Information processing device, method and program

Abstract

To enable a feature point to be appropriately detected in setting a photographing range in an information processing device, method and program.SOLUTION: A processor acquires at least one of a first camera image generated by photographing a subject on a bed with a moving image by a first camera and a second camera image generated by photographing the subject with a moving image by a second camera having a photographing sensitivity higher than the first camera does, selects at least one detection model from a plurality of detection models including a first detection model configured to detect a plurality of feature points on the subject included in the first camera image and a second detection model configured to detect a plurality of feature points on the subject included in the second camera image, detects the plurality of feature points on the subject included in the first camera image or the second camera image by the selected detection model, and specifies a photographing range of the subject on the basis of the plurality of feature points.SELECTED DRAWING: Figure 3

Description

Technical Field

[0001] The present disclosure relates to an information processing apparatus, method, and program.

Background Art

[0002] In recent years, with the progress of medical devices such as CT (Computed Tomography) devices and MRI (Magnetic Resonance Imaging) devices, higher-quality high-resolution three-dimensional images have been increasingly used for image diagnosis.

[0003] When performing imaging of a subject using an imaging device such as a CT device or an MRI device, prior to the main imaging for acquiring a three-dimensional image to determine the imaging range, a scout imaging is performed, and a two-dimensional positioning image (scout image) is acquired. An operator (such as a technician) of the imaging device sets the imaging range at the time of the main imaging while viewing the scout image.

[0004] Prior to the scout imaging, the operator sets the imaging range of the scout imaging for the subject on the bed. For example, the subject is irradiated with a cross-shaped laser, the scan start position of the scout imaging is set, and the scan end position of the scout imaging is set so as to be an imaging range corresponding to the imaging site. At the time of the scout imaging, when the bed is moved from the initial position of the bed to the scan start position, the scan of the scout imaging is started, and when the bed is moved to the scan end position, the scan of the scout imaging is ended. After setting the imaging range of the main imaging based on the scout image acquired by the scout imaging, the operator performs the main imaging to acquire a three-dimensional image.

[0005] Here, when setting the imaging range at the time of the scout imaging, the subject is imaged by a camera provided above the bed, feature points such as both ankles, both waists, both elbows, and both shoulders of the subject are detected, and the imaging range is specified based on the detected feature points. At this time, a learned detection model constructed by machine learning a neural network is used for the detection of the feature points.

[0006] On the other hand, when applying a learned model to medical images, a method has been proposed to efficiently process medical images by preparing a learned model that performs a plurality of different processes and selecting which learned model to use according to the situation (see, for example, Patent Document 1).

Prior Art Documents

Patent Documents

[0007]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0008] By the way, although the imaging device is installed in the examination room, if the brightness of the examination room is insufficient, the noise in the image acquired by the camera increases, so that the feature points cannot be detected accurately. If the feature points cannot be detected accurately, the imaging range cannot be specified accurately.

[0009] The present disclosure has been made in view of the above circumstances, and an object thereof is to be able to appropriately detect feature points according to the brightness of the examination room when setting the imaging range.

Means for Solving the Problems

[0010] The radiation image processing apparatus according to the present disclosure includes at least one processor, The processor acquires at least one of a first camera image generated by video-recording a subject on a bed with a first camera and a second camera image generated by video-recording the subject with a second camera having a higher imaging sensitivity than the first camera, A plurality of detection models including a first detection model constructed to detect a plurality of feature points on a subject included in a first camera image and a second detection model constructed to detect a plurality of feature points on a subject included in a second camera image, select at least one detection model from, Using the selected detection model, detect a plurality of feature points on the subject included in the first camera image or the second camera image, Specify the shooting range of the subject based on the plurality of feature points.

[0011] "High shooting sensitivity" means high shooting performance in a dark place. Therefore, as the second camera, for example, an NIR camera, a night vision camera, or a camera with a higher ISO sensitivity than the first camera is used.

[0012] Note that in the information processing apparatus according to the present disclosure, when the brightness of the environment where the bed is installed is equal to or higher than a reference, the processor selects the first detection model and detects a plurality of feature points on the subject included in the first camera image. When the brightness is less than the reference, the second detection model may be selected to detect a plurality of feature points on the subject included in the second camera image.

[0013] Further, in the information processing apparatus according to the present disclosure, the processor may acquire the first camera image and determine the brightness based on the luminance information derived from the first camera image.

[0014] Further, in the information processing apparatus according to the present disclosure, the processor may acquire the first camera image and determine the brightness based on the noise included in the first camera image.

[0015] Further, in the information processing apparatus according to the present disclosure, the processor may acquire the first camera image, detect feature points from the first camera image using the first detection model, and determine the brightness based on the detection accuracy of the feature points.

[0016] Also, in the information processing apparatus according to the present disclosure, the processor may determine the brightness by a sensor that detects the brightness of the environment.

[0017] Also, in the information processing apparatus according to the present disclosure, the processor acquires a first camera image, selects a first detection model to detect feature points from the first camera image, determines whether the brightness is equal to or higher than a reference based on the first camera image, and if the brightness is equal to or higher than the reference, specifies the shooting range based on the feature points detected using the first detection model, and if the brightness is less than the reference, selects a second detection model and specifies the shooting range based on the feature points detected using the second detection model.

[0018] Also, in the information processing apparatus according to the present disclosure, the first detection model and the second detection model are models that prioritize the frame rate at the time of detecting feature points. The plurality of detection models further includes a third detection model that emphasizes the accuracy at the time of detecting feature points, which is constructed to detect a plurality of feature points on the subject included in the first camera image, and a fourth detection model that emphasizes the accuracy at the time of detecting feature points, which is constructed to detect a plurality of feature points on the subject included in the second camera image. The processor may select a detection model according to the brightness of the environment where the bed is installed and the imaging part of the subject.

