JP7376982B2 - Information processing device, radiography system, information processing method and program - Google Patents

Description

The present invention relates to an information processing apparatus, a radiation imaging system , an information processing method, and a program for capturing and displaying medical images.

Digitalization has become widespread in radiation imaging systems in the medical field, and by generating digital radiation images from radiation emitted by radiation imaging equipment, it has become possible to confirm images immediately after radiography. A digital radiography system improves the workflow compared to conventional film-based radiography methods, and enables faster radiography cycles.

In such a radiography system, when photographing a subject to obtain a digital radiation image, there are cases where imaging failures (hereinafter referred to as "impossible images") occur due to various factors (for example, imaging conditions such as X-ray irradiation conditions). This may occur and re-shooting may be required. Therefore, users such as doctors and radiology technicians must determine whether or not the image is a failure each time a radiography is performed, and if it is determined that the image is a failure, they must enter the reason for the failure and prepare for re-imaging. . As a result of the troublesome and time-consuming process of determining whether the image is defective or inputting the reason for the defect, the inspection time becomes longer and a burden is placed on the user and the subject.

In Patent Document 1, the success or failure of radiography is determined based on the digital radiation image, such as whether the intended imaging object is included, and if it is determined that the radiography has failed, a screen for transitioning to re-imaging is displayed, and the process continues until re-imaging. It has been proposed to reduce the amount of effort required. In Patent Document 2, an image determined to be a poor photograph and a photographing condition are stored in association with each other, and a warning is displayed when the user inputs a photographing condition similar to the stored photographing condition, thereby preventing the photographic failure in advance. It is proposed to do so.

Japanese Patent Application Publication No. 2016-097195 JP2009-268586A

When a user determines whether or not a photo is a failure, the user's thoughts and the intentions of the facility to which the user belongs may be influenced. For example, there is a case where it is required to photograph a knee joint at a particular bending angle from a particular photographing angle. In such a case, it is not possible to appropriately determine photographic defects using a certain standard of the photographing system, as in the techniques described in Patent Documents 1 and 2, and It is not possible to improve the user's effort until re-shooting.

The present invention has been made in view of the above-mentioned problems, and it is an object of the present invention to realize photographic failure determination in accordance with the user's intention.

An information processing device according to one aspect of the present invention has the following configuration. That is,
A user input that affirms or denies the judgment result that the radiographic image is a photo failure by the judgment means for determining whether or not the radiation image is a photo failure, or a user input that negates the judgment result that the radiation image is not a photo failure by the judgment means. and a learning means for learning a function related to a photographic failure determination that determines whether or not the photographic image is a photographic failure by the determination means using the radiation image as an input,
The learning means is configured to determine that the radiation image is a failure when the determination means receives a user input that affirms the determination result that the radiation image is a failure , or when the determination means determines that the radiation image is a failure. If it is determined that the image is not a defective image and a user input that negates the determination result that the image is not a defective image is received , the reason for the image failure is determined using the user-inputted reason for the image failure and inputs the radiation image and imaging conditions. Learn the function to output .

According to the present invention, by learning the function of determining whether or not an image is a failure based on the user's decision as to whether or not the image is a failure, it is possible to perform a failure determination in accordance with the user's intention.

FIG. 1 is a block diagram showing an example of the hardware configuration of a radiation imaging system. FIG. 1 is a block diagram showing an example of a functional configuration of a radiation imaging system according to a first embodiment. 7 is a flowchart showing a photographic failure determination process according to the first embodiment. FIG. 7 is a diagram illustrating an example of a re-shooting transition screen at the time of determining whether the image is defective according to the first embodiment. FIG. 2 is a block diagram illustrating an example of the functional configuration of a radiation imaging system according to a second embodiment. 7 is a flowchart showing a photographic failure determination process according to the second embodiment. FIG. 7 is a diagram illustrating an example of a re-shooting transition confirmation screen at the time of determining whether the image is defective according to the second embodiment. 10 is a flowchart showing a photographic failure determination process according to a third embodiment. 10 is a flowchart illustrating photographic failure determination processing according to the fourth embodiment. FIG. 7 is a block diagram showing an example of the functional configuration of a radiation imaging system according to a fifth embodiment.

Some embodiments of the present invention will be described below with reference to the drawings. Note that the present invention is not limited to the specifically disclosed embodiments, and various modifications and changes can be made without departing from the scope of the claims. Note that in the following embodiments, the term radiation may include, for example, α rays, β rays, γ rays, particle rays, cosmic rays, etc. in addition to X-rays.

<First embodiment>
The system configuration of the first embodiment will be explained using FIGS. 1 and 2. FIG. 1 is a block diagram showing an example of the hardware configuration of a radiation imaging system according to the first embodiment. In the radiation imaging system, an imaging control device 100, a radiation imaging device 110, and a radiation generating device 120 are connected via a network 130. Note that the network 130 may be a wired network or a wireless network.

The imaging control device 100 as an information processing device is constructed from an information processing device such as a computer, and determines image failure by processing radiation images. The imaging control device 100 realizes control of radiography through communication with the radiography apparatus 110. The imaging control device 100 also communicates with the radiation generating device 120 and acquires information (irradiation information) when radiation is irradiated from the radiation generating device 120.

In the imaging control device 100, a network device 101 connects the imaging control device 100 to a network 130. The user input device 102 includes a keyboard, a touch panel, etc., and accepts user operations. The UI display device 103 has, for example, a liquid crystal display, and displays an operation screen and a radiation image. Furthermore, the imaging control device 100 includes a CPU 104 that controls the entire device, a RAM 105 that provides a work space for the CPU 104, and a storage device 106 that stores control programs, radiation images received from the radiation imaging device 110, and the like. The units constituting the photographing control device 100 are connected via a main bus 107, and can mutually transmit and receive data. Note that although the user input device 102 and the UI display device 103 are described as separate devices, these devices may be integrated into an operation/display unit (for example, a liquid crystal panel and a touch panel).

