X-ray apparatus

Abstract

Provided is an X-ray apparatus capable of performing X-ray fluoroscopy or X-ray photography in appropriate conditions more reliably and quickly. The X-ray apparatus includes a condition storage unit that stores image acquisition conditions corresponding to position information of an object in association with the position information; a learning model storage unit that stores a learning model that infers a position of a human body appearing in an image by machine learning and outputs the position; an image analysis unit that infers a position of an object in an input image by inputting the input image to the learning model as an input image and outputs the position; a condition readout unit that reads out image acquisition conditions associated with position information by selecting the position of the object that has been output as the position information; and a main control unit that performs control of at least either of an X-ray tube and an image processing unit in accordance with the read-out image acquisition conditions.

Description

Technical Field

[0001] This invention relates to an X-ray device for irradiating a subject with X-rays to perform X-ray fluoroscopy or X-ray photography. Background Technology

[0002] In medical settings, when using X-ray devices to obtain X-ray images of a subject through fluoroscopy or radiography, it is crucial to set appropriate X-ray exposure conditions or image processing conditions based on the area being imaged or the type of examination being performed. In recent years, X-ray devices equipped with anatomical procedures (APR) have become increasingly common as a means of setting appropriate X-ray exposure or image processing conditions.

[0003] An autopsy procedure is a data structure that pre-associates a series of X-ray irradiation conditions (e.g., tube voltage and tube current) and a series of image processing conditions (e.g., contrast processing conditions) with the radiographic location of the subject or the type of examination. Regarding autopsy procedures, multiple programs are pre-set and stored according to the radiographic location of the subject or the type of examination. For example, in the case of a conventional chest X-ray, information on the X-ray irradiation conditions and image processing conditions suitable for a conventional chest X-ray is stored in association with the conventional chest X-ray.

[0004] When performing X-ray fluoroscopy on the subject, multiple dissection procedures are displayed on the display unit, such as an LCD panel, and the operator selects the appropriate dissection procedure from the multiple dissection procedures (for example, see Patent Documents 1 and 2).

[0005] As an example, in the case of endoscopic retrograde cholangiopancreatography (ERCP) performed on the abdomen of a patient, the operator selects the "Abdomen / ERCP" program from a list of anatomical programs displayed on the monitor, such as "Chest / Routine Radiography," "Abdomen / Routine Radiography," and "Abdomen / ERCP," corresponding to the examination site and type of examination. By making this selection, the X-ray irradiation conditions and image processing parameters, pre-associated with "Abdomen / ERCP" and suitable for abdominal ERCP, are read out and displayed on the monitor. After confirming the displayed X-ray irradiation conditions, the operator begins the examination of the patient.

[0006] Existing technical documents

[0007] Patent documents

[0008] Patent Document 1: Japanese Patent Application Publication No. 2012-143443

[0009] Patent Document 2: Japanese Patent Application Publication No. 2018-191983 Summary of the Invention

[0010] The problem the invention aims to solve

[0011] However, in the case of existing examples with this structure, the following problems exist.

[0012] When performing X-ray imaging using anatomical procedures within existing structures, the operator needs to select the appropriate procedure from a list of displayed options. This manual operation prolongs the examination, posing a particular problem during emergency surgeries. Furthermore, there are concerns that operator errors may lead to inappropriate X-ray exposure parameters during manual operation. Additionally, if the location of the patient exposed to X-rays changes during the procedure or examination, appropriate APR (Automatic Respiratory Procedure) must be reselected based on the changed location, further increasing the operator's workload and prolonging the procedure's duration.

[0013] The present invention was made in view of this situation, and its object is to provide an X-ray apparatus capable of performing X-ray fluoroscopy or X-ray imaging more reliably and quickly under appropriate conditions.

[0014] Solution for solving the problem

[0015] To achieve this objective, the present invention employs the following structure.

[0016] That is, the X-ray device involved in the first method includes: an X-ray tube that irradiates an object with X-rays; an X-ray detector disposed opposite to the X-ray tube for detecting X-rays that have passed through the object; an image processing unit that generates an X-ray image by performing image processing using the detection signal output by the X-ray detector; a condition storage unit that stores image acquisition conditions corresponding to object parts of the object in association with the object parts, the image acquisition conditions including at least one of X-ray irradiation conditions and image processing conditions; and a learning model storage unit that stores a learning model that performs machine learning by using images of the human body as training images to infer the X-ray image. The system includes: a part of the human body reflected in an image and outputting that part; a part inference unit that infers and outputs the part reflected in the input image by inputting at least one of the most recently obtained X-ray image and optical image of the subject into the learning model; a condition readout unit that reads out the image acquisition conditions associated with the object part and stored in the condition storage unit by selecting the part output by the part inference unit as the object part; and a control unit that controls at least one of the X-ray tube and the image processing unit according to the image acquisition conditions read out by the condition readout unit.

[0017] Furthermore, the X-ray apparatus according to the second aspect of the present invention comprises: an X-ray tube that irradiates an examination subject with X-rays; an X-ray detector disposed opposite to the X-ray tube for detecting X-rays that have passed through the examination subject; an image processing unit that generates an X-ray image by performing image processing using a detection signal output by the X-ray detector; a condition storage unit that stores image acquisition conditions corresponding to examination items of the examination subject in association with the examination items, the image acquisition conditions including at least one of X-ray irradiation conditions and image processing conditions; and a learning model storage unit that stores a learning model that infers values ​​by performing machine learning using examination images of the human body as training images. The inspection image is examined to determine the type of inspection and output the type of inspection; the inspection type inference unit infers the type of inspection in the input image and outputs the type of inspection by inputting at least one of the recently obtained X-ray image and optical image of the subject into the learning model; the condition readout unit reads out the image acquisition conditions associated with the inspection item and stored in the condition storage unit by selecting the type of inspection output by the inspection type inference unit as the inspection item; and the control unit controls at least one of the X-ray tube and the image processing unit according to the image acquisition conditions read out by the condition readout unit.

[0018] The effects of the invention

[0019] According to the first aspect of the present invention, the X-ray apparatus automatically sets image acquisition conditions, including at least one of X-ray irradiation conditions and image processing conditions, by using a learning model for inferring object parts reflected in an image and a mechanism that, for example, associates appropriate image acquisition conditions with the object parts in a dissection procedure. The learning model is configured to infer and output the parts of the human body reflected in the image through machine learning by using an image of the human body as a training image. Specifically, the object part inference unit infers and outputs the parts of the subject reflected in the input image by inputting an image of the subject into the learning model. The condition readout unit automatically reads out the image acquisition conditions stored associated with the object part by selecting the already output object part information of the subject. Therefore, when an X-ray image of the subject is acquired, the object part inference unit and the condition readout unit automatically read out the appropriate image acquisition conditions for the irradiation field of the X-ray image. Thus, even if the object part of the subject irradiated by the X-ray changes, the image acquisition conditions suitable for the changed object part are automatically read out. That is, it eliminates the need for operators to manually select the object and set the image acquisition conditions, thus enabling more reliable and rapid X-ray fluoroscopy or X-ray radiography under appropriate conditions.

[0020] According to the second aspect of the present invention, the X-ray apparatus automatically sets image acquisition conditions, including at least one of X-ray irradiation conditions and image processing conditions, by using a learning model that infers the type of examination in an examination image of a human body as an examination item, and a mechanism that associates appropriate image acquisition conditions with the examination item accordingly. The learning model is configured to infer the type of examination in an image and output the type of examination by machine learning using an examination image of the human body as a training image. Specifically, the examination type inference unit infers the type of examination in the input image and outputs the type of examination by inputting an image of the subject into the learning model. The condition readout unit automatically reads out the image acquisition conditions stored associated with the examination item by selecting the output examination type information as the examination item. Therefore, when an X-ray image of the subject is acquired, the examination type inference unit and the condition readout unit automatically read out the appropriate image acquisition conditions for the examination item of the X-ray image. Thus, even if the examination item for irradiating the subject with X-rays changes, the image acquisition conditions suitable for the changed examination item are automatically read out. That is, it eliminates the need for operators to manually select examination items and set image acquisition conditions, thus enabling more reliable and rapid X-ray fluoroscopy or X-ray imaging under appropriate conditions. Attached Figure Description

[0021] Figure 1 This is a front view illustrating the overall structure of the X-ray device involved in Embodiment 1.

[0022] Figure 2 This is a right-side view illustrating the overall structure of the X-ray apparatus involved in Embodiment 1.

[0023] Figure 3 This is a perspective view illustrating the overall structure of the foot switch involved in Embodiment 1.

[0024] Figure 4 This is a diagram illustrating an example of the information displayed by the display unit according to Embodiment 1.

[0025] Figure 5 This is a functional block diagram of the X-ray device involved in Example 1.

[0026] Figure 6 These are diagrams illustrating the APR of Example 1. (a) is a diagram showing the relationship between part information and image acquisition conditions associated with the part information, and (b) is a diagram showing examples of part information related to each part and specific parameters of the image acquisition conditions associated with that part information.

[0027] Figure 7 This is a schematic diagram illustrating a series of steps in Example 1 using a learning model and APR readout image acquisition conditions.

[0028] Figure 8 This is a schematic diagram illustrating a series of steps in Example 1 for reading out image acquisition conditions before the start of X-ray irradiation.

[0029] Figure 9 This is a schematic diagram illustrating a series of steps for reading out image acquisition conditions in Example 1 when the hip joint is the irradiation field.

[0030] Figure 10 This is a schematic diagram illustrating a series of steps for acquiring readout images in Example 1 with the abdomen as the irradiation field.

[0031] Figure 11 This is a schematic diagram illustrating a series of steps for acquiring images in the case of the chest as the irradiation field in Example 1.

[0032] Figure 12 This is a diagram showing the display screen of the APR involved in the existing example.

[0033] Figure 13 This is a diagram illustrating the state of the image acquisition conditions read out using the existing example of APR.

[0034] Figure 14These are diagrams illustrating the APR of Example 1. (a) shows the APR with a routine chest radiograph as an examination item, (b) shows the APR with a chest PCI as an examination item, (c) shows the APR with a routine abdominal radiograph as an examination item, (d) shows the APR with an abdominal ERCP as an examination item, and (e) shows the APR with an abdominal UGI as an examination item.

[0035] Figure 15 The diagram illustrates the APR of Example 2. (a) is a diagram showing the relationship between inspection item information and image acquisition conditions associated with the inspection item information, and (b) is a diagram showing examples of specific parameters of inspection item information related to each part and the image acquisition conditions associated with that inspection item information.

[0036] Figure 16 This is a schematic diagram illustrating a series of steps in Example 2 using a learning model and APR readout image acquisition conditions.

[0037] Figure 17 This is a functional block diagram showing the main parts of the X-ray apparatus involved in Embodiment 3.

[0038] Figure 18 These are diagrams illustrating misparsing that occurs in the learning model in Example 3. (a) shows a case where the learning model performs image parsing correctly, and (b) shows a case where the learning model performs incorrect image parsing.

[0039] Figure 19 This is a flowchart illustrating the main steps of the operation of the X-ray device involved in Embodiment 3.

[0040] Figure 20 This is a schematic diagram illustrating a series of steps in Example 3 where the learning model and APR readout image acquisition conditions are properly performed on the image under the condition that the learning model has performed image parsing.

[0041] Figure 21 This is a schematic diagram illustrating a series of steps in Example 3 where the learning model and APR readout image acquisition conditions are used in the case of incorrect image parsing by the learning model.

[0042] Figure 22 This is a functional block diagram showing the main parts of the X-ray apparatus involved in Embodiment 4.

[0043] Figure 23 This is a diagram showing an example of a screen displaying the approval button in Embodiment 4.

[0044] Figure 24These are flowcharts illustrating the operation of the X-ray apparatus involved in the embodiments. (a) is a flowchart related to Embodiment 1, and (b) is a flowchart related to Embodiment 4.

[0045] Figure 25 This is a diagram showing the relationship between multiple timings and the position of the X-ray irradiation field of each timing in Example 4.

[0046] Figure 26 This is a block diagram showing the control relationships of the main parts of the X-ray apparatus with timing M1 or timing M2 as described in Embodiment 4.

[0047] Figure 27 This is a block diagram showing the control relationships of the main parts of the X-ray apparatus with timing M3 as illustrated in Embodiment 4.

[0048] Figure 28 This is a block diagram showing the control relationships of the main parts of the X-ray apparatus with timing M4 as illustrated in Embodiment 4. Detailed Implementation

[0049] [Example 1]

[0050] Embodiment 1 of the present invention will now be described with reference to the accompanying drawings.