[0019] Also, in the information processing apparatus according to the present disclosure, when the brightness is equal to or higher than the reference, the processor selects either the first detection model or the third detection model to detect a plurality of feature points on the subject included in the first camera image, and when the brightness is less than the reference, the processor selects either the second detection model or the fourth detection model to detect a plurality of feature points on the subject included in the second camera image.

[0020] Further, in the information processing apparatus according to the present disclosure, the processor may select either the first detection model or the third detection model, and either the second detection model or the fourth detection model according to the imaging region of the subject.

[0021] Further, in the information processing apparatus according to the present disclosure, the processor may determine the detection accuracy of the feature points, specify the imaging range based on the feature points when the detection accuracy is equal to or higher than a reference, and issue a warning when the detection accuracy is lower than the reference.

[0022] Further, in the information processing apparatus according to the present disclosure, the first detection model is a model that emphasizes the frame rate at the time of detecting feature points. The plurality of detection models further includes a third detection model that emphasizes the accuracy at the time of detecting feature points, which is constructed to detect a plurality of feature points on the subject included in the first camera image. The processor acquires the first camera image and the second camera image, selects the first detection model, and detects feature points from the first camera image. Determine the detection accuracy of the feature points detected using the first detection model. When the detection accuracy is equal to or higher than a first reference, specify the imaging range based on the feature points detected using the first detection model. When the detection accuracy is lower than the first reference, select the third detection model, detect feature points from the first camera image using the third detection model. Determine the detection accuracy of the feature points detected using the third detection model. When the detection accuracy is equal to or higher than a second reference, specify the imaging range based on the feature points detected using the third detection model. When the detection accuracy is lower than the second reference, select the second detection model, detect feature points from the second camera image using the second detection model. It may be configured to specify the imaging range based on the feature points detected using the second detection model.

[0023] Also, in the information processing apparatus according to the present disclosure, when the detection accuracy is equal to or higher than a first criterion, the processor determines the detection accuracy of feature points detected using a first detection model; when the detection accuracy is equal to or higher than a second criterion, the processor determines the detection accuracy of feature points detected using a third detection model; or the processor determines the detection accuracy of feature points detected using a second detection model. When the detection accuracy is equal to or higher than a third criterion, the imaging range is specified based on the feature points detected using the first detection model when the detection accuracy is equal to or higher than the first criterion, the feature points detected using the third detection model when the detection accuracy is equal to or higher than the second criterion, or the feature points detected using the second detection model. When the detection accuracy is lower than the third criterion, a warning may be issued.

[0024] Also, in the information processing apparatus according to the present disclosure, the first detection model is a model that emphasizes the frame rate at the time of detecting feature points. The plurality of detection models further includes a third detection model that is constructed to detect a plurality of feature points on a subject included in the first camera image and emphasizes the accuracy at the time of detecting the feature points. The processor acquires a first camera image and a second camera image, selects the first detection model, and detects feature points from the first camera image. The processor detects the movement of the subject based on the first camera image. When the movement of the subject is equal to or greater than a first criterion, the imaging range is specified based on the feature points detected using the first detection model. When the movement of the subject is less than the first criterion, the second detection model and the third detection model are selected. Feature points are detected from the first camera image using the third detection model. Feature points are detected from the second camera image using the second detection model. The processor compares the detection accuracy of the feature points detected using the third detection model with the detection accuracy of the feature points detected using the second detection model. When the detection accuracy of the feature points detected using the third detection model is higher, the imaging range is specified based on the feature points detected using the third detection model. When the detection accuracy of the feature points detected using the second detection model is higher, the shooting range may be specified based on the feature points detected using the second detection model.

[0025] Further, in the information processing apparatus according to the present disclosure, the processor determines the detection accuracy of the feature points detected using the first detection model, the detection accuracy of the feature points detected using the third detection model, or the detection accuracy of the feature points detected using the second detection model when the movement is equal to or greater than the first reference, When the detection accuracy is equal to or greater than the second reference, the shooting range is specified based on the feature points detected using the first detection model, the feature points detected using the third detection model, or the feature points detected using the second detection model when the movement is equal to or greater than the first reference, When the detection accuracy is lower than the second reference, a warning may be issued.

[0026] Further, in the information processing apparatus according to the present disclosure, the processor may derive the movement range of the bed based on the shooting range.

[0027] Further, in the information processing apparatus according to the present disclosure, the processor may display a human body image simulating a human body on the display and draw a movement start line and a movement end line of the bed based on the movement range of the bed on the human body image.

[0028] Further, in the information processing apparatus according to the present disclosure, the shooting range may be a shooting range when shooting a positioning image acquired before actually shooting the subject.

[0029] The information processing method according to the present disclosure includes a computer acquiring at least one of a first camera image generated by video-shooting a subject on a bed with a first camera and a second camera image generated by video-shooting the subject with a second camera having a higher shooting sensitivity than the first camera, A plurality of detection models including a first detection model constructed to detect a plurality of feature points on a subject included in a first camera image and a second detection model constructed to detect a plurality of feature points on a subject included in a second camera image, select at least one detection model from, Using the selected detection model, detect a plurality of feature points on the subject included in the first camera image or the second camera image, Specify the shooting range of the subject based on the plurality of feature points.

[0030] The information processing program according to the present disclosure includes a procedure for acquiring at least one of a first camera image generated by video-recording a subject on a bed with a first camera and a second camera image generated by video-recording the subject with a second camera having a higher shooting sensitivity than the first camera, and A procedure for selecting at least one detection model from a plurality of detection models including a first detection model constructed to detect a plurality of feature points on a subject included in the first camera image and a second detection model constructed to detect a plurality of feature points on a subject included in the second camera image, A procedure for using the selected detection model to detect a plurality of feature points on the subject included in the first camera image or the second camera image, A procedure for specifying the shooting range of the subject based on the plurality of feature points is executed by a computer.