After the radiography apparatus 110 transitions to an imaging ready state according to an instruction from the radiography control apparatus 100, it performs radiography in synchronization with the radiation generation apparatus 120. In this way, the radiation imaging device 110 generates a radiation image based on the radiation emitted from the radiation generation device 120. The radiographic apparatus 110 includes a network device 111 connected to a network 130, a CPU 112 that controls the entire apparatus, a RAM 113 that provides a work space for the CPU 112, and a storage device 114 that stores control programs, generated images, and the like. Furthermore, the radiographic apparatus 110 includes a radiation detection panel 115. The radiation detection panel 115 is composed of, for example, an FPD (Flat Panel Detector), and generates a radiation image by generating an electrical signal according to the radiation dose. The various units constituting the radiation imaging apparatus 110 are connected via a main bus 116, and are capable of mutually transmitting and receiving data. Note that in the radiographic system, the number of radiographic apparatuses 110 is not limited to one, and a configuration using a plurality of radiographic apparatuses may be used.

The radiation generating device 120 detects a radiation irradiation instruction from the irradiation switch 125 and generates radiation from the tube 127 under the set irradiation conditions. The irradiation conditions can be set by a user input device 126 (such as an operation panel) that accepts user operations. The radiation generating device 120 further includes a network device 121 that connects to a network 130, a CPU 122 that controls the entire device, a RAM 123 that provides a work space for the CPU 122, and a storage device 124 that stores control programs and the like. Each part constituting the radiation generating device 120 is connected by a main bus 128, and can mutually transmit and receive data.

FIG. 2 is a block diagram showing an example of the functional configuration of the radiation imaging system according to the first embodiment. Each functional unit shown in FIG. 2 is configured such that, for example, a CPU 104, 112, 122 on each device reads a control program stored in a storage device 106, 114, 124 onto a RAM 105, 113, 123 and executes the control program. It is realized by Note that the implementation form of each functional unit is not limited to this, and may be implemented by, for example, dedicated hardware or by cooperation between hardware and software.

The photographing control device 100 includes a communication section 201 , a system control section 202 , an image processing section 203 , a display control section 204 , an image storage section 205 , a photographic failure learning section 206 , and a photographic failure determining section 207 . The communication unit 201 controls the network device 101 and performs communication. The system control unit 202 transmits, via the communication unit 201, irradiation information indicating the irradiation conditions of the radiation generating device 120 (e.g., tube voltage, tube current, etc.) and imaging information of the radiation imaging device 110 (e.g., presence or absence of a grid, imaging (distance, etc.) and manage each status. Further, the system control unit 202 acquires a radiation image from the radiation imaging apparatus 110 via the communication unit 201. Further, the system control unit 202 realizes the basic functions of the imaging control device 100 and controls the operations of each unit.

The image processing unit 203 processes the radiation image acquired via the system control unit 202 and generates an image to be used on the imaging control device 100. The display control unit 204 displays the image generated by the image processing unit 203 via the UI display device 103. Further, the display control unit 204 displays the determination result by the copy failure determination unit 207 via the UI display device 103. Further, the display control unit 204 reflects the processing (enlargement, rotation, etc.) instructed by the system control unit 202 on the image based on the operation from the user input device 102.

The image storage unit 205 stores the image generated by the image processing unit 203 in association with irradiation information of the radiation generating device 120 and imaging information of the radiation imaging device 110 related to the image. Further, the image storage unit 205 stores an image for which the system control unit 202 has instructed to fail the image based on an operation from the user input device 102 (hereinafter also referred to as a failure image) and the reason for the failure at that time.

The defect learning unit 206 uses at least one or more machine learning algorithms to learn radiographic images designated as defective and the reason for the defect. The photo loss learning section 206 includes a feature amount learning section 261 and a feature amount storage section 262. The feature amount learning unit 261 calculates the failure image from the image storage unit 205, the radiation information of the radiation generating device 120 related to the failure image, the imaging information of the radiation imaging device 110, and the inspection information. Learn the image loss features used for determination. The irradiation information used for learning here is information indicating irradiation conditions such as tube voltage and tube current, for example. Similarly, the imaging information used for this learning includes, for example, the presence or absence of a grid in the radiation imaging apparatus 110, the distance between the radiation focus and the radiation detection panel 115, and the like. Furthermore, the examination information used for learning here includes, for example, the body part to be imaged, gender, and physique. Hereinafter, these pieces of information will be collectively referred to as shooting conditions. As will be explained below, radiation images and imaging conditions are used in determining whether the image is defective, generating the reason for the defect, and learning them.

Further, the feature amount learning unit 261 learns not only the above-mentioned image failure feature amount but also the image failure reason feature amount used to generate the image failure reason. Furthermore, the feature quantity learning unit 261 performs the above-described learning for each imaging region (for example, chest imaging, knee joint imaging, etc.). The feature storage unit 262 stores, for each imaged part, the failure feature and the failure reason feature learned by the feature learning unit 261 for each imaged part. Note that there is no restriction on the specific method used for machine learning in the first embodiment. For example, a convolutinoal neural network may be used to learn the defective features, a recurrent neural network may be used to learn the defective features, or a combination of multiple methods may be used.

The defect determination unit 207 determines whether the radiographic image is defective based on the radiographic image and the imaging conditions of the radiographic image. The photographic failure determination unit 207 includes a photographic failure comparison unit 271 and a photographic failure reason generation unit 272. The photographic failure comparison unit 271 compares and evaluates the photographic failure feature quantity recorded in the characteristic quantity storage unit 262 of the photographic failure learning unit 206 and the feature quantity obtained from the image generated by the image processing unit 203, and determines whether the image is photographic. Determine whether it is a loss or not. When determining that the image is defective, the defect comparison unit 271 instructs the display control unit 204 to display a control screen based on the defect determination result. At this time, the failure comparison unit 271 instructs the display control unit 204 to also display the failure reason for the failure image obtained from the failure reason generation unit 272. The failure reason generation unit 272 generates a failure reason for the failure image based on the failure reason feature recorded in the feature storage unit 262.