[0051] <Description of the Overall Structure>

[0052] like Figure 1 and Figure 2 As shown, in Embodiment 1, the X-ray device 1 has an X-ray tube 5 and an X-ray detector 7 arranged facing each other across a top plate 3 for supporting a supine subject M. The top plate 3 is mounted on the upper part of a top plate support 4 configured to be movable vertically. The X-ray tube 5 irradiates the subject M with X-rays. The X-ray detector 7 detects the X-rays that have irradiated the subject M from the X-ray tube 5 and passed through it, converting them into electrical signals for output as X-ray detection signals. An example of the X-ray detector 7 is an FPD (Flat Panel Detector).

[0053] X-ray tube 5 and X-ray detector 7 are respectively disposed at one end and the other end of C-arm 9. C-arm 9 is held by arm holding member 11 and configured to rotate along the arc path of C-arm 9 indicated by reference numeral RA. That is, C-arm 9 rotates along the arc path RA about the axis in the y direction (the direction of the long side of the top plate 3).

[0054] The arm holding member 11 is disposed on the side of the support column 13 and is configured to rotate about an axis (along an arc path RB) parallel to the horizontal axis P parallel to the x-direction (the direction of the short side of the top plate 3). The C-arm 9, held in the arm holding member 11, rotates about the axis in the x-direction along with the arm holding member 11. By configuring the C-arm 9 to rotate freely about two orthogonal axes along the arc paths RA and RB, X-rays can be irradiated onto the subject M from any direction.

[0055] The support column 13 is supported by a support base 15 disposed on the ground, and is configured to be horizontally movable along the upper surface of the support base 15 in both the x and y directions. The arm holding member 11 and C-arm 9, supported by the support column 13, move horizontally in either the x or y direction as the support column 13 moves horizontally. A collimator 17 is disposed on the X-ray tube 5 to confine the X-rays irradiated from the X-ray tube 5 into a predetermined shape. As an example of confining the shape of X-rays, a pyramidal shape is cited.

[0056] An optical camera 19 is mounted on the collimator 17. As an example, the optical camera 19 is a digital camera that acquires an optical image of the subject M by photographing it with visible light. Furthermore, in the X-ray apparatus 1, the positions and orientations of the optical camera 19 and the X-ray tube 5 are adjusted so that the irradiation field of the optical camera 19 is within the same range as the irradiation field of the X-ray tube 5.

[0057] like Figure 2 As shown, a foot switch 21 is mounted on the ground below the top plate 3. The foot switch 21 is connected to a power supply and the CPU of the main control unit 39, which will be described later, via a cable 23. In Embodiment 1, the main control unit 39 is built into the top plate support 4. Figure 1 In the middle, cable 23 is connected to the top plate support 4.

[0058] like Figure 3 As shown, the foot switch 21 includes a main body 25, a switch 27 operated by the surgeon's foot, and a base plate 29. The switch 27 consists of three pedal-type switches: a fluoroscopy switch 27a, a radiography switch 27b, and a top plate movement switch 27c. In Embodiment 1, there are three switches 27, but the number of switches 27 can be appropriately varied.

[0059] The fluoroscopy switch 27a is used to control the start and stop of X-ray fluoroscopy. Specifically, by stepping on the fluoroscopy switch 27a, X-ray fluoroscopy, which involves intermittent exposure to a relatively low dose of X-rays from the X-ray tube 5 according to the X-ray fluoroscopy conditions described later, is initiated. The radiography switch 27b is used to control the start and stop of X-ray radiography. Specifically, by stepping on the radiography switch 27b, X-ray radiography, which involves exposure to a relatively high dose of X-rays from the X-ray tube 5 according to the X-ray radiography conditions described later, is initiated.

[0060] The top plate movement switch 27c is used to control the vertical movement of the top plate 3. That is, by stepping on the top plate movement switch 27c, the top plate support 4 extends or retracts in the z-direction, thereby changing the height of the top plate 3. By operating the top plate movement switch 27c, the top plate 3 moves up and down between a relatively low position (riding / landing position) for the patient M to sit on and descend on the top plate 3 and a relatively high position (treatment position) for the surgeon to perform medical procedures on the patient M in a supine position.

[0061] like Figure 5 As shown, the X-ray device 1 also includes an image processing unit 33, a display unit 35, an arm position detection unit 37, a main control unit 39, an operating table 41, and a storage unit 43.

[0062] The image processing unit 33 is located after the X-ray detector 7 and generates an X-ray image based on the X-ray detection signal output from the X-ray detector 7. The display unit 35 displays the X-ray image generated by the image processing unit 33 and various information related to the X-ray apparatus 1. Examples of display units 35 include liquid crystal monitors or high-precision displays. Examples of structures in which the display unit 35 is mounted include those suspended from the ceiling, those mounted on a mobile trolley, or those mounted on the operating table 41.

[0063] The arm position detection unit 37 detects the rotational movement of C-arm 9 along each of the circular paths RA and RB, and also detects the translational movement of C-arm 9 in the x and y directions, based on a position detector such as a potentiometer or encoder (not shown). By detecting the rotational and translational movements of C-arm 9, the arm position detection unit 37 detects the position of C-arm 9. By detecting the position of C-arm 9, the arm position detection unit 37 can determine the position of the irradiation field of the X-ray tube 5 relative to the subject M.

[0064] As an example, the main control unit 39 includes information processing units such as a central processing unit (CPU). The main control unit 39 uniformly controls various structures of the X-ray device 1, taking the X-ray tube control unit 31, image processing unit 33, and display unit 35 as examples. The main control unit 39 is equivalent to the control unit of this invention.

[0065] The main control unit 39 includes a machine learning unit 45, an image analysis unit 47, and a conditional readout unit 49. The machine learning unit 45 creates a learning model 51 by performing machine learning on pre-acquired X-ray images or optical images. The image analysis unit 47 analyzes the X-ray images or optical images generated by the X-ray device 1 using the learning model 51 and determines the parts of the human body reflected in the images.

[0066] The image analysis unit 47 involved in Example 1 is equivalent to the object part inference unit of the present invention.

[0067] The condition readout unit 49 reads the image acquisition conditions associated with the region determined by the image analysis unit 47 from the condition storage unit 43 by referring to the anatomy procedure 53 (APR 53). Then, the condition readout unit 49 sends the read image acquisition conditions to the X-ray tube control unit 31 or the image processing unit 33, thereby performing X-ray irradiation from the X-ray tube 5 and generating an X-ray image by the image processing unit 33 according to the sent conditions.

[0068] The control panel 41 is used to input operator instructions related to the operation of the X-ray device 1, and the main control unit 39 performs unified control according to the instructions input by the surgical operator to the control panel 41. Examples of operating devices provided with the control panel 41 include keyboard input panels, touch input panels, mice, dials, toggle switches, and push-button switches. In Embodiment 1, regarding the control panel 41, examples include... Figure 1 The structures shown include those attached to the side of the top plate 3, those mounted on the upper part of the support column 13, and those mounted on a mobile trolley.

[0069] Storage unit 43 stores various X-ray images generated by image processing unit 33, as well as various information related to the operation of X-ray device 1. As an example of storage unit 43, non-volatile memory is provided. Storage unit 43 includes learning model storage unit 55 and condition storage unit 57. Learning model storage unit 55 stores learning model 51 generated by machine learning unit 45. Condition storage unit 57 stores APR 53.

[0070] Here, APR 53 related to Example 1 will be described. For example... Figure 6 As shown in (a), APR 53 is a program that associates image acquisition conditions 63 with each part information 61. Part information 61 is information related to parts of the human body that are the objects of X-ray irradiation; examples include the head and neck, chest, abdomen, hip joint, shoulder, knee, and foot. Image acquisition conditions 63 are a series of conditions related to the acquisition of X-ray images, including X-ray irradiation conditions 65 and image processing conditions 67.

[0071] X-ray irradiation conditions 65 include various parameters related to X-ray irradiation, such as X-ray fluoroscopy conditions 68 and X-ray imaging conditions 69. Examples of parameters related to X-ray irradiation include the tube voltage and current applied to the X-ray tube 5, the X-ray irradiation time, and the X-ray irradiation period. Image processing conditions 67 include parameters related to image processing of the electrical signals detected by the X-ray detector 7. Examples include contrast processing values, sharpening processing values, and edge processing values. Furthermore, image acquisition conditions 63 may also include setting conditions for the X-ray detector 7, such as frame rate and gain value.

[0072] X-ray fluoroscopy condition 68 refers to various parameters related to X-ray irradiation during X-ray fluoroscopy. In X-ray fluoroscopy, relatively weak X-rays are intermittently irradiated to acquire X-ray fluoroscopic images (moving images). X-ray radiography condition 69 refers to various parameters related to X-ray irradiation during X-ray radiography. In X-ray radiography, relatively strong X-rays are irradiated for a short period to acquire X-ray radiographic images (still images). Furthermore, in this invention, X-ray images include both X-ray fluoroscopic images and X-ray radiographic images.

[0073] Figure 6 (b) illustrates the specific content of the image acquisition conditions 63 associated with the location information 61 in the APR 53 involved in this embodiment. As an example, the location information 61a of the hip joint in the location information 61 is pre-associated with the image acquisition conditions 63a, which include X-ray irradiation conditions 65a and image processing conditions 67a. The image processing conditions 67a include parameters such as a contrast processing value of 10 and a sharpening processing value of 7. The X-ray fluoroscopy conditions 68a in the X-ray irradiation conditions 65a include parameters such as a tube voltage of 30kV and a tube current of 2.0mA. The X-ray radiography conditions 69a in the X-ray irradiation conditions 65a include parameters such as a tube voltage of 50kV and a tube current of 3.0mA.

[0074] Similarly, the abdominal region information 61b in the region information 61 is pre-associated with the image acquisition condition 63b, which includes X-ray irradiation condition 65b and image processing condition 67b. X-ray irradiation condition 65b includes X-ray fluoroscopy condition 68b and X-ray radiography condition 69b. The chest region information 61c in the region information 61 is pre-associated with the image acquisition condition 63c, which includes X-ray irradiation condition 65c and image processing condition 67c. X-ray irradiation condition 65c includes X-ray fluoroscopy condition 68c and X-ray radiography condition 69c. The head and neck region information 61d in the region information 61 is pre-associated with the image acquisition condition 63d, which includes X-ray irradiation condition 65d and image processing condition 67d. X-ray irradiation condition 65d includes X-ray fluoroscopy condition 68d and X-ray radiography condition 69d. Thus, an APR 53 is pre-set to associate appropriate image acquisition conditions 63 with multiple region information 61, and this APR 53 is stored in the condition storage unit 57.

[0075] Here, use Figure 7 and Figure 24 (a) The structure of automatically setting image acquisition condition 63 in Example 1 will be described. Figure 24 (a) is a flowchart relating to the operation of the X-ray apparatus 1 according to Embodiment 1. First, a learning model 51 is created in the machine learning unit 45 by performing machine learning in advance (step M1). The machine learning unit 45 acquires images of various parts of the human body in advance as the source image R1. The source image R1 is, for example, a group of images including X-ray images, DRR images obtained by projecting three-dimensional CT images in various directions, optical images acquired by an optical camera, etc.

[0076] Then, the machine learning unit 45 performs image processing on the original image R1 to address changes in X-ray conditions, such as increasing or decreasing contrast, increasing or decreasing brightness, and increasing or decreasing noise, thereby obtaining multiple one-dimensional data-enhanced images R2. Then, the machine learning unit 45 performs image processing on the one-dimensional data-enhanced images R2 to address changes in the configuration of the subject M or the position of the C-arm 9, such as rotating, enlarging, and shrinking, thereby obtaining multiple two-dimensional data-enhanced images R3.

[0077] Finally, the machine learning unit 45 performs machine learning on the original image R1, the one-dimensional data-enhanced image R2, and the two-dimensional data-enhanced image R3 as training images to create a learning model 51 for inferring the parts of the human body reflected in the image. That is, by inputting an X-ray image F or an optical image D as input information into the learning model 51, the learning model 51 infers which part of the human body is reflected in the input image and outputs the inferred information about the human body part. The learned learning model 51 is stored in the learning model storage unit 55.

[0078] Furthermore, while Embodiment 1 illustrates a structure for fabricating a learning model 51 within the X-ray apparatus 1, the learning model 51 can also be fabricated in advance using other devices, and the program for the fabricated learning model 51 can be stored in the learning model storage unit 55. In this case, the machine learning unit 45 can be omitted from the X-ray apparatus 1.

[0079] Second, the image analysis unit 47 analyzes the image using the learning model 51 to determine the location of the subject M reflected in the image obtained for the subject M. That is, the image analysis unit 47 reads the learning model 51 stored in the learning model storage unit 55. Then, after the image processing unit 33 generates an X-ray image of the subject M by X-ray irradiation (step M2), the X-ray image sent from the image processing unit 33 is input as an input image to the learning model 51. The learning model 51 analyzes the input image using the constituent elements of the human body reflected in the input X-ray image as clues, and infers the location of the subject M reflected in the input image. The location information obtained by inference is output from the learning model 51 as location information 61 (step M3).