Effect of the Invention

[0031] According to the present disclosure, when setting the shooting range, feature points can be appropriately detected according to the brightness of the examination room.

Brief Description of the Drawings

[0032]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Mode for Carrying Out the Invention

[0033] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings. FIG. 1 is a perspective view showing an overview of a CT apparatus to which the information processing apparatus according to the first embodiment of the present disclosure is applied, and FIG. 2 is a view of the CT apparatus to which the information processing apparatus according to the first embodiment of the present disclosure is applied as seen from the side. As shown in FIGS. 1 and 2, the CT apparatus 1 according to the present embodiment includes a gantry 2, a couch 3, and a console 4.

[0034] The gantry 2 has a tunnel-shaped structure with an opening 5 at its center. Inside the gantry 2, an X-ray source unit that emits X-rays and a detection unit that detects X-rays and generates a radiation image are provided (both are not shown). The X-ray source unit and the detection unit can each rotate along the annular shape of the gantry 2 while maintaining a position relationship facing each other. Also, inside the gantry 2, a control unit that controls the operation of the CT apparatus 1 is provided.

[0035] The examination table 3 has a table portion 3A on which the subject lies, a base portion 3B that supports the table portion 3A, and a drive unit 3C that reciprocates the table portion 3A in the direction of arrow A. The table portion 3A is slidable in the direction of arrow A with respect to the base portion 3B by the drive unit 3C. When taking a CT image, by sliding the table portion 3A, the subject H lying on the table portion 3A is conveyed into the opening 5 of the gantry 2.

[0036] Note that a camera 7 is installed above the examination table 3. The camera 7 is configured by integrating an RGB camera capable of taking an RGB color image by detecting the reflected light of the subject H and an NIR (Near InfraRed) capable of stereo photography. FIG. 3 is a schematic perspective view showing the appearance of the camera 7. As shown in FIG. 3, the camera 7 is configured by attaching an RGB camera 31 and an NIR camera 35 to a base 33. The RGB camera 31 has an RGB sensor 32 including an imaging element such as a lens and a CCD (Charge Cupuled Device). The RGB camera 31 takes a moving image of the subject H on the examination table 3 at a predetermined frame rate to obtain an RGB camera image that is an RGB color moving image and outputs it to the console 4.

[0037] The NIR camera 35 includes a left NIR sensor 36 and a right NIR sensor 37 that contain imaging elements such as a lens and a CCD, as well as an NIR projector 38. The NIR camera 35 irradiates near-infrared rays from the NIR projector 38 toward the subject H, and detects the reflected near-infrared light of the subject H by the left and right NIR sensors 36 and 37 at a predetermined frame rate. Thereby, the NIR camera 35 acquires left and right NIR camera images and outputs them to the console 4. Note that the left and right NIR camera images are monochrome images. Here, since the left NIR sensor 36 and the right NIR sensor 37 are separated, the left and right NIR camera images have parallax. Therefore, the camera 7 derives the depth information of the subject H and other subjects included in the left and right NIR camera images based on the parallax, and outputs the depth information together with the left and right NIR camera images. In the following description, when simply referring to the NIR camera image, it refers to either the left or right NIR camera image.

[0038] Since the NIR camera 35 performs imaging based on near-infrared rays, even if the imaging room where the CT apparatus 1 is installed is dark, it is possible to acquire an NIR camera image in which the subject H can be visually recognized. On the other hand, since the RGB camera 31 performs imaging based on visible light, if the ambient brightness is insufficient, it becomes difficult to visually recognize the subject H in the acquired RGB camera image. Therefore, in order to acquire an RGB camera image in which the subject H can be visually recognized, the imaging room needs to be somewhat bright. The RGB camera 31 is an example of the first camera of the present disclosure. The NIR camera 35 is an example of the second camera of the present disclosure, which has a higher imaging sensitivity than the first camera. The RGB camera image is an example of the first camera image of the present disclosure, and the NIR camera image is an example of the second camera image of the present disclosure.

[0039] Note that in the present embodiment, the camera 7 is assumed to simultaneously acquire an RGB camera image by the RGB camera 31 and an NIR camera image by the NIR camera 35.

[0040] The driving of the gantry 2, the driving of the examination table 3, and the photographing of the subject H by the camera 7 are performed based on the operator's input from the console 4. The console 4 incorporates the information processing apparatus according to the first embodiment.

[0041] Next, the information processing apparatus according to the first embodiment incorporated in the console 4 will be described. First, with reference to FIG. 4, the hardware configuration of the information processing apparatus according to the first embodiment will be described. As shown in FIG. 4, the information processing apparatus 10 is a computer such as a workstation, a server computer, and a personal computer, and includes a CPU (Central Processing Unit) 11, a nonvolatile storage 13, and a memory 16 as a temporary storage area. Further, the information processing apparatus 10 includes a display 14 such as a liquid crystal display, an input device 15 such as a keyboard and a mouse, and an interface 17 such as a network I / F (InterFace) connected to the CT apparatus 1. The CPU 11, the storage 13, the display 14, the input device 15, the memory 16, and the network I / F 17 are connected to a bus 18. Note that the CPU 11 is an example of the processor in the present disclosure. The display 14 and the input device 15 are also illustrated in FIGS. 1 and 2.

[0042] The storage 13 is realized by an HDD (Hard Disk Drive), an SSD (Solid State Drive), a flash memory, or the like. The information processing program 12 installed in the information processing apparatus 10 is stored in the storage 13 as a storage medium. The CPU 11 reads the information processing program 12 from the storage 13, expands it in the memory 16, and executes the expanded information processing program 12.