The radiation imaging apparatus 110 includes an imaging execution section 211, an image generation section 212, an image correction section 213, and a communication section 214. The imaging execution unit 211 acquires a radiation image by irradiating radiation from the radiation detection panel 115. The imaging execution unit 211 also acquires a dark image (dark output) when no radiation is irradiated from the radiation detection panel 115. The image generation unit 212 generates image information with a small data size based on the radiation image generated by the imaging execution unit 211. In this embodiment, a reduced image is used in which a plurality of adjacent pixels of a radiation image are treated as one pixel, and the pixel value is an average value of the plurality of pixels. The image correction unit 213 corrects the radiation image with a dark image and generates a corrected radiation image. The communication unit 214 controls the network device 111 to perform communication. For example, the communication unit 214 transmits the image information generated by the image generation unit 212 and the radiation image generated by the image correction unit 213 to the imaging control device 100.

The radiation generator 120 includes a generator control section 221 and a communication section 222. The generator control unit 221 generates radiation from the tube 127 based on the irradiation conditions set by the user input device 126 in response to detection of a radiation irradiation instruction from the irradiation switch 125 . The communication unit 222 controls the network device 121 and performs communication. For example, the communication unit 222 notifies the radiation imaging apparatus 110 of an irradiation start notification in response to the generation device control unit 221 detecting a radiation irradiation instruction from the irradiation switch 125. After the communication unit 222 receives an irradiation permission notification from the radiation imaging apparatus 110 in response to this notification, the generator control unit 221 performs control to generate radiation. In this way, the radiation irradiation by the radiation generating device 120 and the imaging operation by the radiation imaging device 110 perform synchronous communication. Furthermore, the generator control unit 221 transmits irradiation information indicating the irradiation conditions of the radiation generator 120 to the imaging control device 100 via the communication unit 222 .

Note that in the first embodiment and the second to fifth embodiments described later, the imaging control device 100 and the radiation imaging device 110 are separate bodies; It may also be executed.

Next, the photographic failure determination process of the photographing control device 100 in the first embodiment will be explained with reference to the flowchart of FIG. 3. Further, the structure of the re-shooting transition screen when determining whether the image is defective will be explained with reference to FIG. 4.

FIG. 3 is a flowchart showing the photographic failure determination process of the photographing control device 100 according to the first embodiment. It is assumed that the photographic failure determination process shown in the flowchart of FIG. 3 is started in response to the start of an examination in which the photographing control device 100 performs photographing control based on a user operation.

In step S300, the system control unit 202 transmits an instruction to prepare for imaging to the radiation imaging apparatus 110 via the communication unit 201. When the radiography apparatus 110 is ready, the radiography apparatus 110 sends a return preparation completion notification to the radiography control apparatus 100 via the communication unit 214. After receiving the notification of completion of preparation, the system control unit 202 puts the photographing control device 100 into a photographing ready state and begins to accept photographing operations (step S301). In this embodiment, conditions for driving the radiation imaging apparatus 110 based on information on the subject to be imaged and examination information such as the body part to be imaged inputted by the user are transmitted as preparation instructions.

In step S301, when the user presses the irradiation switch 125 of the radiation generating device 120 upon receiving the ready state of the imaging control device 100, imaging starts. When imaging starts, the radiation generating device 120 generates radiation from the tube 127. The radiation that has passed through the subject is detected by the radiation detection panel 115 of the radiation imaging apparatus 110. The imaging execution unit 211 acquires a radiation image from the detection signal of the radiation detection panel 115. Thereafter, the image generation unit 212 generates image information, the image correction unit 213 generates a corrected radiation image, and the communication unit 214 transmits the image information and the corrected radiation image to the imaging control device 100. In addition, in parallel with these processes, the radiation generating device 120 transmits irradiation information for the imaging to the imaging control device 100 via the communication unit 222.

In step S302, the system control unit 202 receives image information, radiation images, and irradiation information via the communication unit 201, and the image processing unit 203 generates an image to be used on the imaging control device 100 (generated image). This generated image is a radiographic image suitable for a user to judge whether the image is defective or for medical diagnosis. Thereafter, the display control unit 204 displays the generated image on the radiation image display screen displayed on the UI display device 103. FIG. 4 shows an example of a radiation image display screen 420 that displays an image 421 that is a generated image. Hereinafter, the generated image will be explained as a radiation image acquired by the imaging control device 100.

In step S303, the photographic failure determination unit 207 determines whether or not the radiographic image is a photographic failure, using a photographic failure determination using a radiation image and radiographic image imaging conditions as input (photographic failure determination). More specifically, the image failure comparison unit 271 of the image failure determination unit 207 receives as input the image generated by the image processing unit 203, the examination information (including the area to be imaged) used for the imaging preparation instruction, and the imaging irradiation information. , determine whether the radiographic image is defective or not. This photographic failure determination is performed based on the photographic failure feature amount for each imaging region used in the photographing preparation instruction of the feature amount learning unit 261 in the photographic failure learning unit 206. If it is determined in step S303 that the image has not been photographed, the process proceeds to step S304, and if it is determined that the image has not been photographed, the process proceeds to step S305.

In step S304, the system control unit 202 determines whether the user operation received via the display control unit 204 is a copy failure instruction. This means that the image 421 displayed on the radiation image display screen 420 has been determined by the user to be a failure, although the failure determination unit 207 has not determined that the image 421 is a failure. In this way, the process branches depending on the user input that affirms or negates the determination result by the copy failure determination unit 207. The user can notify the system control unit 202 that the image has been determined to be a failure, for example, by pressing the failure button 422. In step S304, if it is determined that the instruction is a defective copy, the process proceeds to step S305, and if it is not determined that the instruction is a defective copy (in this embodiment, the OK button 423 is pressed), the process proceeds to step S305. The process advances to step S311.

The process of step S305 is a process performed when it is determined in step S303 that the result is a failure, or when it is determined that a failure instruction has been given in step S304. The failure reason generation unit 272 generates a radiographic image when the radiation image is determined to be a failure by the failure determination unit 207, or when the radiation image is not determined to be a failure by the failure determination unit 207 but is specified as a failure by user input. , a radiation image and imaging conditions are input to generate the reason for the failure. More specifically, in step S305, the failure reason generation unit 272 inputs the radiation image determined to be failure, the examination information used for the imaging preparation instruction, and the irradiation information for the imaging, and generates the feature quantity learning unit. A reason for failure to photograph is generated based on the reason for failure to photograph for each body part in H.261. The failure reason generation unit 272 compares and evaluates the failure reason feature stored in the feature storage unit 262 and the failure reason feature obtained from the input information, and generates a failure reason. do. At this time, the failure reason generation unit 272 generates a plurality of failure reason candidates based on the failure reason feature amount corresponding to the photographed body part. In the present embodiment, the reason for failure is generated based on the degree of match between the failure reason feature acquired from the input information and the failure reason feature stored in the feature storage unit 262. shall be.