[0080] Third, the condition readout unit 49 uses the part information 61 obtained by the image analysis unit 47 and the APR 53 to set appropriate image acquisition conditions 63. The condition readout unit 49 inputs the part information 61 obtained by the image analysis unit 47 into the APR 53, reads the image acquisition conditions 63 stored in the APR 53 associated with the part information 61, and outputs the image acquisition conditions 63 (step M4). The condition readout unit 49 sets the output image acquisition conditions 63 as the conditions to be used in the subsequent acquisition of the X-ray image (step M5).

[0081] Image acquisition conditions 63, set as the conditions for X-ray image acquisition, are sent to the X-ray tube 5 and the image processing unit 33, thereby generating a new X-ray image of the subject M according to the various parameters of the set image acquisition conditions 63 (step M6). In this way, the X-ray apparatus 1 automatically sets appropriate image acquisition conditions 63 using the learning model 51 and APR 53. Furthermore, after a predetermined time has elapsed (step M7), the process returns to step S2 to generate an input image, thereby using the learning model 51 to infer location information 61. As an example of the predetermined time, the time for generating 10 frames of X-ray images is listed. In this case, the learning model 51 performs the inference every time 10 frames of X-ray images are generated.

[0082] <Description of Actions>

[0083] Here, the procedure of examining the subject M using the X-ray device 1 will be specifically described when the learning model 51 and APR 53 are stored in the storage unit 43. In Example 1, as an example of the examination, a case of percutaneous coronary intervention (PCI) will be described. In the coronary intervention described in Example 1, a catheter Ch is inserted through the groin, and while the catheter Ch is checked step by step by X-ray fluoroscopy, the catheter Ch is guided to the coronary artery and a stent is placed in the coronary artery. That is, as... Figures 8 to 11 As shown, the X-ray-irradiated parts of the subject M shift in the order of hip joint La, abdomen Lb, and chest Lc.

[0084] When initiating an examination of subject M based on coronary intervention, first set the image acquisition conditions 63 for obtaining the initial X-ray image. The operator, such as... Figure 8 With the subject M placed on the top plate 3 as shown, the C-arm 9 is moved so that the irradiation field is positioned at the hip joint La. Then, the hip joint La of the subject M is photographed using the optical camera 19. Through this photographing, the optical camera 19 generates an optical image D1 of the hip joint La. The data of the generated optical image D1 is sent to the image analysis unit 47 (refer to reference numeral T1 in the attached drawing).

[0085] The image analysis unit 47 inputs the optical image D1 as input to the learning model 51. The learning model 51 analyzes the optical image D1, estimates the location reflected in the optical image D1 as input information, and outputs the location. In this case, the learning model 51 estimates that the location reflected in the optical image D1 is most likely the hip joint, based on clues such as the outline of the subject M reflected in the optical image D1 and the presence of the foot. As a result, the learning model 51 outputs the location information 61a of the hip joint. The location information 61a output from the learning model 51 is sent to the conditional readout unit 49 (refer to reference numeral T2 in the attached drawing).

[0086] The condition readout unit 49 searches for appropriate image acquisition conditions 63 for the hip joint La involved in the location information 61a. That is, by inputting the location information 61a into the APR 53, the image acquisition conditions 63 stored in the APR 53 in association with the location information 61a, i.e., image acquisition conditions 63a, are read out and output. As a result, the X-ray irradiation conditions 65a included in the image acquisition conditions 63a are set as control parameters for the X-ray tube 5, and the image processing conditions 67a are set as control parameters for the image processing unit 33.

[0087] Thus, by performing the operation of acquiring the optical image D1 using the optical camera 19, the image acquisition conditions 63a are automatically set using the learning model 51 and APR 53. The image acquisition conditions 63a set by the input of the optical image D1 are equivalent to the image acquisition conditions 63 used to acquire the initial X-ray image.

[0088] After confirming that the image acquisition condition 63a has been set, the operator begins X-ray fluoroscopy by stepping on the fluoroscopy switch 27a of the foot switch 21. While confirming the X-ray fluoroscopic image of the hip joint La obtained through X-ray fluoroscopy, the operator inserts the catheter Ch into the groin of the subject M.

[0089] At this time, the main control unit 39 controls the X-ray tube 5 according to the X-ray fluoroscopy conditions 68a included in the image acquisition conditions 63a. That is, under conditions such as a tube voltage of 30kV and a tube current of 2.0mA, such as... Figure 9 As shown, X-ray tube 5 irradiates the hip joint La with X-rays. Then, the main control unit 39 controls the image processing unit 33 according to the image acquisition condition 67a included in the image acquisition condition 63a. That is, the image processing unit 33 performs various image processing on the detection signal of the X-ray detector 7 under conditions such as contrast value 10 and sharpness 7, and generates an X-ray fluoroscopic image (X-ray image F1) of the hip joint La as the target area. The operator begins to operate the catheter Ch while confirming the position of the catheter Ch reflected in the X-ray image F1.

[0090] In addition, the X-ray device 1 automatically sets the image acquisition conditions 63 at predetermined time intervals. Figure 9 The process of automatically setting image acquisition conditions 63 by irradiating the hip joint La with X-rays is shown. By irradiating the hip joint La with X-rays from the X-ray tube 5, the image processing unit 33 generates an X-ray image F1 with the hip joint La as the object. The data of the generated X-ray image F1 is sent to the image analysis unit 47.

[0091] The image analysis unit 47 inputs the X-ray image F1 as input to the learning model 51. The learning model 51 estimates the location reflected in the X-ray image F1 as input information and outputs the location. In this case, the learning model 51 uses the pelvis Ba and femur Bc of the subject M reflected in the X-ray image F1 as clues to estimate that the location reflected in the X-ray image F1 is most likely the hip joint. As a result, the learning model 51 outputs information that the location is the hip joint, i.e., location information 61a. The location information 61a output from the learning model 51 is sent to the conditional readout unit 49.

[0092] The condition readout unit 49 searches for appropriate image acquisition conditions 63 for the hip joint La involved in the location information 61a. That is, by inputting the location information 61a into the APR 53, the image acquisition conditions 63 stored in the APR 53 in association with the location information 61a, i.e., the image acquisition conditions 63a that are pre-set to be appropriate for generating an X-ray image of the hip joint La, are read out and output. As a result, the X-ray tube 5 irradiates X-rays based on the X-ray irradiation conditions 65a, and the image processing unit 33 performs image processing based on the image processing conditions 67a. Thus, when the position of the C-arm 9 is determined in such a way that the X-ray irradiation field is located at the hip joint La, the image acquisition conditions 63a are automatically set as the X-ray image acquisition conditions, and the X-ray image F1 is continuously generated. While confirming the position of the catheter Ch reflected in the X-ray image F1, the operator manipulates the catheter Ch to move it toward the heart.

[0093] Furthermore, the generated X-ray image F1 is sent to the image resolution unit 47 and displayed on the display unit 35. Figure 4 The information displayed in the display unit 35 is shown. The display unit 35 has an image display area K1, a resolution result display area K2, a selected area display area K3, and an application condition display area K4. The image display area K1 is used to display the most recently acquired X-ray image F or optical image D. In the case where the hip joint La has been irradiated with X-rays, the most recently acquired image is X-ray image F1, therefore X-ray image F1 is displayed in the image display area K1.

[0094] The analysis result display area K2 is used to display the information output by the learning model 51 as the analysis result. The learning model 51 outputs information about the parts reflected in the input image (in this case, X-ray image F1) along with their accuracy. For example, the following information is output: the accuracy for the hip joint in X-ray image F1 is 91.4%, for the head and neck is 1.4%, for the chest is 4.2%, and for the abdomen is 0.1%. The image analysis unit 47 selects the part with the highest accuracy as part information 61 and sends part information 61 to the conditional readout unit 49. Furthermore, the image analysis unit 47 displays a predetermined number of parts in the analysis result display area K2 in descending order of accuracy. In Embodiment 1, three parts are displayed in the analysis result display area K2 in descending order of accuracy. That is, the information for the "hip joint," "chest," and "head and neck" is displayed together with their accuracy values. The number of parts displayed in region K2 in the analysis results is not limited to three, and can be changed appropriately.

[0095] The selection area display region K3 is used to display the selected area as area information 61. Here, since the area information 61a of the hip joint La is selected, it is displayed as "hip joint" in the selection area display region K3. The application condition display region K4 is used to display the image acquisition condition 63 applied at the current time point. Here, since the image acquisition condition 63a associated with the area information 61a of the hip joint La is applied, the image acquisition condition 63a is displayed in the application condition display region K4. Furthermore, for ease of explanation, in Figure 4 In the image acquisition condition 63a, the tube voltage and tube current information are displayed in the application condition display area K4. An adjustment key NB is also displayed in the application condition display area K4. The operator uses a mouse or similar device to input appropriate information into the adjustment key NB, thereby enabling adjustments to increase or decrease the tube voltage value from the initial value determined in the image acquisition condition 63a.

[0096] The operator directs their gaze toward the display unit 35 to check the content of the image acquisition conditions 63 displayed in the application condition display area K4, the content of the location displayed in the selection area display area K3, and the accuracy information displayed in the resolution result display area K2. This allows the operator to determine whether the image acquisition conditions 63, automatically set according to the learning model 51 and APR 53, are appropriate. Simultaneously, while checking the X-ray image F1 displayed on the display unit 35, the operator guides the catheter Ch toward the coronary artery.

[0097] The operator moves the catheter Ch and moves the position of the C-arm 9 in the y-direction to match the position of the catheter Ch, thereby changing the position of the X-ray irradiation field. Therefore, as the catheter Ch travels from the region of the hip joint La to the region of the abdomen Lb, the object being irradiated by X-rays is displaced from the hip joint La to the abdomen Lb. In the X-ray apparatus 1, because the object being irradiated by X-rays is displaced, the image acquisition conditions 63 are newly set. Figure 10 The process of automatically setting image acquisition conditions 63 by irradiating the abdominal Lb with X-rays is shown.

[0098] Since X-rays are irradiated onto the abdomen Lb through the X-ray tube 5, the image processing unit 33 generates an X-ray image F2 of the abdomen Lb as the target area. The data of the generated X-ray image F2 is sent to the image analysis unit 47.

[0099] The image analysis unit 47 takes the X-ray image F2 as input and feeds it into the learning model 51 to analyze the X-ray image F2. Specifically, in the image analysis unit 47, the learning model 51 estimates the location reflected in the X-ray image F2, which is the input information, and outputs that location. In this case, the learning model 51 uses the stomach Ga and lumbar vertebrae Bh of the subject M reflected in the X-ray image F2 as clues to estimate that the location reflected in the X-ray image F2 is most likely the abdomen. As a result, the learning model 51 outputs abdominal location information 61b. The location information 61b output from the learning model 51 is sent to the conditional readout unit 49.

[0100] The condition readout unit 49 searches for appropriate image acquisition conditions 63 for the abdomen Lb related to the location information 61b. That is, by inputting the location information 61b into the APR 53, it reads out the image acquisition conditions 63, i.e., image acquisition conditions 63b, stored in the APR 53 associated with the location information 61b, and outputs the image acquisition conditions 63b. As a result, the X-ray tube 5 irradiates X-rays based on the X-ray irradiation conditions 63b, and the image processing unit 33 performs image processing based on the image processing conditions 67b.

[0101] The operator operates the fluoroscopy switch 27a with their foot, thus the X-ray tube 5 irradiates X-rays according to the X-ray fluoroscopy condition 68b to perform X-ray fluoroscopy. Furthermore, when X-ray imaging is required, the operator changes the switch to the imaging switch 27b, thereby the X-ray tube 5 irradiates X-rays according to the X-ray imaging condition 69b to perform X-ray imaging. Through this X-ray irradiation, an X-ray image of the abdomen Lb, generated under appropriate conditions, can be obtained.

[0102] Thus, with the position of C-arm 9 determined so that the X-ray irradiation field is located in the abdomen Lb, image acquisition conditions 63b are automatically set and X-ray image F2 is continuously generated. When the X-ray image F generated by the image processing unit 33 changes from X-ray image F1 reflecting the hip joint La to X-ray image F2 reflecting the abdomen Lb, the image acquisition conditions 63 are quickly changed by the image analysis unit 47 and the condition readout unit 49. That is, the image acquisition conditions 63a suitable for the hip joint La are automatically and quickly changed to image acquisition conditions 63b suitable for the abdomen Lb. Therefore, the operator can confirm the X-ray image F2 generated under appropriate conditions in the display unit 35 without interrupting the operation of the catheter Ch.