[0043] The information processing program 12 is stored in a storage device of a server computer connected to a network or in a network storage in a state accessible from the outside, and is downloaded and installed on the computer constituting the information processing apparatus 10 in response to a request. Alternatively, it is recorded on a recording medium such as a DVD (Digital Versatile Disc) or a CD-ROM (Compact Disc Read Only Memory) and distributed, and is installed on the computer constituting the information processing apparatus 10 from the recording medium.

[0044] Next, the functional configuration of the information processing apparatus according to the first embodiment will be described. FIG. 5 is a diagram showing the functional configuration of the information processing apparatus according to the first embodiment. As shown in FIG. 5, the information processing apparatus 10 includes a photographing control unit 20, a camera control unit 21, a selection unit 22, a feature point detection unit 23, and a photographing range specifying unit 24. Then, by executing the information processing program 12, the CPU 11 functions as the photographing control unit 20, the camera control unit 21, the selection unit 22, the feature point detection unit 23, and the photographing range specifying unit 24.

[0045] The photographing control unit 20 controls the photographing unit, the detection unit, and the control unit provided in the gantry 2 to photograph the subject H in accordance with an instruction from the input device 15. Note that, at the time of CT photographing, in order to determine the photographing range, a scout photographing is performed prior to the main photographing for acquiring a three-dimensional CT image. The scout photographing is performed by photographing the subject H with the photographing unit and the detection unit fixed.

[0046] At the time of scout imaging, as will be described later, the imaging range of the subject H is set, and the bed 3 is moved to the opening 5 of the gantry 2 so that the set imaging range is imaged, and the scout imaging is performed. The scout image obtained by the scout imaging is a two-dimensional X-ray image including the imaging range set for the subject H. The scout image is displayed on the display 14. The operator views the scout image displayed on the display 14 and sets the imaging range for performing the main imaging. After setting the imaging range, when the operator gives an instruction for the main imaging from the input device 15, the main imaging is performed, and a three-dimensional CT image of the subject H is obtained. The obtained scout image and CT image are stored in the storage 13.

[0047] The camera control unit 21 controls the imaging of the subject H on the bed 3 by the camera 7. The imaging of the subject H by the camera 7 is performed to set the imaging range when performing the scout imaging. The imaging of the subject H by the camera 7 is performed from the pre-imaging preparation stage. That is, the camera control unit 21 starts the imaging by the camera 7 from the time before the subject H lies supine on the bed 3, and causes the camera 7 to acquire a camera image. Then, when an instruction to start the scout imaging is given by the operator, the camera control unit 21 stops the imaging by the camera 7. The acquired camera image is stored in the memory 16 for specifying the imaging range described later. In the following description, the RGB camera image and the NIR camera image may sometimes be simply referred to as the camera image.

[0048] In the first embodiment, the selection unit 22 selects one detection model from a first detection model 22A constructed to detect a plurality of feature points on the subject H included in the RGB camera image acquired by the camera 7 and a second detection model 22B constructed to detect a plurality of feature points on the subject H included in the NIR camera image. The first detection model 22A and the second detection model 22B are constructed by machine learning of a neural network.

[0049] Since the first detection model 22A detects feature points from an RGB camera image, a model constructed to detect feature points from a color image is used. Since the second detection model 22B detects feature points from an NIR camera image, a model constructed to detect feature points from a monochrome image is used.

[0050] For the training of the first detection model 22A, a color image including the whole body of a human and having 17 known feature points is used as a teacher image. Specifically, an RGB camera image acquired by the RGB camera 31 is used. For the training of the second detection model 22B, a monochrome image including the whole body of a human and having 17 known feature points is used as a teacher image. Actually, either the left or right NIR camera image acquired by the NIR camera 35 is used as a teacher image. Note that the teacher image is acquired, for example, by photographing a person wearing an inspection gown with the camera 7 in the same manner as in the actual inspection.

[0051] Note that neural networks having the same structure may be used for the first detection model 22A and the second detection model 22B, or neural networks having a structure suitable for detecting feature points from each of the RGB camera image and the NIR camera image may be used.

[0052] In the first embodiment, the selection unit 22 selects either the first detection model 22A or the second detection model 22B according to the brightness of the photographing room in which the bed 3 is installed. The determination of the brightness may be made, for example, based on the luminance information derived from the RGB camera image. In this case, the selection unit 22 may derive the luminance at each pixel of the RGB camera image, derive the average value of the luminances at all the pixels of the RGB camera image as the luminance information, determine that it is bright when the luminance information is equal to or greater than a threshold value, and determine that it is dark when the luminance information is less than the threshold value. Note that the luminance of each pixel can be derived by the following formula (1) based on the signal values of R, G, and B of each pixel. In formula (1), Y is the luminance. Y = 0.299×R + 0.587×G + 0.114×B (1)

[0053] On the other hand, when the brightness is insufficient, the RGB camera image will contain a lot of noise. Therefore, the selection unit 22 may derive the noise characteristics of the RGB camera image and determine the brightness based on the noise characteristics. As the noise characteristics, the standard deviation of the pixel values of each pixel in the RGB camera image can be used. The standard deviation may be calculated by the RMS (root mean square) of the pixel values of each image in the RGB camera image. In this case, the selection unit 22 calculates the average value N 2 m of the squares of the pixel values of each pixel in the RGB camera image, and the average value N 2 m is opened to the square root (that is, √N 2 m). The selection unit 22 may determine that it is dark if the standard deviation is equal to or greater than a predetermined threshold value, and determine that it is bright if the standard deviation is less than the threshold value. Note that, instead of the standard deviation, the variance of the pixel values of the RGB camera image or the difference between the maximum value and the minimum value may be calculated as the noise characteristics.

[0054] Alternatively, a brightness sensor for detecting the brightness near the hospital bed 3 in the CT apparatus 1 or the imaging room may be installed, and the brightness may be determined based on the output of the brightness sensor. In this case, the selection unit 22 may determine that it is bright if the detected value of the brightness sensor is equal to or greater than a predetermined threshold value, and determine that it is dark if the detected value of the brightness sensor is less than the threshold value.