In step S306, the display control unit 204 displays the determination result by the copy failure determination unit 207. More specifically, the display control unit 204 changes the radiographic image display screen 420 displayed on the UI display device 103 to a re-imaging transition screen 400 (FIG. 4) that reflects the reason for the failure generated by the failure determination unit 207. Display on top. Note that in this embodiment, other processing operation instructions and imaging completion instructions from the user on the radiation image display screen 420 are not accepted until the user completes the processing on the re-imaging transition screen 400.

FIG. 4 shows an example of a re-shooting transition screen 400 when the shooting control device 100 determines that the image is defective. As shown in FIG. 4, the re-imaging transition screen 400 is displayed on the radiographic image display screen 420 when it is determined that the image has failed. The image 421 is a radiation image that is subject to image failure determination. The reshooting transition screen 400 includes an input area 401 for inputting the reason for shooting failure, an area for displaying candidate reasons for shooting failure (hereinafter referred to as candidate display area 402), a confirmation button 403 for confirming execution of reshooting, and a confirmation button 403 for confirming the execution of reshooting. It has a cancel button 404 for instructing cancellation of shooting.

The input area 401 is an area that accepts the reason for copy failure that is input by user operation (text input operation) via the display control unit 204. In addition, when displaying the reshooting transition screen 400, the display control unit 204 determines whether the failure reasons generated by the failure determination unit 207 match the failure reason features stored in the feature storage unit 262. The reason for the failure with the highest degree of failure is displayed in advance in the input area 401. The user may use the reason for the failure in the image that was previously input into the input area 401 when the reshooting transition screen 400 is displayed, or may input a new reason for the failure in the image via the display control unit 204. Furthermore, a reason for copy failure selected from candidates for copy failure reasons displayed in candidate display area 402 may be displayed in input area 401 .

When the reshooting transition screen 400 is displayed, the candidate display area 402 displays the reasons for failure other than the reason for failure displayed in the input area 401 among the failure reason candidates generated by the failure reason generation unit 272. Display as a possible reason. In the candidate display area 402, the plurality of failure reasons are displayed in descending order of degree of coincidence with the failure reason feature quantity, that is, in order of likelihood. The confirm button 403 is a button for confirming that the image on the radiation image display screen (not shown) is a failure, and for transitioning to examination start processing for re-imaging. At this time, the feature amount learning unit 261 links the reason for the failure of the image input in the input area 401 as the reason for the failure of the image. The cancel button 404 is a button for determining that the image on the radiation image display screen 420 is not a defective image and for transitioning the display back to the radiation image display screen.

Note that the reshooting transition screen 400 is not limited to the above format. In other words, it is sufficient to have an area for inputting the reason for failure in photographing through a user operation, an area for displaying a plurality of reasons for failure in photography from the failure judgment unit 207, and a scanning unit for instructing whether or not to control re-photography.

In step S307, the system control unit 202 determines whether or not the radiographic image is a failure in response to a user input that affirms or negates the determination result by the failure determination unit 207. More specifically, in this embodiment, the system control unit 202 determines whether the user operation received via the display control unit 204 is a reshooting instruction (whether the confirm button 403 was pressed or the cancel button 404 was pressed). ) is determined. If the user operation is an instruction to reshoot by pressing the confirm button 403 (if the judgment result of the defect judgment is affirmative), the process returns to step S300 through the learning process from steps S308 to S309, and the system control unit 202 begins preparations for reshooting. If the user operation is an instruction to cancel reshooting by pressing the cancel button 404, the process advances to step S310. In steps S308 and S310, the imaging failure determination function of the imaging failure determination unit 207 is trained based on the radiographic image, the imaging conditions, and the user's decision as to whether or not the image is a failure. Further, in step S309, the function of generating a reason for failure in an image in the failure reason generation unit 272 is learned based on the radiation image, the imaging conditions, and the determined reason for failure to image. Learning will be further explained below.

In step S308, the feature amount learning unit 261 in the image loss learning unit 206 learns the image loss feature amount for each imaging region based on the radiation image, the examination information used for the imaging preparation instruction, and the imaging irradiation information. conduct. Thereafter, the defective feature quantity is stored in the feature quantity storage unit 262 for each photographed region. Subsequently, in step S309, the feature amount learning unit 261 in the image failure learning unit 206 performs the following operations based on the radiation image, the examination information used for the imaging preparation instruction, the imaging irradiation information, and the determined reason for the image failure. The feature values for reasons for image failure are learned for each area to be imaged. The determined reason for copy failure is the reason for copy failure written in the input area 401. Thereafter, the feature amount for the reason for photographing failure is stored in the feature amount storage unit 262 for each imaging site used in the imaging preparation instruction. When the learning process from steps S308 to S309 is thus completed, the process returns to step S300 and the system control unit 202 starts preparations for re-photographing.

As described above, in steps S306 to S309, the re-imaging transition screen 400 functions as an input unit for inputting the reason for failure when the radiation image is designated as failure. In steps S307 and S308, at least one or more machine learning algorithms are used to learn the radiation image designated as defective and the reason for the defect. As a result, the defect determination unit 207 in step S303 uses the result of the above learning to determine whether the newly captured radiation image is a defect.

On the other hand, if the user's operation is an instruction to cancel reshooting by pressing the cancel button 404 (if the determination result of the defect determination is negative), the process proceeds from step S307 to step S311 via the learning process of step S310. In step S310, the feature amount learning unit 261 performs learning processing equivalent to step S308. However, in step S308, learning is performed as a feature amount that is a defective image (a defective feature amount), whereas in step S310, learning is performed as a feature amount that is not a defective image. In this way, the photographic failure learning unit 206 learns the function of the photographic failure determining unit 207 for determining whether or not the photographic image is a photographic failure, based on the radiation image, the imaging conditions, and the user's determination as to whether the photographic image is a photographic failure. Note that if it is determined in step S304 that there is an instruction to take a bad picture, but no re-photographing instruction is given in step S307, the learning in step S310 may be skipped.