[0103] The operator continues manipulating catheter Ch to advance it towards the vicinity of the heart. Then, the position of C-arm 9 is moved in the y-direction accordingly to the advancement of catheter Ch. Therefore, as catheter Ch advances from the region of the abdomen Lb towards the region of the chest Lc where the heart is located, the object being irradiated by X-rays shifts from the abdomen Lb towards the chest Lc. Because the object is shifted towards the chest Lc, new image acquisition conditions 63 are set. Figure 11 The process of automatically setting image acquisition conditions 63 by irradiating the chest Lc with X-rays is shown.

[0104] X-rays are irradiated onto the chest Lc by X-ray tube 5, thereby generating an X-ray image F3 of the chest Lc as the target area by image processing unit 33. The data of the generated X-ray image F3 is sent to image analysis unit 47.

[0105] The image analysis unit 47 inputs the X-ray image F3 as input to the learning model 51. The learning model 51 estimates the location reflected in the X-ray image F3 as input information and outputs the location. In this case, the learning model 51 uses the heart H, lungs Lu, and thoracic vertebrae (not shown) of the subject M reflected in the X-ray image F3 as clues to estimate that the location reflected in the X-ray image F3 is most likely the chest. As a result, the learning model 51 outputs chest location information 61c. The location information 61c output from the learning model 51 is sent to the conditional readout unit 49.

[0106] The condition readout unit 49 searches for appropriate image acquisition conditions 63 for the chest Lc involved in the location information 61c. That is, by inputting the location information 61c into the APR 53, the image acquisition conditions 63 stored in the APR 53 associated with the location information 61c, i.e., image acquisition conditions 63c, are read out and output. Then, the read image acquisition conditions 63c are set as X-ray image acquisition conditions, the X-ray irradiation conditions 65c in the image acquisition conditions 63c are sent to the X-ray tube 5, and the image processing conditions 67c are sent to the image processing unit 33.

[0107] As a result, X-ray tube 5 irradiates X-rays based on X-ray irradiation condition 65c, and image processing unit 33 performs image processing on X-ray image F based on image processing condition 67c. Therefore, since the target area is displaced towards the chest Lc, the image acquisition condition 63c is quickly and automatically set, thus rapidly improving the visibility of X-ray image F3 using appropriate image acquisition condition 63c. While confirming X-ray image F3, the operator places the stent in the coronary artery, thereby concluding the PCI procedure.

[0108] Thus, the X-ray device 1 involved in Embodiment 1 has the following structure: by combining the learning model 51 with the APR 53, the X-ray device 1 automatically changes the image acquisition conditions 63 according to the changes in the object part that becomes the object of X-ray irradiation.

[0109] In existing APR (Area-of-Reflection) devices, the image acquisition conditions associated with information about the object region are read out by the operator manually selecting the object region to be irradiated by X-rays. As an example, existing devices such as... Figure 12 As shown, a touch panel TP is provided for operation, displaying multiple icon groups Ac for specifying target areas. The operator selects the icon from the icon group Ac for specifying the target area to be irradiated by X-rays (for example, icon Ab for specifying the abdomen) and presses the icon.

[0110] By pressing the icon Ab used to specify the abdomen, such as... Figure 13 As shown, the image acquisition condition Na associated with the object location information of the "abdomen" is read and set, and the information of the image acquisition condition Na is displayed on the touch panel TP. Figure 13 In the existing example shown, conditions such as tube voltage of 40kV and tube current of 2.5mA are displayed as image acquisition condition Na. Thus, in order to set the image acquisition conditions using APR in the existing device, the operator must manually perform the operation to determine the object location.

[0111] Therefore, as an example, when performing PCI using existing equipment, whenever the target area to be irradiated by X-rays changes to the hip joint, abdomen, or chest, the operator needs to manually specify the target area using the touch panel TP. Since each such manual operation interrupts the catheter procedure, the following problems arise: the PCI procedure becomes prolonged, and it is difficult for the operator to concentrate on the procedure.

[0112] As a method to reduce the number of manual operations, one approach is to select the chest APR from the beginning and read out the image acquisition conditions suitable for the chest. However, since X-ray images are also acquired using image acquisition conditions suitable for the chest when the target area for X-ray irradiation is the hip joint or abdomen, it is difficult to acquire highly visible X-ray images in the early or middle stages of PCI.

[0113] Compared to this existing structure, the X-ray apparatus 1 according to Embodiment 1 includes an image analysis unit 47 and a conditional readout unit 49. The image analysis unit 47 analyzes recently acquired images using a learning model 51 and estimates the location of the subject M reflected in the image. The conditional readout unit 49 reads out appropriate image acquisition conditions 63 for the location estimated by the image analysis unit 47 using an APR 53. The learning model 51 learns in advance by using images of the human body as training information to infer the location of the human body reflected in the image. Therefore, the learning model 51 infers the location of the subject M reflected in the input image (X-ray image F or optical image D) and outputs it as location information 61.

[0114] In APR 53, image acquisition conditions 63 are associated with the part information 61 used to determine the object part, which enables the appropriate acquisition of X-ray images of the object part. Therefore, by sending the image of the subject M recently acquired using the X-ray tube 5 or the optical camera 19 as input information to the image analysis unit 47 and the condition readout unit 49, appropriate image acquisition conditions 63 are automatically set for the parts of the object that become the image of the subject M.

[0115] Thus, the X-ray apparatus 1 according to Embodiment 1 has the following structure: by using a learning model 51 for estimating the parts of the human body reflected in the image and an APR 53 that associates image acquisition conditions 63 with part information 61, the image acquisition conditions 63 are automatically changed according to changes in the object part irradiated by X-rays. Therefore, even in the case of performing an inspection such as PCI, where the object part irradiated by X-rays changes within a short period of time, the image acquisition conditions 63 automatically change to appropriate parameters according to the change in the object part. Therefore, the operator can focus on the inspection while confirming the highly visible X-ray image acquired under appropriate conditions.

[0116] Furthermore, by combining the learning model 51 with APR 53, a structure can be achieved that uses the image of the subject M as input information and the appropriate image acquisition condition 63 as output information. In current machine learning, it is difficult to accurately infer the image acquisition condition 63 that can appropriately generate the image using an image of the human body as training information. That is, it is difficult to use the learning model to directly infer the image acquisition condition based on the image of the subject.

[0117] Therefore, the inventors conducted in-depth research and discovered that the following machine learning method can be performed: using an image of the subject M as input information, the location reflected in the image can be accurately inferred. Moreover, by combining the APR 53, which establishes a correlation between the location and the image acquisition condition 63, with the learning model 51, the structure of accurately inferring the appropriate image acquisition condition 63 from the image of the subject M is realized.

[0118] [Example 2]

[0119] Next, Embodiment 2 of the present invention will be described. The overall structure of the X-ray device 1A involved in Embodiment 2 is similar to that described above. Figure 1 The overall structure of the X-ray apparatus 1 involved in Embodiment 1 is basically the same. Therefore, in Embodiment 2, the same reference numerals are used to refer to the same structures as in Embodiment 1, and detailed descriptions are omitted.

[0120] The difference between the X-ray apparatus 1A in Example 2 and Example 1, which uses the learning model 51 to infer the object portion of the subject reflected in the image, is that after inferring the examination item in the image using the learning model 51A, the image acquisition condition 63 corresponding to that examination item is read out. The structure for automatically reading out the image acquisition condition 63 in Example 2 will be described below.

[0121] Generally, the parameters of image acquisition conditions 63 suitable for acquiring X-ray images differ depending on the object region depicted in the X-ray image. However, even if the object region in the X-ray image is the same, the parameters of image acquisition conditions 63 suitable for acquiring the X-ray image will differ if the type of examination performed on that object region (also known as the examination item or procedure) is different. For example, the image acquisition conditions 63 suitable for acquiring X-ray images differ between a routine X-ray of the abdomen without any insertion and an ERCP procedure with an endoscope inserted into the abdomen.

[0122] Therefore, the learning model 51A according to Embodiment 2 is configured to take an image as input information and output inspection items. The similarity between the learning model 51A according to Embodiment 2 and the learning model 51 according to Embodiment 1 is that the original image R1, the one-dimensional data-enhanced image R2, and the two-dimensional data-enhanced image R3 are used as training images for machine learning in the machine learning unit 45. However, the learning model 51 according to Embodiment 1 identifies the skeleton, organs, contours, etc. of the human body reflected in the image, and infers the parts of the human body reflected in the image based on the structure of the human body.

[0123] On the other hand, the learning model 51A involved in Embodiment 2 is configured to detect not only the human body structure such as bones and organs reflected in the image, but also examination equipment such as catheters or endoscopes and examination reagents such as contrast agents. Furthermore, in addition to using information related to the structure of the human body as clues, the learning model 51A also uses information such as the presence or absence of examination equipment or examination reagents as clues to infer which examination item the image pertains to. That is, the learning model 51A is configured to use an image of the subject M as input information and output information used to determine the examination item pertaining to the image as examination item information 71.

[0124] Figure 14 The figures shown are diagrams illustrating the relationship between the input image and the inference results obtained by the learning model 51A. In the following... Figure 14 When the X-ray image W1 shown in (a) is input to the learning model 51A, the learning model 51A analyzes the X-ray image W1 to detect the heart (H), lungs (Lu), thoracic vertebrae (not shown), and the outline of the human body, etc. Then, based on information such as the presence of the heart (H), lungs (Lu), etc., and the absence of examination equipment such as catheters, the learning model 51A infers that the X-ray image W1 is a conventional X-ray image of the chest. As a result, when the X-ray image W1 is set as input information, the learning model 51A outputs the following examination item information 71 (examination item information 71a): the examination item of the input image is "conventional chest radiography".

[0125] In the like Figure 14When the X-ray image W2 shown in (b) is input into the learning model 51A, the learning model 51A analyzes the X-ray image W2 to detect the heart (H), lungs (Lu), catheters (Ch), etc., and infers from these that the X-ray image W2 is an image of a chest catheterization procedure, namely PCI, as the examination item. As a result, when the X-ray image W2 is set as input information, the learning model 51A outputs the following examination item information 71 (examination item information 71b): the examination item of the input image is "chest PCI".

[0126] In the like Figure 14 When the X-ray image W3 shown in (c) is input into the learning model 51A, the learning model 51A analyzes the X-ray image W3 to detect lumbar spine Bh, gastric Ga, etc. Then, combining information such as the presence of lumbar spine Bh, gastric Ga, etc., and information such as the absence of examination equipment such as catheters, it infers that the X-ray image W3 is an image of a routine abdominal X-ray radiograph as the examination item. As a result, when the X-ray image W3 is set as input information, the learning model 51A outputs examination item information 71 such as "routine abdominal radiograph" (examination item information 71c).

[0127] In the like Figure 14 When the X-ray image W4 shown in (d) is input into the learning model 51A, the learning model 51A detects the lumbar spine (Bh), stomach (Ga), endoscopy (Es), etc. Using these as clues, it infers that the X-ray image W4 is an image of an abdominal endoscopic procedure (ERCP). As a result, when the X-ray image W4 is used as input information, the learning model 51A outputs examination item information 71 (examination item information 71d) such as "abdominal ERCP".

[0128] In the like Figure 14 When the X-ray image W5 shown in (e) is input into the learning model 51A, the learning model 51A combines the lumbar spine Bh, stomach Ga, and stomach shape to detect the retained contrast agent Ct, etc. Using these as clues, it infers that the X-ray image W5 is an image of an examination involving gastroscopy, i.e., an upper gastrointestinal X-ray imaging (UGI) series. As a result, when the X-ray image W5 is used as input information, the learning model 51A outputs examination item information 71 such as "abdominal UGI" (examination item information 71e).

[0129] In this way, by taking X-ray images F, optical images D, etc., as input information, the learning model 51A is machine learning in which it infers what kind of inspection item the input images are involved in, and outputs the inspection item information 71 obtained through inference. The learned learning model 51A is stored in the learning model storage unit 55.

[0130] Here, the APR 53A involved in Example 2 will be described. Since the learning model 51A takes an image as input information and outputs inspection item information 71, APR 53A, as... Figure 15 As shown in (a), image acquisition conditions 63 are associated with each inspection item information 71. That is, by inputting inspection item information 71 into APR 53A, image acquisition conditions 63 corresponding to the inspection items involved in the inspection item information 71 are output from APR 53A.