[0055] The feature point detection unit 23 detects feature points from the camera image using the detection model selected by the selection unit 22. FIG. 6 is a diagram for explaining the feature points. As shown in FIG. 6, in the present embodiment, the first and second detection models 22A and 22B are constructed so as to detect 17 feature points on the subject H included in the camera image. The 17 feature points are both eyes, nose, both ears, both shoulders, both elbows, both hands, both waists, both knees, and both feet. Note that the midpoint of the collarbone may be added to these 17 feature points to use 18 feature points.

[0056] The first detection model 22A and the second detection model 22B derive the probability representing the possibility of each of the 17 feature points at each pixel of the camera image. Then, the feature point detection unit 23 detects, as a feature point, the pixel with the highest probability derived for each of the 17 feature points. For example, when detecting the right eye as a feature point, the feature point detection unit 23 compares the probabilities of the entire pixels of the camera image derived by the first detection model 22A and the second detection model 22B being the right eye, and detects the pixel with the largest probability as the feature point of the right eye.

[0057] The shooting range specifying unit 24 specifies the shooting range of the subject H when performing scout shooting based on the 17 feature points detected by the feature point detection unit 23. For this purpose, the shooting range specifying unit 24 first determines the detection accuracy of the feature points detected by the feature point detection unit 23. As described above, the feature point detection unit 23 detects, as a feature point, the pixel with the highest probability derived for each of the 17 feature points. The higher the probability output by the detection model for the detected feature point, the better the detection accuracy. Therefore, the shooting range specifying unit 24 compares the representative value of the probability derived by the detection model for the 17 feature points with a predetermined threshold value, and determines that the detection accuracy is good when the representative value is equal to or greater than the threshold value, and the detection accuracy is poor when the representative value is less than the threshold value. As the representative value, an average value, a median value, a weighted average value according to the shooting part, etc. can be used, but it is not limited to these.

[0058] When the shooting range specifying unit 24 determines that the detection accuracy is good, it performs a process of specifying the shooting range. On the other hand, when it determines that the detection accuracy is poor, the shooting range specifying unit 24 performs a warning display on the display 14. Here, if the subject H moves too much, a thick blanket is put on the subject H, or the entire body of the subject H is not included in the shooting range of the camera 7, the feature points cannot be accurately detected using any detection model, and the detection accuracy deteriorates. In such a case, the shooting range specifying unit 24 determines that the detection accuracy is poor.

[0059] When a warning is displayed, the operator manually sets the shooting range for scout shooting. That is, the operator measures the distance from the initial position of the bed to the scan start line, further measures the distance between the scan start line and the scan end line, and inputs the measured distances from the input device 15. Based on the input distances, the movement of the bed 3 during scout shooting is controlled.

[0060] Hereinafter, the process of specifying the shooting range by the shooting range specifying unit 24 will be described. For example, when the shooting part is the head, the scout image has a shooting range from the top of the head to the tip of the chin. Therefore, the shooting range specifying unit 24 sets a line connecting both eyes or both ears among the feature points detected by the feature point detecting unit 23, and further sets a scan start line at the top of the head and a scan end line between the chin and both shoulders. Then, based on the distance relationship between the line connecting both eyes or both ears and the scan start line and the scan end line, the distance D1 between the scan start line and the scan end line is derived. The range of the distance D1 between the scan start line and the scan end line is the shooting range.

[0061] Here, before scout shooting, the bed 3 is in the initial position and the subject H is lying supine on the bed 3, so the distance from the end of the bed 3 to the top of the subject H's head can be known from the camera image. Therefore, the shooting range specifying unit 24 calculates the distance D2 from the end of the bed 3 to the top of the subject H's head as the movement amount of the bed from the initial position to the scan start line, that is, the movement amount of the bed 3 until the top of the subject H's head reaches the scan position in the CT apparatus 1.

[0062] When the imaging range is specified, the imaging range specifying unit 24 displays the scan start line and the scan end line on the schema, which is a human body diagram displayed on the display 14. FIG. 7 is a diagram showing the schema on which the start line and the end line are displayed. As shown in FIG. 7, the schema 40 shows a scan start line 41 and a scan end line 42. The operator checks the scan start line and the scan end line displayed on the display 14. After the check, if it is okay, the operator uses the input device 15 to give an instruction to start the scout imaging.

[0063] Upon receiving the instruction to start the scout imaging, information on the distances D1 and D2 is output to the CT apparatus 1. The imaging control unit 20 controls the CT apparatus 1 such that after the driving unit 3C moves the bed 3 by the distance D2, the scan of the scout imaging is started, and when the driving unit 3C moves the bed 3 by the distance D1, the scan of the scout imaging is ended.

[0064] The scout image obtained by the scout imaging is displayed on the display 14. The operator checks the scout image displayed on the display 14 and sets the imaging range of the main imaging with respect to the scout image. Then, the main imaging is performed by giving an imaging instruction of the main imaging using the input device 15, and a three-dimensional CT image of the imaging region of the subject H is obtained within the set imaging range of the main imaging.

[0065] Next, the processing performed in the first embodiment will be described. FIG. 8 is a flowchart showing the processing performed in the first embodiment. The processing is started when an instruction to start imaging is given from the input device 15, and the selection unit 22 determines the brightness of the imaging room in which the CT apparatus 1 is installed (step ST1). Then, the selection unit 22 selects either the first detection model 22A or the second detection model 22B according to the determined brightness (detection model selection; step ST2).