In step S311, the system control unit 202 receives a user's operation via the display control unit 204, and performs processing based on the operation. The user can perform image enlargement/reduction, gradation processing, examination information input processing, etc. via the operation section 424 of the radiation image display screen 420 by the system control section 202. Further, the user can instruct the end of the examination from the operation unit 424 of the radiation image display screen 420. When the end of the examination is instructed, the system control unit 202 ends the examination, and saves the image generated by the image processing unit 203 in the storage device 106 via the image storage unit 205, or stores it in an external device (not shown). Performs processing such as transferring images to storage.

As described above, in the first embodiment, the imaging control device 100 collects the radiographic image that has been determined to be a failure and has been instructed to be re-exposed, the examination information used for the imaging preparation instruction, the irradiation information for imaging, and the reason for the failure. Based on this, we learn the failure features and the failure reason features. Similarly, the imaging control device 100 calculates the imaging failure feature amount based on the radiation images that were determined to be imaging failures but for which re-imaging was not instructed, the examination information used for imaging preparation instructions, and the irradiation information for the imaging. learn. Based on the photographic failure feature quantity and the photographic failure reason feature quantity that have been learned in this way, the photographing control device 100 performs a photographic failure judgment using the photographed image, the inspection information used for the photographing preparation instruction, and the photographing irradiation information as input. , and generates a reason for failure when determining whether the image is defective. The fact that re-shooting has been executed means that the user has affirmed the defect judgment by the photo-defective determination unit 207 and the reason for the defect, and the fact that re-photographing has not been executed means that they have been denied. That's true.

That is, according to the first embodiment, it is determined whether the radiographic image is a defective image or not in response to a user input affirming or denying the determination result of the defective image determination (steps S304, S307). Then, based on the radiation image, the imaging conditions, and the user's decision as to whether or not the image is a failure, the function of determining the failure in the above-mentioned determination function is learned. In this way, the user's judgment is reflected in the learning of the function of determining whether the image is defective or the function which generates the reason for the defective image. Therefore, it is possible to determine whether a photograph has been taken in accordance with the thoughts of the user or the intentions and concerns of the facility to which the user belongs, and it is possible to reduce the time and effort required to re-photograph due to the occurrence of a photograph failure.

In addition, using the radiographic image areas included in the imaging conditions, learning to determine whether the radiographic image is defective is performed for each image area, and whether or not the radiographic image is defective is determined using the learned image defect determination function for each image area. By determining this, the accuracy of determining whether the image is defective can be improved. In addition, by providing a failure reason generation unit 272 that generates a reason for failure when a radiation image is determined to be failure by the failure determination, the failure reason is automatically generated. The user's effort can be further reduced. Furthermore, by learning the function of the reason for failure generation unit 272 using the reason for failure determined by the user's re-shooting instruction, it becomes possible to automatically generate a reason for failure to photograph that is appropriate to the user's intention. Furthermore, by learning the function for generating the reason for failure in photographing for each body part, it is possible to generate the reason for failure in photography with higher accuracy.

<Second embodiment>
Next, a second embodiment will be described. In the configuration of the first embodiment, on the reshooting transition screen 400, the user can input any reason for the failure of the shot via the input area 401. Then, the feature quantity learning unit 261 learns the copy failure reason feature based on the input copy failure reason, and stores the result in the feature quantity storage unit 262. This means that, for example, learning is performed even if the reason for failing a photograph entered by the user is not an appropriate reason, and the learning accuracy of the photography control device 100 decreases, resulting in a reason for failing a photograph that is not intended by the user. may cause this. Therefore, a problem arises that does not lead to reducing the effort required to re-photograph due to the occurrence of photographic errors.

Therefore, in the second embodiment, the photographing control device 100 determines the validity of the reason for failure when the user inputs the reason for failure, and suppresses execution of learning using an inappropriate reason for failure. The hardware configuration of the radiographic system according to the second embodiment is the same as that of the first embodiment (FIG. 1). Hereinafter, differences from the first embodiment will be mainly explained using FIGS. 5 to 7.

FIG. 5 is a block diagram showing an example of the software configuration of the radiation imaging system according to the second embodiment. The photographic failure determination unit 207 of the photographing control device 100 includes a photographic failure reason comparison unit 500 that functions as an evaluation unit that evaluates the photographic failure reason input by the user to the re-photography transition screen 400. If the evaluation value of the reason for failure is less than or equal to a predetermined value, the failure reason comparison unit 500 displays a confirmation screen for inquiring the user whether or not to adopt the determined reason for failure. This will be explained in more detail below.

For example, the failure reason comparison unit 500 compares the failure reason feature recorded in the feature storage unit 262 of the failure learning unit 206 with the feature amount of the failure reason input by the user into the input area 401. Then, the reason for copy failure input by the user is evaluated. For example, the degree of similarity between the reason for copying failure generated by the copying failure reason generation unit 272 and the text of the copying failure reason input by the user is evaluated. If the evaluation value obtained by evaluating the reason for failure in the image is less than a predetermined threshold, the reason for failure in image comparison unit 500 instructs the display control unit 204 to display a control screen based on the evaluation result. In addition, in the second embodiment, when evaluating the reason for copy failure, the judgment is made by evaluating the similarity of the sentence with the reason for copy failure generated by the copy failure reason generation unit 272. There are no restrictions on the specific method. For example, using an evaluation index for caption generation such as CIDer or METEOR, the percentage of terms included in the sentence of the reason for failure input by the user in the sentence of the reason for failure generated by the failure reason generation unit 272 is calculated. On the other hand, the determination may be made using a certain percentage as the evaluation value.

FIG. 6 is a flowchart showing the photographic failure determination process of the photographing control device 100 according to the second embodiment. In FIG. 6, the same reference numerals as in the first embodiment are given to the same processing steps as the photographic defect determination process (FIG. 3) of the first embodiment.

In step S600, the system control unit 202 determines whether the user has input any reason for copy failure via the input area 401. If it is determined that the user has inputted the reason for copy failure, the process proceeds to step S601, and if it is determined that the user has not input the reason for copy failure, the process proceeds to step S308.