[0131] Figure 15 (b) illustrates the specific content of the image acquisition conditions 63 associated with the examination item information 71 in APR 53A according to Embodiment 2. As an example, examination item information 71a, which includes a chest X-ray, is associated in advance with image acquisition condition 63e, which includes image processing condition 67e, X-ray fluoroscopy condition 68e, and X-ray imaging condition 69e. Examination item information 71b, which includes a chest PCI, is associated in advance with image acquisition condition 63f. Examination item information 71c, which includes an abdominal X-ray, is associated in advance with image acquisition condition 63g. Examination item information 71d, which includes an abdominal ERCP, is associated in advance with image acquisition condition 63h. Thus, in Embodiment 2, APR 53A is pre-set to associate appropriate image acquisition conditions 63 with each of the multiple examination item information 71, and this APR 53A is stored in the condition storage unit 57.

[0132] In Embodiment 2, the image analysis unit 47 determines the examination items involved in the image obtained for the subject M using a learning model 51A. Specifically, by inputting an image such as an X-ray image sent from the image processing unit 33 into the learning model 51A, the learning model 51A infers the examination items to be performed in the X-ray image based on clues such as the constituent elements of the human body reflected in the input X-ray image. The information obtained through inference is output from the learning model 51A as examination item information 71. The condition readout unit 49 inputs the examination item information 71 obtained by the image analysis unit 47 into the APR 53A to set appropriate image acquisition conditions 63 for the examination items involved in the examination item information 71.

[0133] The image analysis unit 47 involved in Example 2 is equivalent to the inspection type inference unit of the present invention.

[0134] Here, use Figure 16 The series of steps in automatically setting the image acquisition condition 63 in Example 2 will be described. As a specific example of an inspection item, the case in which the image acquisition condition 63 is set while PCI is being performed with the catheter Ch propelled to the chest Lc will be described.

[0135] With the catheter Ch propelled to the chest Lc, X-rays are irradiated onto the chest Lc through the X-ray tube 5, thereby generating an X-ray image W2 of the chest Lc as the target area by the image processing unit 33. The data of the generated X-ray image W2 is sent to the image analysis unit 47.

[0136] The image analysis unit 47 inputs the X-ray image W2 as the input image to the learning model 51A. The learning model 51A estimates the location reflected in the X-ray image W2 as input information and outputs that location. In this case, the learning model 51A uses clues such as the heart H, lung Lu, and duct Ch of the subject M reflected in the X-ray image W2 to estimate that the X-ray image W2 is most likely an image for chest PCI. As a result, the learning model 51A outputs information such as chest PCI as the examination item of the input image, i.e., examination item information 71b. The examination item information 71b output from the learning model 51A is sent to the conditional readout unit 49.

[0137] The condition readout unit 49 searches for appropriate image acquisition conditions 63 for acquiring X-ray images of the chest PCI involved based on the examination item information 71b. That is, by inputting the examination item information 71b, which is selected as an examination item, into the APR 53A, the image acquisition conditions 63, i.e., image acquisition conditions 63f, stored in the APR 53A in association with the examination item information 71b are read out and the image acquisition conditions 63f are output.

[0138] As a result, the X-ray tube 5 irradiates X-rays based on X-ray irradiation conditions 65f, and the image processing unit 33 performs image processing on the X-ray image F based on image processing conditions 67f. Therefore, even without manual operation to select examination items, the image acquisition conditions 63f can be quickly and automatically set, thus rapidly improving the visibility of the X-ray image W2 using appropriate image acquisition conditions 63f. The operator confirms the X-ray image W2 while placing the stent in the coronary artery, thereby concluding the PCI procedure.

[0139] In Example 2, by using the learning model 51A and APR 53A, the recently obtained image of the subject M can be used as input information to infer the examination performed in the image, and the image acquisition conditions 63 suitable for the examination can be read out and set. That is, the image acquisition conditions 63 are automatically and appropriately changed not only according to changes in the object irradiated by X-rays, but also according to operations such as the insertion of endoscope Es or the injection of contrast agent Ct.

[0140] Therefore, when using examination equipment or reagents, the burden on the operator and the patient M can be reduced. For example, in the case of an imaging examination using contrast agent Ct, the image acquisition condition 63 corresponding to ordinary X-ray imaging is automatically selected before the injection of contrast agent Ct, and automatically changed to the image acquisition condition 63 corresponding to contrast imaging after the injection of contrast agent Ct. That is, there is no need to manually select the examination item before and after the injection of contrast agent Ct, so the operator can focus on the imaging procedure. Furthermore, when continuously changing the X-ray imaging target site by tracking the injection of contrast agent Ct into the digestive tract or blood vessels, the examination item also changes with the change of target site, and therefore the image acquisition condition 63 automatically changes accordingly. Therefore, the risk of missing the contrast agent due to manually setting the image acquisition condition 63 can be more reliably avoided.

[0141] [Example 3]

[0142] Next, Embodiment 3 of the present invention will be described. For example... Figure 17 As shown, the X-ray device 1B involved in Embodiment 3 differs from that in Embodiment 1 in that the main control unit 39 also has a misanalysis detection unit 73.

[0143] The misinterpretation detection unit 73 is disposed after the image analysis unit 47 and before the conditional readout unit 49. The misinterpretation detection unit 73 is configured to receive the position information of the C-arm 9 constantly detected by the arm position detection unit 37, and to receive the part information 61 output from the image analysis unit 47. Furthermore, the misinterpretation detection unit 73 detects whether misinterpretation of the learning model 51 has occurred based on whether the position information of the C-arm has changed and whether the part information 61 has changed.

[0144] Figure 18 This diagram illustrates the situation where the learning model 51 misinterprets in the image parsing unit 47. Figure 18(a) shows the state in which the learning model 51 normally resolves the X-ray image F2 that reflects the abdomen. In this case, the learning model 51 normally resolves the X-ray image F2 obtained at the specified time Ta, and outputs the abdominal location information 61b after determining that the X-ray image F2 is an image reflecting the abdomen.

[0145] On the other hand, due to noise or fluctuations in pixel values ​​generated in the X-ray image F used as input information, the learning model 51 sometimes incorrectly identifies the parts reflected in the input image. For example, ... Figure 18 As shown in (b), suppose an X-ray image F4 of the abdomen showing a ring-shaped artifact At is acquired at time Tb after time Ta. When the X-ray image F4 of the abdomen showing a ring-shaped artifact At is input into the learning model 51, the following situation occurs: the learning model 51 misidentifies the image of the stomach Ga and the artifact At as an image of the lungs, and the learning model 51 actually performs parsing that the X-ray image F4 showing the abdomen is misidentified as an image showing the chest. In the case of such misparsing, even though the X-ray image F4 of the abdomen is input, the learning model 51 will output chest location information 61c.

[0146] When the location information 61, which was output due to this misinterpretation, is sent to the conditional readout unit 49, the conditional readout unit 49 outputs and sets image acquisition conditions 63c suitable for acquiring X-ray images of the chest. As a result, inappropriate image acquisition conditions 63c were applied when acquiring X-ray images of the abdomen, thus reducing the visibility of the subsequently generated X-ray image F4 of the abdomen.

[0147] Therefore, the X-ray apparatus 1B according to Embodiment 3 prevents the generation of an X-ray image F by applying inappropriate image acquisition conditions 63 due to misinterpretation of the learning model 51 by having a misinterpretation detection unit 73. Figure 19 This is a flowchart of the process by which the misinterpretation detection unit 73 detects whether a misinterpretation of the learning model 51 has occurred.

[0148] The misinterpretation detection unit 73 compares the position information of the X-ray irradiation field with the part information 61 output by the learning model 51 for each frame of the X-ray image F. That is, for the X-ray image acquired at a predetermined time T1, the position information of the X-ray irradiation field when the X-ray image was generated and the part information 61 output by the learning model 51 are sent to the misinterpretation detection unit 73 (step S1).

[0149] In Embodiment 3, the position of the X-ray irradiation field is determined based on the position of C-arm 9. That is, the position of the X-ray irradiation field is detected by the arm position detection unit 37, which detects the position of C-arm 9, and the position information of the X-ray irradiation field is sent from the arm position detection unit 37 to the misinterpretation detection unit 73. Additionally, the part information 61 output by the learning model 51 is sent from the image parsing unit 47 to the misinterpretation detection unit 73. The arm position detection unit 37 corresponds to the irradiation position detection unit of the present invention.

[0150] Then, for the X-ray image acquired at time T2, which is equivalent to one frame from time T1, the location information of the X-ray irradiation field and the part information 61 output by the learning model 51 are sent to the misinterpretation detection unit 73 (step S2).

[0151] The misinterpretation detection unit 73 first compares the location information 61 of the X-ray image acquired at time T1 with the location information 61 of the X-ray image acquired at time T2 to determine whether the location information 61 has changed between time T1 and time T2 (step S3). If the location information 61 has not changed, the misinterpretation detection unit 73 determines that no misinterpretation of the learning model 51 has occurred, and sends the location information 61 obtained at time T2 to the condition readout unit 49 to read out the image acquisition condition 63 (step SR1). Therefore, the X-ray tube 5 and the image processing unit 33 are controlled using the image acquisition condition 63 obtained at time T2 to perform the acquisition of X-ray images after time T2.

[0152] On the other hand, if the location information 61 changes between time T1 and time T2, step S4 is entered. When step S4 is entered, the misinterpretation detection unit 73 compares the position information of the X-ray irradiation field acquired at time T1 with the position information of the X-ray irradiation field acquired at time T2 to determine whether the position of the X-ray irradiation field has changed between time T1 and time T2.

[0153] If it is determined that the X-ray irradiation field has shifted between time T1 and time T2, the misinterpretation detection unit 73 determines that no misinterpretation of the learning model 51 has occurred. Then, the misinterpretation detection unit 73 sends the part information 61 obtained at time T2 to the condition readout unit 49 to read out the image acquisition condition 63 (step SR1). That is, as a result of the X-ray irradiation field shifting and changing the target part of the X-ray irradiation, it is assumed that the part information 61 output by the learning model 51 has changed, so the misinterpretation detection unit 73 determines that no misinterpretation of the learning model 51 has occurred. Therefore, the X-ray tube 5 and the image processing unit 33 are controlled using the image acquisition condition 63 obtained at time T2 to perform the acquisition of X-ray images after time T2.

[0154] On the other hand, if it is determined in step S4 that no displacement of the X-ray irradiation field has occurred between time T1 and time T2, the misinterpretation detection unit 73 determines that a misinterpretation of the learning model 51 has occurred (step S5). Then, the misinterpretation detection unit 73 does not use the part information 61 acquired at time T2, but instead sends the part information 61 acquired at time T1 to the condition readout unit 49 to read out the image acquisition condition 63 (step SR2). Therefore, the X-ray tube 5 and the image processing unit 33 are controlled using the image acquisition condition 63 acquired at time T1 to perform the acquisition of X-ray images after time T2.

[0155] If the X-ray irradiation field does not shift, the location information 61 output by the learning model 51 should also remain unchanged since the irradiated area does not change. Therefore, the situation where the location information 61 output by the learning model 51 changes even though the X-ray irradiation field does not shift between time T1 and time T2 is considered a misinterpretation by the learning model 51. Therefore, if it is determined in step S4 that the position information of the X-ray irradiation field has not changed, it can be determined that the location information 61 obtained at time T2 is incorrect. In this way, by determining whether the X-ray irradiation field has shifted and whether the location information 61 has changed in each frame, the misinterpretation detection unit 73 can detect whether the learning model 51 has misinterpreted.

[0156] use Figure 20 and Figure 21 Specifically, the process of judgment by the misinterpretation detection unit 73 in Embodiment 3 will be described. Figure 20 The procedure is shown in the case where no misinterpretation occurs. At time T1, C-arm 9 is located at the abdomen Lb, and an X-ray image F2 reflecting the abdomen is generated. The image analysis unit 47 inputs the X-ray image F2 as input information to the learning model 51, thereby outputting the abdominal location information 61b. The abdominal location information 61b is sent to the misinterpretation detection unit 73 as the location information 61 involved at time T1. At the same time, the arm position detection unit 37 sends the location information that the position of the X-ray irradiation field at time T1 means that the abdomen Lb is located to the misinterpretation detection unit 73.

[0157] Then, at time T2, corresponding to the next frame after time T1, arm C9 moves towards the chest Lc, generating an X-ray image F3 reflecting the chest. Image parsing unit 47 inputs X-ray image F3 as input to learning model 51, thereby outputting chest location information 61c. Chest location information 61c is sent to misinterpretation detection unit 73 as location information 61 related to time T2. Then, arm position detection unit 37 sends location information to misinterpretation detection unit 73, indicating that the X-ray irradiation field at time T2 is located at chest Lc.