[0066] Subsequently, the camera control unit 21 starts shooting by the camera 7 to acquire a camera image (step ST3), and the feature point detection unit 23 detects feature points from the camera image using the detection model selected by the selection unit 22 (step ST4). Subsequently, the shooting range specifying unit 24 determines the detection accuracy of the feature points (step ST5). If it is determined that the detection accuracy is good, the shooting range specifying unit 24 specifies the shooting range during scouting shooting based on the feature points (step ST6), and based on the specified shooting range, draws the scan start line and the scan end line during scouting shooting on the schema displayed on the display 14 (line drawing; step ST7). On the other hand, if it is determined that the detection accuracy is bad, the shooting range specifying unit 24 performs a warning display (step ST8). When the warning display is performed, the information processing apparatus 10 ends the shooting range specifying process. In this case, as described above, the operator manually sets the shooting range of the scouting shooting.

[0067] Subsequently, it is determined whether or not an instruction to start scouting shooting has been given by the operator (step ST9). If step ST9 is negated, the process returns to step ST3, and the processes after step ST3 are repeated. If step ST9 is affirmed, the camera control unit 21 stops shooting by the camera 7 (step ST10), and the information processing apparatus 10 ends the shooting range specifying process.

[0068] Thereafter, scouting shooting is performed by the shooting control unit 20 to acquire a scouting image, which is displayed on the display 14. After the operator checks the scouting image, the operator sets the shooting range of the main shooting. Then, when the operator gives an instruction for the main shooting from the input device 15, the main shooting is performed, and a three-dimensional CT image of the subject H is acquired.

[0069] As described above, in the first embodiment, the detection model used for detecting feature points is selected according to the brightness of the inspection room. Therefore, when the inspection room is bright, the first detection model 22A that detects feature points from the RGB camera image can be selected, and when the inspection room is dark, the second detection model 22B that detects feature points from the NIR camera image can be selected. Thus, when setting the shooting range, feature points can be appropriately detected according to the brightness of the inspection room, and as a result, the shooting range at the time of scouting shooting can be appropriately specified using the detected feature points.

[0070] In addition, in the above first embodiment, both the RGB camera image and the NIR camera image are acquired by the camera 7, but the present invention is not limited to this. When the selection unit 22 determines that it is bright, only the RGB camera image may be acquired by the RGB camera 31 of the camera 7, and when the selection unit 22 determines that it is dark, only the NIR camera image may be acquired by the NIR camera 35 of the camera 7.

[0071] Next, a second embodiment of the present disclosure will be described. Since the hardware configuration and functional configuration of the information processing apparatus according to the second embodiment are the same as those of the first embodiment described above, detailed description thereof will be omitted here. The information processing apparatus 10 according to the second embodiment first selects the first detection model 22A to detect feature points from the RGB camera image, determines whether the brightness is equal to or higher than a reference based on the RGB camera image, and when the brightness is equal to or higher than the reference, the shooting range is specified based on the feature points detected from the RGB camera image using the first detection model 22A. When the brightness is lower than the reference, the second detection model 22B is selected, and the shooting range is specified based on the feature points detected from the NIR camera image using the second detection model 22B. This is different from the first embodiment.

[0072] Next, the processing performed in the second embodiment will be described. FIG. 9 is a flowchart showing the processing performed in the second embodiment. For example, when an instruction to start shooting is given from the input device 15, the processing is started, and the camera control unit 21 starts shooting by the camera 7 to acquire an RGB camera image (step ST21). Next, the selection unit 22 selects the first detection model 22A (step ST22), and the feature point detection unit 23 detects feature points from the RGB camera image using the first detection model 22A selected by the selection unit 22 (step ST23).

[0073] Subsequently, the selection unit 22 derives luminance information from the RGB camera image and determines the brightness of the examination room where the CT apparatus 1 is installed based on the luminance information (step ST24). If it is determined that the brightness is low, the selection unit 22 selects the second detection model 22B instead of the first detection model 22A (step ST25). Subsequently, the feature point detection unit 23 detects feature points from the NIR camera image using the second detection model 22B (step ST26), and the shooting range specifying unit 24 determines the detection accuracy of the feature points (step ST27). If it is determined in step ST24 that the brightness is high, the process proceeds to step ST27, and the shooting range specifying unit 24 determines the detection accuracy of the feature points detected from the RGB camera image in step ST23.

[0074] If it is determined that the detection accuracy is good, the shooting range specifying unit 24 specifies the shooting range at the time of scout shooting based on the feature points (step ST28), and based on the specified shooting range, draws the scan start line and the scan end line at the time of scout shooting on the schema displayed on the display 14 (line drawing; step ST29). On the other hand, if it is determined that the detection accuracy is poor, the shooting range specifying unit 24 performs a warning display (step ST30). When the warning display is performed, the information processing apparatus 10 ends the shooting range specifying process. In this case, as described above, the operator manually sets the shooting range of the scout shooting.

[0075] Subsequently, it is determined whether or not an operator has given an instruction to start scouting photography (step ST31). If step ST31 is negative, the process returns to step ST21, and the processes after step ST21 are repeated. If step ST31 is affirmative, the camera control unit 21 stops the photography by the camera 7 (step ST32), and the information processing apparatus 10 ends the photographing range specifying process.

[0076] As described above, in the second embodiment, the feature points are detected from the RGB camera image using the first detection model 22A, the brightness is determined using the RGB camera image, the detected feature points are used as they are when it is bright, and the second detection model 22B is used to detect the feature points using the NIR camera image when it is dark. Therefore, when setting the photographing range, the feature points can be appropriately detected according to the brightness of the photographing room. As a result, the photographing range at the time of scouting photography can be appropriately specified using the detected feature points.