In step S601, the failure reason comparison unit 500 of the failure determination unit 207 evaluates the failure reason input by the user. If the evaluation value of the reason for copy failure input by the user is less than or equal to the predetermined threshold, the process proceeds to step S602, and if the evaluation value is greater than the predetermined threshold, the process proceeds to step S308.

In step S602, the photographic failure determination unit 207 generates a confirmation screen to continue reshooting. The display control unit 204 displays the generated confirmation screen on the reshoot transition screen 400 displayed on the UI display device 103. In the second embodiment, other control operation instructions from the user on the radiation image display screen and imaging completion processing are not accepted until the user completes the processing on the confirmation screen to continue re-imaging.

FIG. 7 is a diagram showing an example of the configuration of a confirmation screen 700 for continuing reshooting. The confirmation screen 700 has a continue reshooting process button 701 and a cancel reshooting process button 702. The re-imaging process continuation button 701 is a button for confirming that the image on the radiation image display screen is a failure and transitioning to examination start processing for re-imaging. At this time, the reason for the failure of the image input in the input area 401 is linked as the reason for the failure of the image. The re-shooting process cancel button 702 is a button for controlling the display to return to the re-shooting transition screen 400 after determining that the reason for the failure of the shot input by the user via the input area 401 is not appropriate.

In step S603, the system control unit 202 determines whether a reshooting instruction has been issued by the user operation received via the display control unit 204 (whether the reshooting process continue button 701 has been pressed or the reshooting process cancel button has been pressed). 702 is pressed). If the user operation is an instruction to press the re-imaging process continuation button 701, the system control unit 202 executes the processes from step S308 to step S309, and shifts the process to step S300 in order to perform re-imaging. If the user operation is an instruction to press the re-shooting process cancel button 702, the system control unit 202 shifts the process to step S306 and displays the re-shooting transition screen 400 again. This allows the user to re-enter the reason for copy failure.

As described above, in the second embodiment, the photographing control device 100 evaluates the reason for failure of photography input by the user, thereby reducing learning accuracy when the reason for failure of photography input by the user is not appropriate as a reason. It becomes possible to prevent In other words, it is possible to prevent the occurrence of a problem that does not lead to reducing the effort required to re-photograph due to the occurrence of a defective image due to the generation of a reason for a defective image that is not intended by the user.

<Third embodiment>
Next, a third embodiment will be described. In the first embodiment, a photographed image is determined to be a failure based on a user's operation, and the failure feature of the image is learned by the feature storage unit 262. This is done even if, for example, the user inadvertently determines that the image has failed, and the photographing control device 100's criteria for determining the failure of the image is lowered. Therefore, a problem arises in that even if the user determines that the photographed image is not a good photograph, the photographing control device 100 determines that the photographed image is not a bad photograph.

Therefore, in the third embodiment, it is possible to switch the learning mode in the photographing control device 100, and when the learning mode is set, the processing for learning the feature quantity of photographic failure is changed. The hardware configuration and functional configuration of the radiation imaging system according to the third embodiment are the same as those of the first embodiment (FIGS. 1 and 2). Hereinafter, differences from the first embodiment will be mainly explained using FIG. 8.

FIG. 8 is a flowchart showing the photographic failure determination process of the photographing control device 100 according to the third embodiment. Note that in FIG. 8, the same reference numerals as in the first embodiment are given to processing steps similar to the copy failure determination process (FIG. 3) of the first embodiment.

In step S800, the system control unit 202 determines whether the learning mode of the copy failure determination unit 207 is valid. In the third embodiment, the user can select in advance whether to enable a learning mode in which learning by the feature amount learning unit 261 is performed on a radiation imaging setting screen (not shown) provided by the system control unit 202. shall be taken as a thing. When the learning mode of the photo failure determination unit 207 is set to be valid, learning of the photo failure feature amount and the photo failure reason feature amount is performed in steps S308 and S309. On the other hand, if the learning mode of the photographic defect determination unit 207 is set to invalid, the process directly proceeds to step S300 in order to prepare for re-photographing.

Similarly, in step S801, the system control unit 202 determines whether the learning mode of the copy failure determination unit 207 is valid. When the learning mode of the photographic defect determination unit 207 is set to be valid, learning of the photographic defect feature amount (learning when the photographic defect is not a defective image) in step S310 is performed. On the other hand, if the learning mode of the copy failure determination unit 207 is set to invalid, the process advances to step S311.

As described above, in the third embodiment, the imaging control device 100 determines whether or not to perform learning by the feature quantity learning unit 261 based on the learning mode setting by the user. That is, even if the user inadvertently determines that the image has failed, the photographing control device 100 does not perform learning, thereby making it possible to reduce the possibility of erroneously determining that the image is a failure.

<Fourth embodiment>
Next, a fourth embodiment will be described. In the configuration of the first embodiment, the photographed image is determined to be a failure based on the user's operation, and re-photographing is performed. Therefore, the failure learning unit 206 has sufficiently learned, and the accuracy of the failure judgment by the failure judgment unit 207 and the accuracy of the failure reason generated by the failure reason generation unit 272 are on par with the user's judgment. Even if this occurs, reshooting will not be performed unless the user performs an operation. This means that even if the image capturing control device 100 can clearly determine that the image is a defective image on the system, a re-photographing control process is required by the user's operation, which does not lead to reducing the effort required to re-photograph due to the occurrence of a defective image. This will cause problems.

Therefore, in the fourth embodiment, an automatic re-photographing function is provided for automatically shifting the photographing control device 100 to re-photographing when a photographic failure occurs. The hardware configuration and functional configuration of the radiation imaging system according to the fourth embodiment are the same as those of the first embodiment (FIGS. 1 and 2). Hereinafter, differences from the first embodiment will be mainly explained using FIG. 9.

FIG. 9 is a flowchart showing the photographic failure determination process of the photographing control device 100 according to the fourth embodiment. Note that in FIG. 9, the same reference numerals as in the first embodiment are given to processing steps similar to the photographic failure determination process (FIG. 3) of the first embodiment.