[0158] The misinterpretation detection unit 73 compares the location information 61 related to timing T1 and timing T2 with the position information of the X-ray irradiation field. First, by comparing the location information 61b related to timing T1 with the location information 61c related to timing T2, the misinterpretation detection unit 73 determines that the location information 61 has changed between timing T1 and timing T2 (step S3). Next, by comparing the position information Lb of the X-ray irradiation field related to timing T1 with the position information Lc of the X-ray irradiation field related to timing T2, the misinterpretation detection unit 73 determines that the position of the X-ray irradiation field has changed between timing T1 and timing T2 (step S4).

[0159] Since both the location information 61 and the position information of the X-ray irradiation field have changed, the misinterpretation detection unit 73 determines that the learning model 51 has not misinterpreted and accepts the interpretation result (location information 61c) of the learning model 51 involved in timing T2. Then, the misinterpretation detection unit 73 sends the location information 61c to the condition readout unit 49 (step SR1). In other words, when entering step SR1, the misinterpretation detection unit 73 sends the image acquisition condition 63 output by the learning model 51 after the change in the content of the location information 61 to the condition readout unit 49. As a result, the condition readout unit 49 reads the image acquisition condition 63c and uses the image acquisition condition 63c to generate the X-ray image after timing T2.

[0160] Figure 21 The procedure in the event of misinterpretation is illustrated. At time T1, C-arm 9 is located at the abdomen Lb, generating an X-ray image F2 reflecting the abdomen. The learning model 51 takes the X-ray image F2 as input and outputs abdominal location information 61b. The abdominal location information 61b is sent to the misinterpretation detection unit 73. The arm position detection unit 37 sends position information to the misinterpretation detection unit 73, indicating that the position of C-arm 9 at time T1 is the abdomen Lb.

[0161] Then, at time T2, C-arm 9 does not move from the abdomen Lb, generating an X-ray image F4 with the abdomen Lb as the X-ray irradiation field. Here, the learning model 51 takes the X-ray image F4 as input information, and due to artifacts such as At, it incorrectly resolves the image reflected in the X-ray image F4, setting it as chest location information 61c as output. The chest location information 61c output due to misresolution is sent to the misresolution detection unit 73 as location information 61 involved in time T2. Then, the arm position detection unit 37 sends the location information that the position of the X-ray irradiation field at time T2 means the abdomen Lb.

[0162] The misinterpretation detection unit 73 compares the location information 61 related to timing T1 and timing T2 with the position information of the X-ray irradiation field. First, by comparing the location information 61b related to timing T1 with the location information 61c related to timing T2, the misinterpretation detection unit 73 determines that the location information 61 has changed between timing T1 and timing T2 (step S3). Next, by comparing the position information Lb of the X-ray irradiation field related to timing T1 with the position information Lb of the X-ray irradiation field related to timing T2, the misinterpretation detection unit 73 determines that the position of the X-ray irradiation field has not changed between timing T1 and timing T2 (step S4).

[0163] Because the location information 61 has changed while the position information of the X-ray irradiation field has not changed, the misinterpretation detection unit 73 determines that a misinterpretation has occurred in the learning model 51, and ignores or discards the interpretation result (location information 61c) of the learning model 51 involved in timing T2 as erroneous information (step S5). Then, the misinterpretation detection unit 73 sends the location information 61b obtained at timing T1, which is closest to timing T2, to the condition readout unit 49 (step SR2). In other words, when entering step SR2, the misinterpretation detection unit 73 sends the image acquisition condition 63 output by the learning model 51 before the content of the location information 61 changes to the condition readout unit 49. As a result, the condition readout unit 49 reads the image acquisition condition 63b and uses the image acquisition condition 63b to generate an X-ray image after timing T2.

[0164] In Embodiment 3, a structure is included that detects the occurrence of misinterpretation of the learning model 51 by also having a misinterpretation detection unit 73. The misinterpretation detection unit 73 determines whether the output result of the learning model 51 has changed and whether the position of the X-ray irradiation field has changed. Then, if it detects that the output result of the learning model 51 has changed but the position of the X-ray irradiation field has not changed, it determines that a misinterpretation has occurred in the learning model 51, and discards the latest output result of the learning model 51 (here, the location information 61 obtained at time T2). Then, the most recently obtained output result of the learning model 51 (here, the location information 61 obtained at time T1) is sent to the condition readout unit 49 to read out the image acquisition condition 63.

[0165] The misinterpretation detection unit 73 continuously determines whether the position of the X-ray irradiation field has changed and whether the output of the learning model 51 has changed. Therefore, even if the learning model 51 misinterprets and outputs incorrect location information 61 due to noise in the input information, this incorrect location information 61 can be ignored. Thus, even if the learning model 51 misinterprets due to noise in the input information, it can avoid reading inappropriate image acquisition conditions 63 based on the incorrect location information 61 output due to misinterpretation. Therefore, the impact of misinterpretation by the learning model 51 caused by fluctuations can be reduced, and appropriate image acquisition conditions 63 can be automatically read.

[0166] [Example 4]

[0167] Next, Embodiment 4 of the present invention will be described. In Embodiments 1 and 2, the image acquisition condition 63 read by the condition readout unit 49 is automatically set to generate the X-ray image F. That is, the latest image acquisition condition 63 read by the condition readout unit 49 is automatically sent to the X-ray tube 5 or the image processing unit 33, thereby always controlling the X-ray tube 5 or the image processing unit 33 according to the latest image acquisition condition 63.

[0168] On the other hand, in Embodiment 4, the configuration is such that, under predetermined conditions, an approval step occurs as a pre-step before the step of controlling the X-ray tube 5 or the image processing unit 33 using the image acquisition condition 63 read by the condition readout unit 49. The approval step is the step in which the operator decides whether to set the latest image acquisition condition 63 read by the condition readout unit 49.

[0169] like Figure 22 As shown, the X-ray apparatus 1C according to Embodiment 4 differs from Embodiment 1 in that the main control unit 39 also includes an approval condition determination unit 75, an approval key display control unit 77, and a condition setting control unit 79.

[0170] The approval condition determination unit 75 determines whether a specified condition (approval condition) requiring an approval step has occurred. In Embodiment 4, a change in the content of the information output by the learning model 51 is defined as a specified condition. If the approval condition determination unit 75 determines that the specified condition has been met, it sends a message indicating that an approval step is required to the approval key display control unit 77.

[0171] The approval key display control unit 77 is located after the approval condition determination unit 75 and controls the display unit 35 to display the approval key GK. By sending information indicating that an approval step is required from the approval condition determination unit 75, the approval key display control unit 77 controls the display unit 35 to display the approval key GK. Figure 23An example of how the approval key GK in the display unit 35 is displayed is shown. Details of the approval key GK will be described later.

[0172] The condition setting control unit 79 is located before the condition reading unit 49 and after the approval condition determination unit 75. When a specified condition requiring an approval step is generated, the condition setting control unit 79 controls whether the latest image acquisition condition 63 read by the condition reading unit 49 can be set. The condition setting control unit 79 is configured to control the condition reading unit 49 by being triggered by the operator operating the approval key GK, thereby setting the latest image acquisition condition 63.

[0173] In this embodiment, the condition setting control unit 79 is configured to send a signal (blocking signal ST) to the condition readout unit 49 to prevent the latest image acquisition condition 63 from being set as an X-ray image acquisition condition, under the control of the approval condition determination unit 75. Furthermore, the condition setting control unit 79 is controlled by the operator operating the approval key GK to stop sending the blocking signal ST.

[0174] Figure 24 (b) is a flowchart illustrating a series of operations of the X-ray apparatus 1C according to Embodiment 4. In Embodiment 1, etc., as... Figure 24 As shown in (a), when the condition readout unit 49 reads out the image acquisition condition 63 in step M4, the process automatically and unconditionally proceeds to step M5. That is, the latest readout image acquisition condition 63 is automatically set as the X-ray image acquisition condition and sent to the X-ray tube 5 or the image processing unit 33. Then, the process proceeds to step M6, where the X-ray image F for the next frame is generated according to the latest image acquisition condition 63.

[0175] On the other hand, in Embodiment 4, there is an approval condition determination step Q1 between step M4 and step M5. The approval condition determination step Q1 is a step to determine whether an approval condition has been generated. If it is determined that no approval condition has been generated, that is, if the content of the information output by the learning model 51 in this embodiment has not changed, then step M5 is entered. That is, if no approval condition has been generated, the latest image acquisition condition 63 read by the condition readout unit 49 is set as the condition for X-ray image acquisition.

[0176] If, in the approval condition determination step Q1, the approval condition determination unit 75 determines that an approval condition has been generated, the process proceeds from step Q1 to step M4a to display the approval key GK. As an example, the approval key GK is as follows: Figure 23The approval condition generation area K5 in the display unit 35 is shown as indicated. In addition to displaying the approval key GK and the rejection key DK, the approval condition generation area K5 also displays information J1 indicating that an approval condition has been generated and information J2 indicating whether to approve the latest image acquisition condition 63. The presence or absence of information J1 and information J2, as well as the content of information J1 and information J2, can be appropriately selected.

[0177] After the approval key GK is displayed, the process proceeds to step Q2. In step Q2, the operator selects whether to approve the latest image acquisition condition 63. If the latest image acquisition condition 63 is approved, the operator operates the control panel 41, etc., to select the approval key GK. Examples of operations for selecting the approval key GK include clicking the approval key GK with the mouse on the control panel 41, or touching the approval key GK displayed on the display unit 35, which is a touch panel. The control panel 41 corresponds to the approval instruction input unit of this invention.

[0178] After selecting the approval key GK, the process proceeds from step Q2 to step M5, where the latest image acquisition condition 63 is set as the X-ray image acquisition condition. Then, the process proceeds to step M6, where the X-ray image is generated using the set image acquisition condition 63. That is, the X-ray image F for the next frame is generated using the latest image acquisition condition 63.

[0179] On the other hand, there is a situation where the operator determines that the visibility of the X-ray image F is reduced when the latest image acquisition condition 63 is used. In addition, there is a situation where sufficient visibility is obtained in the X-ray image F obtained by using the image acquisition condition 63 (the most recently set image acquisition condition 63) that is set as the X-ray image acquisition condition at the current time, so the operator determines that it is not necessary to approve the latest image acquisition condition 63.

[0180] Thus, if the operator determines that the latest image acquisition condition 63 is not acceptable, the operator can either suspend the operation by not selecting the "Accept" key GK or select the "Reject" key DK. If the "Reject" key DK is selected, the "Accept" key GK and other information displayed in the acceptance condition generation area K5 will become invisible.

[0181] If the approval key GK is not selected, proceed to step M5a. Upon entering step M5a, the latest image acquisition condition 63 is not set as an X-ray image acquisition condition; instead, the most recently set image acquisition condition 63 is continued to be set as an X-ray image acquisition condition. That is, the image acquisition condition 63, which has been recently set as an X-ray image acquisition condition, continues to be sent to the X-ray tube 5 or the image processing unit 33. Then, proceed to step M6, and generate an X-ray image using the set image acquisition condition 63. That is, the X-ray image F for the next frame onwards is generated by continuing to use the most recently set image acquisition condition 63.

[0182] Here, specific examples are given and used. Figures 25 to 28 The series of steps for setting image acquisition condition 63 in Example 4 will be described. Figure 25 The positions of the X-ray irradiation fields at each timing are shown. In Example 4, as a specific example, X-ray images F are acquired from timing P1 to timing P4. Furthermore, it is assumed that the X-ray irradiation field shifts from the abdomen (Lb) to the chest (Lc) immediately before timing P3. That is, at timings P1 and P2, the X-ray irradiation field is located at the abdomen (Lb), and at timings P3 and P4, the X-ray irradiation field is located at the chest (Lc). Additionally, it is assumed that the approval key GK is activated at timing P4.

[0183] First, the process of setting image acquisition condition 63 at time points P1 and P2 will be explained. The processes up to step M4 are similar to... Figure 10 The procedures in Example 1 are the same. That is, X-rays are irradiated from the X-ray tube 5 to acquire an X-ray image F2 of the abdomen Lb (step M2), and the learning model 51 uses the X-ray image F2 as input and outputs site information 61b (step M3). Then, APR 53 reads out the image acquisition condition 63b associated with the site information 61b as the latest image acquisition condition 63 (step M4).

[0184] In Example 4, as Figure 24 As shown in (b), the process proceeds from step S4 to step Q1. During timing P1 and timing P2, the position of the X-ray irradiation field remains unchanged. Therefore, during timing P1 and timing P2, the approval condition determination unit 75 determines that no approval condition has been generated. In the case where no approval condition has been generated, as... Figure 26 As shown, the approval condition determination unit 75 does not perform special control over the approval key display control unit 77 and the condition setting control unit 79. Therefore, the approval key display control unit 77 does not perform special control over the display unit 35, and thus the approval key GK is not displayed.