[0077] In the second embodiment described above, the brightness is determined based on the luminance information derived from the RGB camera image, but it is not limited thereto. The feature points may be detected from the RGB camera image using the first detection model 22A, and the brightness may be determined based on the detection accuracy of the feature points. Here, when the inspection room is dark, it becomes difficult for the subject H to be visually recognized in the RGB camera image, so the detection accuracy of the feature points decreases. Therefore, the selection unit 22 compares the representative value of the probability derived by the detection model for 17 feature points with a predetermined threshold value. When the representative value is equal to or greater than the threshold value, the detection accuracy is good, so the inspection room is bright. When the representative value is less than the threshold value, the detection accuracy is poor, so the inspection room may be determined to be dark.

[0078] Next, a third embodiment of the present disclosure will be described. Since the hardware configuration of the information processing apparatus according to the third embodiment is the same as that of the information processing apparatus according to the first embodiment described above, detailed description thereof will be omitted here. FIG. 10 is a diagram showing a functional configuration of the information processing apparatus according to the third embodiment. In FIG. 10, the same components as those in FIG. 5 are given the same reference numerals, and detailed description thereof will be omitted. The information processing apparatus 10A according to the third embodiment has a third detection model 22C and a fourth detection model 22D in addition to the first detection model 22A and the second detection model 22B, and is different from the first embodiment in that a detection model used for detecting feature points is selected from the first detection model 22A, the second detection model 22B, the third detection model 22C, and the fourth detection model 22D.

[0079] In the third embodiment, the first detection model 22A is a model that emphasizes frame rate and detects feature points from RGB camera images, and the second detection model 22B is a model that emphasizes frame rate and detects feature points from NIR camera images. The third detection model 22C is a model that emphasizes accuracy and detects feature points from RGB camera images, and the fourth detection model 22D is a model that emphasizes accuracy and detects feature points from NIR camera images.

[0080] "Emphasizing the frame rate" means, for example, reducing the amount of data to be processed or omitting calculations to improve the processing speed for detecting feature points. Since the first detection model 22A and the second detection model 22B emphasize the frame rate, models with a small amount of calculation and a high processing speed for detecting feature points are used.

[0081] "Emphasizing accuracy" means improving the accuracy of detecting feature points of operations that require time without reducing the amount of data or omitting operations. Since the third detection model 22C and the fourth detection model 22D emphasize accuracy, they have a large amount of computation and are slower in processing speed than the first detection model 22A and the second detection model 22B. However, models with higher accuracy for detecting feature points than the first detection model 22A and the second detection model 22B are used.

[0082] For the first detection model 22A and the second detection model 22B, a neural network with a structure that has a small amount of computation and can perform the process of detecting feature points at high speed is used. For the third detection model 22C and the fourth detection model 22D, although they have a large amount of computation, a neural network with a structure that can detect feature points more accurately than the first detection model 22A and the second detection model 22B is used. For all neural networks, a teacher image that includes the whole body of a human body and has 17 known feature points, which is obtained by photographing the human body with the camera 7, is used for learning. Note that the teacher image is obtained, for example, by photographing a person wearing an inspection suit with the camera 7 in the same way as in the actual inspection. In the third embodiment, RGB camera images are used as teacher images for the learning of the first detection model 22A and the third detection model 22C. Also, NIR camera images, which are gray images, are used as teacher images for the learning of the second detection model 22B and the fourth detection model 22D.

[0083] Note that the same neural network structure may be used for the first detection model 22A and the second detection model 22B, and the third detection model 22C and the fourth detection model 22D. In this case, the neural network is trained using different teacher images for the first detection model 22A and the second detection model 22B, and the third detection model 22C and the fourth detection model 22D. For example, the first teacher image used for training the first detection model 22A and the second detection model 22B has a relatively low resolution. On the other hand, the second teacher image used for training the third detection model 22C and the fourth detection model 22D has a higher resolution compared to the first teacher image.

[0084] In the third embodiment, the selection unit 22 selects a detection model according to the brightness of the examination room and the imaging site of the subject H. Here, the imaging sites of the subject H include the head, chest, abdomen, lower limbs, and the whole body. The head is likely to move during imaging, while the chest or abdomen is less likely to move during imaging. Therefore, when the examination room is bright and the imaging site is the head or the whole body including the head, the selection unit 22 selects the first detection model 22A that emphasizes the frame rate for detecting feature points from the RGB camera image. On the other hand, when the examination room is dark and the imaging site is the head or the whole body including the head, the selection unit 22 selects the second detection model 22B that emphasizes the frame rate for detecting feature points from the NIR camera image.

[0085] Also, when the imaging site is the abdomen or the lower limbs, imaging is often performed with a blanket covering. In such a case, when the examination room is bright, the selection unit 22 selects the third detection model 22C that emphasizes accuracy for detecting feature points from the RGB camera image. When the examination room is dark, the selection unit 22 selects the fourth detection model 22D that emphasizes accuracy for detecting feature points from the NIR camera image.

[0086] Note that which part of the subject H is to be imaged is included in the examination order provided by the doctor during imaging, and the operator sets it from the input device 15 according to the examination order.

[0087] Next, the processing performed in the third embodiment will be described. FIG. 11 is a flowchart showing the processing performed in the third embodiment. For example, when an instruction to start shooting is given from the input device 15, the processing is started, and the selection unit 22 determines the brightness of the shooting room where the CT apparatus 1 is installed and the shooting site included in the inspection order (step ST41). Then, the selection unit 22 selects one of the first detection model 22A to the fourth detection model 22D according to the determined brightness (detection model selection; step ST42).

[0088] When the selection unit 22 determines that it is bright and the shooting site is a site where movement is easy, the selection unit 22 selects the first detection model 22A. When it determines that it is dark and the shooting site is a site where movement is easy, the selection unit 22 selects the second detection model 22B. Further, when the selection unit 22 determines that it is bright and the shooting site is a site where movement is difficult, the selection unit 22 selects the third detection model 22C. When it determines that it is dark and the shooting site is a site where movement is difficult, the selection unit 22 selects the fourth detection model 22D.