In step S900, the system control unit 202 determines whether the automatic reshooting function is valid. In the fourth embodiment, on a radiography setting screen (not shown) provided by the system control unit 202, it is possible to select whether to enable or disable the automatic re-imaging function. The automatic re-shooting function automatically transitions to preparation for re-shooting (step S300) without the user making a decision via the re-shooting transition screen 400 when the photo-defective determination unit 207 determines that the photo is defective. It is a function. If it is determined in step S900 that the automatic re-shooting function is valid, the system control unit 202 starts preparations for re-shooting without waiting for a user operation on the re-shooting transition screen 400. That is, when the automatic difference shooting function is set to be valid, the process proceeds from step S900 to step S308, and through learning of the failure feature quantity (step S308), learning of the failure reason feature quantity (step S309), Return to step S300.

On the other hand, if the automatic re-shooting function is set to be disabled, the process proceeds from step S900 to step S306, and the system control unit 202 performs display processing of the re-shooting transition screen 400 via the display control unit 204.

As described above, in the fourth embodiment, the photographing control device 100 determines whether to transition to the re-photographing transition screen 400 based on the user's setting of the automatic re-photographing function of the photographic failure determination unit 207. That is, when the photographing control device 100 determines that the image is a failed image, preparations for re-photographing can be immediately started, and the effort required to re-photograph due to the occurrence of a failed image can be further reduced.

<Fifth embodiment>
Next, a fifth embodiment will be described. In the configuration of the first embodiment, a photographed image is determined to be a failure based on a user's operation, and the failure feature of the image is learned by the feature learning unit 261. The machine learning algorithm used in the first embodiment requires large-scale learning data to improve accuracy. Therefore, in a state where the photographing control device 100 is not operating sufficiently and cannot sufficiently learn the feature amount of photographic failure, there is a possibility that appropriate photographic failure judgment cannot be made due to insufficient learning data. That is, the photographing control device 100 determines that a photographed image that is not a photographic failure is a photographic failure, or determines that a photographic image that is a photographic failure is not a photographic failure, thereby reducing the effort required for re-photographing due to the occurrence of a photographic failure. The problem is that it is not possible to

Therefore, in the fifth embodiment, the photographing control device 100 is communicably connected to an external device, and is configured so that the external device can specify and obtain learning results of the feature quantities of photographic failure and/or photographic failure reasons to be provided. do. The hardware configuration of the radiographic system according to the fifth embodiment is the same as that of the first embodiment (FIG. 1). Hereinafter, differences from the first embodiment will be mainly explained using FIG. 10.

FIG. 10 is a block diagram showing an example of the functional configuration of a radiographic system according to the fifth embodiment. The information processing apparatus 1000 is composed of, for example, a computer having a CPU and a memory, and includes a communication section 1001 and a copy loss learning section 1002. The communication unit 1001 controls the network device of the information processing apparatus 1000 and performs communication.

The photo loss learning unit 1002 includes a feature learning unit 1010, a feature storage unit 1011, and a feature communication unit 1012. The feature amount learning section 1010 has the same function as the feature amount learning section 261. In addition, the feature amount learning unit 1010 uses the information about image failures received from a plurality of image capture control devices as image failure feature values and image failure reason characteristics for each individual image capture control device and for the plurality of image capture control devices as a whole. Learn as quantity. The feature storage unit 1011 has the same function as the feature storage unit 262. In addition, the feature amount storage unit 1011 stores the image failure feature amount and image failure reason feature amount for each individual image capture control device and for the plurality of image capture control devices as a whole, which are generated by the feature amount learning unit 1010. The feature quantity communication unit 1012 receives information about photographic failure from the photographing control device 100. Further, the feature quantity communication unit 1012 acquires from the feature quantity storage unit 1011 and transmits to the photographing control apparatus 100 the photographic failure characteristic quantity and the photographic failure reason characteristic quantity requested by the photographing control apparatus 100 .

In the photographing control device 100, the photographic failure learning section 206 includes a feature communication section 1100. The feature quantity communication unit 1100 transmits to the information processing apparatus 1000 the information on the photographic failure, which becomes the learning data for the photographic failure feature quantity and the photographic failure reason feature quantity which the photographic failure learning unit 206 performs in steps S308 and S309. Further, the feature quantity communication unit 1100 acquires the image failure feature quantity and the image failure reason feature quantity generated by the feature quantity learning unit 1010 from the information processing device 1000, and stores them in the feature quantity storage unit 262. Note that in the fifth embodiment, the system control unit 202 provides a screen (communication setting screen) for setting up communication with the information processing apparatus 1000. Further, on the communication setting screen, the user can select in advance whether to send/receive the information on photo failure to the information processing device 1000, and which shooting control device (or all shooting control devices) to acquire the feature values when receiving the information. shall be taken as a thing. For example, in the imaging control device 100, the user determines which imaging control device to acquire the imaging failure feature amount and imaging failure reason feature amount, and which imaging region's imaging failure feature amount and imaging failure reason feature amount. You can specify whether to obtain the . Then, the feature amount communication unit 1100 can obtain from the information processing device 1000 the failure feature amount and failure reason feature amount of the designated imaging region from another imaging control device designated by the user.

As described above, in the fifth embodiment, the photographing control device 100 performs the photographic failure determination using the photographic failure feature amount and the photographic failure reason feature amount based on the plurality of photographic control devices learned by the information processing device 1000. It can be carried out. That is, even when the imaging control device 100 is not fully operating, it is possible to use feature amounts using large-scale learning data from another imaging control device. Therefore, the photographing control device 100 can reduce the number of times the photographing control device 100 determines that a photograph has failed due to insufficient learning, and can further reduce the effort required to re-photograph due to the occurrence of a photographic failure.

(Other examples)
The present invention provides a system or device with a program that implements one or more of the functions of the embodiments described above via a network or a storage medium, and one or more processors in the computer of the system or device reads and executes the program. This can also be achieved by processing. It can also be realized by a circuit (for example, ASIC) that realizes one or more functions.