[0185] Furthermore, in the absence of an approval condition, the condition setting control unit 79 is not controlled by the approval condition determination unit 75, and therefore no blocking signal ST is sent from the condition setting control unit 79 to the condition readout unit 49. Consequently, the latest image acquisition condition 63 (here, image acquisition condition 63b) read by the condition readout unit 49 is set as the X-ray image acquisition condition (step M5). That is, at timings P1 and P2, the latest image acquisition condition 63b is sent to the X-ray tube 5 or the image processing unit 33, thereby acquiring the X-ray image F according to the image acquisition condition 63b (step M6).

[0186] Next, the process of setting image acquisition condition 63 at time P3 will be explained. The processes up to step M4 are similar to those described above. Figure 11 The procedures in Example 1 are the same. That is, X-rays are irradiated from the X-ray tube 5 to acquire an X-ray image F3 of the chest Lc (step M2), and the learning model 51 uses the X-ray image F3 as input and outputs site information 61c (step M3). Then, APR 53 reads out the image acquisition condition 63c associated with the site information 61c as the latest image acquisition condition 63 (step M4).

[0187] However, the procedures following step M4 in timing P3 differ from those following step M4 in timing P1 and timing P2. The X-ray irradiation field for timing P2 is located in the abdomen (Lb), while the X-ray irradiation field for timing P3 is located in the chest (Lc). That is, as... Figure 25 As indicated by reference numeral L1 in the attached diagram, at time P2, the information output by the learning model 51 is location information 61b; on the other hand, at time P3, the information output by the learning model 51 changes to location information 61c. Therefore, at time P3, the approval condition determination unit 75 determines in step Q1 that an approval condition has been generated.

[0188] If an approval condition is determined to have been met, the approval condition determination unit 75 sends a control signal to the approval key display control unit 77 and the condition setting control unit 79. The approval key display control unit 77 then causes the display unit 35 to display the approval key GK according to the control signal from the approval condition determination unit 75 (step M4a). Specifically, as... Figure 23 As shown, the approval key GK and the rejection key DK are displayed in the approval condition generation area K5 of the display unit 35.

[0189] Furthermore, although an approval condition is generated, the approval key GK is not activated at time P3 (step Q2). Therefore, the condition setting control unit 79 sends a blocking signal ST to the condition readout unit 49 according to the control signal from the approval condition determination unit 75. Consequently, the latest image acquisition condition 63c read by the condition readout unit 49 is not set as an X-ray image acquisition condition, but instead the most recently timed image acquisition condition 63 is set as an X-ray image acquisition condition (step M5a). In other words, the image acquisition condition 63c read at time P3 is not sent to the X-ray tube 5 or the image processing unit 33 (see reference). Figure 27 (The attached figure is labeled RP).

[0190] Furthermore, the most recent timing of timing P3 is timing P2. Therefore, the image acquisition condition 63b, which is set as the X-ray image acquisition condition at timing P2, continues to be set as the X-ray image acquisition condition at timing P3. That is, as Figure 25 As shown by reference numeral L2 in the attached figure, although the image acquisition condition 63c is read out by the condition readout unit 49 at timing P3, the image acquisition condition 63b is set as the X-ray image acquisition condition, and the X-ray image F is acquired using the image acquisition condition 63b (step M6).

[0191] Finally, the process of setting image acquisition condition 63 at time P4 will be explained. The process up to step M4 is the same as the process up to step M4 at time P3. That is, the learning model 51 outputs part information 61c (step M3). Then, APR 53 reads out the image acquisition condition 63c associated with the part information 61c as the latest image acquisition condition 63 (step M4).

[0192] However, no approval key GK is operated during time P3, but during time P4, the operator confirms the approval key GK displayed on the display unit 35 and operates the approval key GK using the operation panel 41 (step M4a, step Q2). By operating the approval key GK, an instruction is input to approve the content of the latest read image acquisition condition 63.

[0193] By operating the approval key GK in step Q2, such as Figure 28 As shown, a signal AN indicating that the blocking signal ST should be stopped is sent to the condition setting control unit 79 via the operation console 41. The condition setting control unit 79 stops sending the blocking signal ST by receiving the signal AN. Therefore, by operating the approval key GK at timer P4, the latest image acquisition condition 63c read by the condition readout unit 49 is set as the X-ray image acquisition condition (step M5). Therefore, the latest image acquisition condition 63c is sent from the condition readout unit 49 to the X-ray tube 5 or the image processing unit 33, thereby acquiring the X-ray image F according to the image acquisition condition 63c (step M6).

[0194] Thus, in Embodiment 4, as a pre-stage step before setting the image acquisition condition 63 as the condition for X-ray image acquisition, an approval step including steps Q1, M4a, and Q2 is performed. In APR 53, the image acquisition condition 63 associated with the site information 61 is determined to be a parameter suitable for cases where the subject's physique is within a general range. Therefore, depending on various conditions such as the subject's physique, there may be cases where the image acquisition condition 63 read out using APR 53 is not actually the most suitable parameter for generating an X-ray image of the subject. In such cases, when the image acquisition condition 63 read from the condition readout unit 49 is automatically set as the condition for X-ray image acquisition, the visibility of the X-ray image is actually reduced for the operator.

[0195] Therefore, in this embodiment, an approval step is set, triggered by the operator's approval of the setting of the latest image acquisition condition 63, to generate an X-ray image F using the parameters involved in the latest image acquisition condition 63. By setting an approval step, the situation where the image acquisition condition 63, which is not actually the most suitable, is automatically set as the X-ray image acquisition condition against the operator's will can be avoided.

[0196] <Effects obtained from the structure of the implementation method>

[0197] (First item) The X-ray apparatus (1) according to this embodiment includes: an X-ray tube (5) that irradiates an object (M) with X-rays; an X-ray detector (7) that detects X-rays that have passed through the object (M); an image processing unit (33) that generates an X-ray image by performing image processing using the detection signal output by the X-ray detector (7); a condition storage unit (57) that stores image acquisition conditions (63) corresponding to object parts of the object (M) in association with the object parts, wherein the image acquisition conditions (63) include at least one of X-ray irradiation conditions (65) and image processing conditions (67); and a learning model storage unit (55) that stores a learning model (51) that performs machine learning by using images of the human body as training images. The learning model (51) is used to infer and output the part of the human body reflected in the image; the object part inference unit (47) infers and outputs the part by inputting at least one of the recently obtained X-ray image (F) and optical image (D) of the subject (M) into the learning model (51); the condition readout unit (49) reads out the image acquisition condition (63) associated with the part information stored in the condition storage unit (57) by selecting the part output by the object part inference unit (47) as part information; and the control unit (39) controls at least one of the control of the X-ray tube (5) and the control of the image processing unit (33) according to the image acquisition condition (63) read out by the condition readout unit (49).

[0198] According to the X-ray apparatus 1 described in the first claim, image acquisition conditions 63, including at least one of X-ray irradiation conditions 65 and image processing conditions 67, are automatically set by using a learning model 51 for inferring object parts reflected in an image and an APR 53 that is associated with appropriate image acquisition conditions corresponding to the object parts. The learning model 51 is configured to infer the parts of the human body reflected in the image and output the parts by using machine learning with an image of the human body as a training image. That is, in the image analysis unit 47, the object parts of the subject M reflected in the input image are inferred and output by inputting an image of the subject M into the learning model 51. The condition readout unit 49 automatically reads out the image acquisition conditions 63 associated with the output object parts of the subject M as part information 61. Therefore, when an X-ray image F of the subject is acquired, the image analysis unit 47 and the condition readout unit 49 automatically read out the appropriate image acquisition conditions 63 for the irradiation field of the X-ray image F. Therefore, even if the object area of ​​the subject M to be irradiated by X-rays changes, the image acquisition conditions 63 suitable for the changed object area are automatically read out. That is, the operator does not need to manually select the object area and set the image acquisition conditions 63, so X-ray fluoroscopy or X-ray radiography can be performed more reliably and quickly under appropriate conditions.

[0199] (Second item) In addition, the X-ray apparatus described in the first item also includes: an irradiation position detection unit (37) that continuously detects the position of the X-ray irradiation field relative to the subject (M); and a misinterpretation detection unit (73) that, when the content of the part output by the learning model (51) changes and the irradiation position detection unit (37) detects the displacement of the X-ray irradiation field, reads out the image acquisition condition (63) from the condition readout unit (49) by selecting the part output by the learning model (51) after the change as the target part, and when the content of the part output by the learning model (51) (61) changes and the irradiation position detection unit (37) does not detect the displacement of the X-ray irradiation field, reads out the image acquisition condition (63) from the condition readout unit (49) by selecting the part output by the learning model (51) before the change.

[0200] According to the X-ray apparatus 1B described in the second item, the misinterpretation detection unit 73 compares the position of the X-ray irradiation field relative to the subject M with the content of the location information 61 output by the learning model 51 to detect whether misinterpretation has occurred in the learning model 51. If the content of the location information 61 output by the learning model 51 changes and the X-ray irradiation field shifts, it can be determined that the change in the location information 61 output by the learning model 51 is not caused by misinterpretation by the learning model 51. Therefore, the misinterpretation detection unit 73 selects the changed location information 61 to read the image acquisition condition 63 from the condition readout unit 49.

[0201] On the other hand, if the content of the location information 61 output by the learning model 51 changes but the X-ray irradiation field remains unchanged, it can be determined that the change in the location information 61 output by the learning model 51 is caused by a misinterpretation of the learning model 51. Therefore, the misinterpretation detection unit 73 does not select the changed location information 61, but instead selects the location information 61 output before the change to read the image acquisition condition 63 from the condition readout unit 49. By having such a misinterpretation detection unit 73, even if the content of the location information 61 output by the learning model 51 changes due to a misinterpretation of the learning model 51 caused by fluctuations such as noise, the content of the location information 61 output due to the misinterpretation can be reliably ignored. Therefore, the situation where an inappropriate image acquisition condition 63 is read out due to a misinterpretation of the learning model 51 can be avoided.

[0202] (Third item) In addition, the X-ray apparatus described in the first or second item also includes an approval instruction input unit (41) for inputting an operator's approval instruction for the image acquisition conditions (63) read by the condition readout unit (49). The control unit (39) is configured to control at least one of the following controls when the approval instruction input unit (41) has been input to approve the image acquisition conditions (63): controlling the X-ray tube (5) to irradiate X-rays according to the X-ray irradiation conditions (65) read by the condition readout unit (49), and controlling the image processing unit (33) to generate an X-ray image according to the image processing conditions (67).

[0203] According to the X-ray apparatus 1C described in the third item, the operator approves the image acquisition condition 63 read by the condition readout unit 49, and thereby controls the X-ray tube 5 or the image processing unit 33 according to the content of the image acquisition condition 63. Due to reasons such as the subject's body size not being within the normal range, there may be cases where the image acquisition condition 63 associated with the part information 61 output by the learning model 51 is actually unsuitable as a condition for acquiring X-ray images of the subject. In this case, by not approving the image acquisition condition 63 read by the condition readout unit 49, the operator can avoid the situation where, against the operator's will, the image acquisition condition 63, which is not actually the most suitable, is automatically set as the X-ray image acquisition condition.

[0204] (Fourth item) The X-ray apparatus (1A) according to this embodiment includes: an X-ray tube (5) that irradiates an object (M) with X-rays; an X-ray detector (7) that detects X-rays that have passed through the object (M); an image processing unit (33) that generates an X-ray image by performing image processing using the detection signal output by the X-ray detector (7); a condition storage unit (57) that stores image acquisition conditions (63) corresponding to the examination items (71) of the object (M) in association with the examination items (71), wherein the image acquisition conditions (63) include at least one of the X-ray irradiation conditions (65) and the image processing conditions (67); and a learning model storage unit (55) that stores a learning model (51A) that uses examination images of the human body as training images. Machine learning is performed to infer the type of inspection in the inspection image and output the type of inspection; the inspection type inference unit (47) infers the type of inspection in the input image and outputs the type of inspection by inputting at least one of the X-ray image (F) and optical image (D) of the subject into the learning model (51A); the condition readout unit (49) reads out the image acquisition condition (63) associated with the inspection item (71) stored in the condition storage unit (57) by selecting the type of inspection output by the inspection type inference unit (47) as the inspection item (71); and the control unit (39) controls at least one of the control of the X-ray tube (5) and the control of the image processing unit (33) according to the image acquisition condition (63) read out by the condition readout unit (49).

