claim 9, wherein The execution module (503) is configured to, when the positioning state represents that the positioning is not abnormal, store the positioning information; perform X-ray photography on the breast; display the stored positioning information before performing X-ray photography on the breast again; send a prompt message for prompting positioning of the breast based on the positioning information; and perform X-ray photography on the breast again.

12. A control host of a mammography system, characterized in that comprising: a processor (601); a memory (602) configured to store executable instructions of the processor (601); the processor (601) is configured to read the executable instructions from the memory (602) and execute the executable instructions to implement the breast X-ray photography method of any one of claims 1-7.

13. A computer readable storage medium having stored thereon computer instructions, wherein, the computer instructions, when executed by a processor, implement the breast X-ray photography method of any one of claims 1-7.

14. A computer program product, characterised in that, a computer program, when executed by a processor, implements the breast X-ray photography method of any one of claims 1-7.

Images