100: Radiography control device, 110: Radiography device, 120: Radiation generation device, 201: Communication unit, 202: System control unit, 203: Image processing unit, 204: Display control unit, 205: Image storage unit, 206: Photographer Loss learning section, 207: Photo loss judgment section

Claims (18)

In response to a user input that affirms or denies the judgment result that the radiation image is a photo failure by the determination means for determining whether or not the radiation image is a photograph failure, or a user input that negates the determination result that the radiation image is not a photograph failure by the determination means. , comprising a learning means for learning a function related to photographic failure determination that determines whether or not the photographic image is a photographic failure by the determination means using the radiation image as input;
The learning means is configured to determine that the radiation image is a failure when the determination means receives a user input that affirms the determination result that the radiation image is a failure, or when the determination means determines that the radiation image is a failure. When it is determined that the image is not a defective image and a user input that negates the judgment result that the image is not a defective image is received, the reason for the image failure is output using the radiographic image and the imaging conditions as input, using the user input reason for the defective image. An information processing device that learns functions. The learning means learns a function related to the photographic failure determination in the determination means for each imaged part,
The information processing apparatus according to claim 1, wherein the determination unit determines whether or not the radiation image is a failure using a function learned for the imaging site. 3. The information processing apparatus according to claim 1, further comprising a display unit that receives the radiographic image and the imaging condition as input and displays the reason for failure outputted by the function of outputting the reason for failure. The imaging conditions include the imaging site of the radiographic image,
The learning means learns a function related to outputting the reason for photographing failure for each part to be photographed,
4. The information processing apparatus according to claim 3, wherein the reason for failure to capture the radiographic image is output using a function related to outputting the reason for failure to capture, which is learned for the imaging site included in the imaging condition. The function related to outputting the reason for image failure outputs a plurality of candidate reasons for image failure corresponding to the imaged region for the radiographic image,
5. The information processing apparatus according to claim 4, wherein the display means displays the plurality of rejection reason candidates. 6. The information processing apparatus according to claim 5, wherein the display means displays the plurality of copy failure reason candidates in order of likelihood. Further comprising an evaluation means for evaluating the reason for copy failure input by the user,
3. The display means, if the evaluation value of the evaluation by the evaluation means is less than or equal to a predetermined value, displays a confirmation screen for inquiring the user whether or not to adopt the user-inputted reason for copy failure. 7. The information processing device according to any one of 6 to 6. 2. The evaluation means evaluates the user-input reason for failure based on a comparison between the reason for failure input by the user and the reason for failure output by the function related to the output of the reason for failure. 7. The information processing device according to 7. further comprising a first setting means for setting whether or not learning can be performed by the learning means,
Claims 1 to 8, wherein the learning means executes learning of the function related to the copy failure determination and the function related to the output of the copy failure reason when execution of learning is set to be enabled by the first setting unit. The information processing device according to any one of the above. When the determination means determines that the radiation image is a failure and a user input affirming the determination result that the radiation image is a failure is received, or the determination means determines that the radiation image is not a failure. The information according to any one of claims 1 to 9, further comprising a re-photographing unit that starts processing for re-photographing when a user input denying the determination result that the image is not a photo failure is received. Processing equipment. Further comprising a second setting means for setting whether the automatic re-shooting function is enabled or disabled,
When the automatic re-imaging function is set to be valid, the re-imaging means starts processing for the re-imaging in response to the determining means determining that the radiation image is a failure. The information processing device according to claim 10. It is possible to communicate with an external device having a failure learning unit that receives a failed image and a reason for failure from the outside and learns a function related to failure determination and a function for outputting a failure reason.
The learning means acquires the result learned by the copy failure learning unit of the external device from the external device via the communication, and outputs the function related to the copy failure of the information processing device and the reason for the copy failure. The information processing device according to any one of claims 1 to 11, wherein the information processing device retains the information as a result of learning used by the function. When a radiation image is designated as a failure, the reason for failure is determined using the input reason for failure and at least one machine learning algorithm, using the radiation image designated as failure and the imaging conditions as input. a learning means for learning a function that outputs
means for outputting a reason for failure using the newly taken radiation image as an input, using the learning result by the learning means when a newly taken radiation image is designated as failure; Information processing device. an imaging means that generates a radiation image by generating an electrical signal according to the radiation dose;
determining means for determining whether or not the radiographic image is a failure;
In response to a user input affirming or denying the judgment result of the judgment means that the image is a failure, or a user input denying the judgment result of the judgment means that the image is a failure, the judgment means receives the radiation image as an input. a learning means for learning a function related to defect judgment for determining whether or not the photo is defective,
The learning means is configured to determine that the radiation image is a failure when the determination means receives a user input that affirms the determination result that the radiation image is a failure, or when the determination means determines that the radiation image is a failure. When it is determined that the image is not a defective image and a user input that negates the judgment result that the image is not a defective image is received, the reason for the image failure is output using the radiographic image and the imaging conditions as input, using the user input reason for the defective image. A radiography system that learns functions. an imaging means that generates a radiation image by generating an electrical signal according to the radiation dose;
When the radiation image is designated as a failure, output the reason for failure using the input and at least one machine learning algorithm with the radiation image designated as failure and imaging conditions as input. A learning means for learning functions;
means for outputting a reason for failure using the newly taken radiation image as an input, using the learning result by the learning means when a newly taken radiation image is designated as failure; Radiography system. In response to user input that affirms or denies the determination result that the radiation image is a photographic defect obtained by determining whether or not the radiation image is a photographic defect, or in response to a user input that negates the determination result that the radiation image is not a photographic defect obtained by the determination. , a step of learning a function of determining whether or not the radiation image is a defective image by inputting the radiation image;
When the radiographic image is determined to be a defective image according to the determination and a user input affirming the determination result that the radiation image is a defective image is received, or when it is determined that the radiographic image is not a defective image according to the determination, and If a user input denying the determination result that the image is not a defect is received, a step of learning a function of outputting the reason for the image failure using the radiographic image and the imaging conditions as input, using the user input reason for the image failure; An information processing method, including . When a radiation image is designated as a failure, the reason for failure is determined using the input reason for failure and at least one machine learning algorithm, using the radiation image designated as failure and the imaging conditions as input. a process of learning a function that outputs
Information processing, including a step of outputting a reason for failure using the learned result and inputting the newly acquired radiation image when a newly acquired radiation image is designated as failure. Method. A program that causes a computer to execute the information processing method according to claim 16 or 17.

Images