[0205] According to the X-ray apparatus 1A described in the fourth item, the image acquisition condition 63, which includes at least one of the X-ray irradiation condition 65 and the image processing condition 67, is automatically set by using a learning model 51A for inferring the type of examination in an examination image and an APR 53A that is associated with appropriate image acquisition conditions 63 corresponding to the examination item information 71 related to the type of examination. The learning model 51A is configured to infer the type of examination in the examination image and output the type of examination by using machine learning to image the examination image of the human body as a training image. That is, in the image parsing unit 47, the type of examination in the input image is inferred and the type of examination is output by inputting the image of the subject M as the input image. The condition reading unit 49 automatically reads the image acquisition condition 63 stored in association with the examination item information 71 by selecting the type of examination of the subject M that has been output as the examination item information 71. Therefore, when an X-ray image F of the subject is acquired, the image analysis unit 47 and the condition readout unit 49 automatically read out the appropriate image acquisition conditions 63 for the examination items of the X-ray image F. Thus, even if the examination item information 71 of the subject M changes, the image acquisition conditions 63 suitable for the changed examination item information 71 are automatically read out. That is, the operator does not need to manually select the object area and set the image acquisition conditions 63, thus enabling more reliable and rapid performance of X-ray fluoroscopy or X-ray radiography under appropriate conditions.

[0206] (Fifth item) In addition, the X-ray apparatus described in the fourth item also includes: an irradiation position detection unit (37) that continuously detects the position of the X-ray irradiation field relative to the subject (M); and a misinterpretation detection unit (73) that, when the content of the inspection item (71) output by the learning model (51) changes and the irradiation position detection unit (37) detects the displacement of the X-ray irradiation field, reads the image acquisition condition (63) from the condition readout unit (49) by selecting the inspection item (71) output after the change; and when the content of the inspection item (71) output by the learning model (51) changes and the irradiation position detection unit (37) does not detect the displacement of the X-ray irradiation field, reads the image acquisition condition (63) from the condition readout unit (49) by selecting the inspection item (71) output before the change.

[0207] According to the X-ray apparatus 1B described in the fifth item, the misinterpretation detection unit 73 compares the position of the X-ray irradiation field relative to the subject M with the content of the inspection item information 71 output by the learning model 51 to detect whether misinterpretation has occurred in the learning model 51. If the content of the inspection item information 71 output by the learning model 51 changes and the X-ray irradiation field shifts, it can be determined that the change in the inspection item information 71 output by the learning model 51 is not caused by misinterpretation by the learning model 51. Therefore, the misinterpretation detection unit 73 selects the changed inspection item information 71 to read the image acquisition condition 63 from the condition readout unit 49.

[0208] On the other hand, if the content of the inspection item information 71 output by the learning model 51 changes but the X-ray irradiation field does not shift, it can be determined that the change in the inspection item information 71 output by the learning model 51 is caused by a misinterpretation of the learning model 51. Therefore, the misinterpretation detection unit 73 does not select the changed inspection item information 71, but instead selects the inspection item information 71 output before the change to read the image acquisition condition 63 from the condition readout unit 49. By having such a misinterpretation detection unit 73, even if the content of the inspection item information 71 output by the learning model 51 changes due to a misinterpretation of the learning model 51 caused by fluctuations such as noise, the content of the inspection item information 71 output due to the misinterpretation can be reliably ignored. Therefore, the situation of reading out an inappropriate image acquisition condition 63 due to a misinterpretation of the learning model 51 can be avoided.

[0209] (Sixth item) In addition, the X-ray apparatus described in the fourth or fifth item also includes an approval instruction input unit (41) for inputting an operator's approval instruction for the image acquisition conditions (63) read by the condition readout unit (49). The control unit (39) is configured to perform at least one of the following controls when the approval instruction input unit (41) has been input to approve the image acquisition conditions (63): controlling the X-ray tube (5) to irradiate X-rays according to the X-ray irradiation conditions (65) read by the condition readout unit (49), and controlling the image processing unit (33) to generate an X-ray image (F) according to the image processing conditions (67).

[0210] According to the X-ray apparatus 1C described in the sixth item, the operator approves the image acquisition condition 63 read by the condition readout unit 49, and thereby controls the X-ray tube 5 or the image processing unit 33 according to the content of the image acquisition condition 63. Due to reasons such as the subject's body size not being within the normal range, there may be cases where the image acquisition condition 63 associated with the part information 61 output by the learning model 51 is actually unsuitable as a condition for acquiring an X-ray image of the subject. In this case, by not approving the image acquisition condition 63 read by the condition readout unit 49, the operator can avoid the situation where, against the operator's will, the image acquisition condition 63, which is not actually the most suitable, is automatically set as the X-ray image acquisition condition.

[0211] <Other Implementation Methods>

[0212] Furthermore, the embodiments disclosed herein are illustrative in all respects and not restrictive. The scope of the invention includes the claims and all modifications within the meaning and scope of their equivalents. As an example, the invention can be modified and implemented as follows.

[0213] (1) In the above embodiments, the image acquisition condition 63 is not limited to a structure that includes both X-ray irradiation condition 65 and image processing condition 67, but may include either one. In addition, the X-ray irradiation condition 65 is not limited to a structure that includes both X-ray fluoroscopy condition 68 and X-ray imaging condition 69, but may include either one.

[0214] (2) In the above embodiments, the X-ray image F used as input information to the learning model 51 can be either an X-ray fluoroscopy image or an X-ray radiograph. Alternatively, X-ray radiography can be performed using the image acquisition condition 63 read from the learning model 51 and APR 53 with the X-ray fluoroscopy image as input information. Alternatively, X-ray fluoroscopy can be performed using the image acquisition condition 63 read from the learning model 51 and APR 53 with the X-ray radiograph image as input information.

[0215] (3) In the above embodiments, an X-ray fluoroscopic imaging device with a C-arm 9 is used as an example of X-ray device 1, but it is not limited thereto. The structure of the present invention can be applied to any radiographic imaging device, such as an X-ray imaging device for ordinary X-ray imaging or a tomographic imaging device.

[0216] (4) In the above embodiment 3, a structure for detecting the displacement of the X-ray irradiation field based on the position of the C-arm 9 detected by the arm position detection unit 37 was shown, but it is not limited to a structure in which the displacement of the X-ray irradiation field is detected by the arm position detection unit 37. As an example, when the X-ray irradiation field is displaced due to the horizontal movement of the top plate 3, a structure for determining the position of the top plate 3 is needed when detecting the displacement of the X-ray irradiation field.

[0217] (5) In the above embodiment 4, the acceptance condition is not limited to the case where the information output by the learning model 51 changes. Other examples of acceptance conditions include cases where the position of the irradiation field relative to the subject changes or cases where the S / N ratio of the most recently obtained X-ray image is below a predetermined threshold. As a necessary condition for the position of the irradiation field to change, cases where the position of the C-arm 9 or the top plate 3 changes.

[0218] (6) In the above embodiment 4, a structure was illustrated in which the approval key GK is displayed together with the X-ray image when the approval key GK is displayed on the display unit 35, but it is not limited to this. When the display unit 35 is equipped with multiple monitors, the X-ray image and analysis results can be displayed on one monitor and the approval key GK can be displayed on other monitors. In addition, the monitor or touch panel for displaying the approval key GK can be arranged on the operation table 41 to display the approval key GK.

[0219] (7) In the above embodiment 4, the structure for inputting the instruction to approve the image acquisition condition 63 is not limited to the structure for operating the approval key GK displayed on the display unit 35, etc., and other structures such as switches or buttons may also be used appropriately. In addition, the method by which the condition setting control unit 79 controls the condition reading unit 49 when the approval condition is generated is not limited to the structure of using the blocking signal ST to block the transmission of the image acquisition condition 63 from the condition reading unit 49 to the X-ray tube 5, etc. When the approval condition is generated, as long as the control is performed such that the latest image acquisition condition 63 is set as the condition for X-ray image acquisition triggered by the input of the instruction to approve the latest image acquisition condition 63, the structure controlled by the condition setting control unit 79 can be appropriately changed. As an example, the following structure is listed: by having the operator input the instruction to approve the latest image acquisition condition 63, the condition setting control unit 79 sends a signal to the condition reading unit 49 to instruct the transmission of the latest image acquisition condition 63 to the X-ray tube 5, etc.

[0220] Explanation of reference numerals in the attached figures

[0221] 1: X-ray device; 3: Top plate; 5: X-ray tube; 7: X-ray detector; 9: C-arm; 17: Collimator; 19: Optical camera; 21: Foot switch; 33: Image processing unit; 35: Display unit; 37: Arm position detection unit; 39: Main control unit; 41: Operating console; 43: Storage unit; 45: Machine learning unit; 47: Image analysis unit; 49: Condition readout unit; 51: Learning model; 53: APR; 55: Learning model storage unit; 57: Condition storage unit; 61: Location information; 63: Image acquisition conditions; 65: X-ray irradiation conditions; 67: Image processing conditions; 68: X-ray fluoroscopy conditions; 69: X-ray imaging conditions; 71: Inspection item information; 73: Misinterpretation detection unit; 75: Approval condition determination unit; 77: Approval key display control unit; 79: Condition setting control unit.

Claims

1. An X-ray device comprising: An X-ray tube irradiates the subject with X-rays. An X-ray detector that detects X-rays that pass through the subject; The image processing unit generates an X-ray image by performing image processing using the detection signal output by the X-ray detector; The condition storage unit stores image acquisition conditions corresponding to the object parts of the subject in association with the object parts, the image acquisition conditions including at least one of X-ray irradiation conditions and image processing conditions; The learning model storage unit stores learning models that perform machine learning by using images of the human body as training images to infer and output the parts of the human body reflected in the images. The object part inference unit infers the part reflected in the input image by inputting at least one of the X-ray image and optical image of the subject into the learning model and outputting the part. The condition readout unit reads out the image acquisition conditions associated with the object part and stored in the condition storage unit by selecting the part output by the object part inference unit as the object part; The control unit controls at least one of the X-ray tube and the image processing unit according to the image acquisition conditions read out by the condition readout unit. The irradiation position detection unit continuously detects the position of the X-ray irradiation field relative to the subject. as well as The error detection unit reads the image acquisition conditions from the condition readout unit when the content of the region output by the learning model changes and the irradiation position detection unit detects the displacement of the X-ray irradiation field, by selecting the region output after the change as the target region; and when the content of the region output by the learning model changes and the irradiation position detection unit does not detect the displacement of the X-ray irradiation field, by selecting the region output before the change as the target region.

2. The X-ray apparatus according to claim 1, characterized in that, It also includes an approval instruction input unit, which is used to input the operator's approval of the image acquisition conditions read by the condition readout unit. The control unit is configured to perform at least one of the following controls when an instruction to approve the image acquisition conditions is input through the approval instruction input unit: controlling the X-ray tube to irradiate X-rays according to the X-ray irradiation conditions read by the condition readout unit, and controlling the image processing unit to generate the X-ray image according to the image processing conditions.

3. An X-ray device comprising: An X-ray tube irradiates the subject with X-rays. An X-ray detector that detects X-rays that pass through the subject; The image processing unit generates an X-ray image by performing image processing using the detection signal output by the X-ray detector; The condition storage unit stores image acquisition conditions corresponding to the examination items of the subject in association with the examination items, the image acquisition conditions including at least one of X-ray irradiation conditions and image processing conditions; The learning model storage unit stores learning models that perform machine learning by using examination images of the human body as training images to infer the type of examination in the examination images and output the type of examination. The examination type inference unit infers the type of examination in the input image and outputs the type of examination based on information including the presence or absence of examination equipment and examination reagents reflected in the input image, by inputting at least one of the X-ray image and optical image of the subject as input images into the learning model. The condition readout unit reads out the image acquisition conditions stored in the condition storage unit in association with the inspection item by selecting the inspection type output by the inspection type inference unit as the inspection item; as well as The control unit controls at least one of the X-ray tube and the image processing unit according to the image acquisition conditions read by the condition readout unit.

4. The X-ray apparatus according to claim 3, characterized in that, It also has: An irradiation position detection unit continuously detects the position of the X-ray irradiation field relative to the subject; and The error detection unit, when the content of the inspection type output by the learning model changes and the irradiation position detection unit detects the displacement of the X-ray irradiation field, reads the image acquisition conditions from the condition readout unit by selecting the inspection type output after the change as the inspection item; when the content of the inspection type output by the learning model changes and the irradiation position detection unit does not detect the displacement of the X-ray irradiation field, it reads the image acquisition conditions from the condition readout unit by selecting the inspection type output before the change as the inspection item.

5. The X-ray apparatus according to claim 3 or 4, characterized in that, It also includes an approval instruction input unit, which is used to input the operator's approval of the image acquisition conditions read by the condition readout unit. The control unit is configured to perform at least one of the following controls when an instruction to approve the image acquisition conditions is input through the approval instruction input unit: controlling the X-ray tube to irradiate X-rays according to the X-ray irradiation conditions read by the condition readout unit, and controlling the image processing unit to generate the X-ray image according to the image processing conditions.

Images