[0089] Subsequently, the camera control unit 21 starts shooting by the camera 7 and acquires a camera image (step ST43), and the feature point detection unit 23 detects feature points from the camera image using the detection model selected by the selection unit 22 (step ST44).

[0090] That is, when the first detection model 22A is selected, the feature point detection unit 23 detects feature points from the RGB camera image using the first detection model 22A. When the second detection model 22B is selected, the feature point detection unit 23 detects feature points from the NIR camera image using the second detection model 22B. Further, when the third detection model 22C is selected, the feature point detection unit 23 detects feature points from the RGB camera image using the third detection model 22C. When the fourth detection model 22D is selected, the feature point detection unit 23 detects feature points from the NIR camera image using the fourth detection model 22D.

[0091] Subsequently, the shooting range specifying unit 24 determines the detection accuracy of the feature points (step ST45). If it is determined that the detection accuracy is good, the shooting range specifying unit 24 specifies the shooting range during scout shooting based on the feature points (step ST46), and based on the specified shooting range, draws the scan start line and the scan end line during scout shooting on the schema displayed on the display 14 (line drawing; step ST47). On the other hand, if it is determined that the detection accuracy is bad, the shooting range specifying unit 24 performs a warning display (step ST48).

[0092] Subsequently, it is determined whether or not an instruction to start scout shooting has been given by the operator (step ST49). If step ST49 is negated, the process returns to step ST43, and the processes after step ST43 are repeated. If step ST49 is affirmed, the camera control unit 21 stops the shooting by the camera 7 (step ST50), and the information processing apparatus 10 ends the shooting range specifying process.

[0093] After this, scout shooting is performed by the shooting control unit 20 to obtain a scout image, which is displayed on the display 14. After the operator checks the scout image, the shooting range for the main shooting is set. Then, when the operator gives an instruction for the main shooting from the input device 15, the main shooting is performed, and a three-dimensional CT image of the subject H is obtained.

[0094] As described above, in the third embodiment, in addition to the brightness of the examination room, the detection model used for detecting the feature points is selected according to the imaging region. For this reason, the feature points can be detected using an appropriate detection model according to the brightness of the examination room and the imaging region. Therefore, the shooting range during scout shooting can be appropriately specified using the detected feature points.

[0095] In the above-described third embodiment, both the RGB camera image and the NIR camera image are acquired by the camera 7, but the present invention is not limited thereto. When the selection unit 22 determines that it is bright, only the RGB camera image may be acquired by the RGB camera 31 of the camera 7, and when the selection unit 22 determines that it is dark, only the NIR camera image may be acquired by the NIR camera 35 of the camera 7.

[0096] Next, a fourth embodiment of the present disclosure will be described. Note that since the hardware configuration and the functional configuration of the information processing apparatus according to the fourth embodiment are the same as those of the third embodiment described above, detailed description thereof will be omitted here. In the fourth embodiment, first, the first detection model 22A that emphasizes the frame rate is selected to detect feature points from the RGB camera image, and according to the detection accuracy of the feature points, the third detection model 22C that emphasizes accuracy is used to detect feature points from the RGB camera image, or the fourth detection model 22D that emphasizes accuracy is used to detect feature points from the NIR camera image. This is different from the third embodiment.

[0097] Next, the processing performed in the fourth embodiment will be described. FIGS. 12 and 13 are flowcharts showing the processing performed in the third embodiment. For example, when an instruction to start shooting is input from the input device 15, the processing is started, and the camera control unit 21 starts shooting by the camera 7 to acquire a camera image including an RGB camera image and an NIR camera image (step ST61). Next, the selection unit 22 selects the first detection model 22A that emphasizes the frame rate (step ST62), and the feature point detection unit 23 detects feature points from the RGB camera image using the detection model selected by the selection unit 22 (step ST63).

[0098] Next, the shooting range specifying unit 24 determines the detection accuracy of the feature points (step ST64). If the shooting range specifying unit 24 determines that the detection accuracy is poor, the selection unit 22 selects the accuracy - focused third detection model 22C instead of the first detection model 22A (step ST65). Subsequently, the feature point detection unit 23 detects feature points from the RGB camera image (step ST66), and the shooting range specifying unit 24 determines the detection accuracy of the feature points (step ST67).

[0099] If the shooting range specifying unit 24 determines that the detection accuracy is poor, the selection unit 22 selects the accuracy - focused fourth detection model 22D that detects feature points from the NIR camera image instead of the third detection model 22C (step ST68). Note that instead of the fourth detection model 22D, the frame rate - focused second detection model 22B that detects feature points from the NIR camera image may be selected. Subsequently, the feature point detection unit 23 detects feature points from the NIR camera image (step ST69), and the shooting range specifying unit 24 determines the detection accuracy of the feature points (step ST70).

[0100] If it is determined that the detection accuracy is good, the shooting range specifying unit 24 specifies the shooting range during scout shooting based on the feature points (step ST71), and based on the specified shooting range, draws the scan start line and the scan end line during scout shooting on the schema displayed on the display 14 (line drawing; step ST72). On the other hand, if it is determined that the detection accuracy is poor, the shooting range specifying unit 24 performs a warning display (step ST73). When the warning display is performed, the information processing device 10 ends the shooting range specifying process. In this case, as described above, the operator manually sets the shooting range of the scout shooting.

[0101] If in step ST64 and step ST67, the shooting range specifying unit 24 determines that the detection accuracy is good, the process proceeds to step ST71, and the processes after step ST71 are performed.

[0102] Subsequently, it is determined whether or not an instruction to start scouting shooting has been given by the operator (step ST74). If step ST74 is negative, the process returns to step ST61, and the processes after step ST61 are repeated. If step ST74 is affirmative, the camera control unit 21 stops shooting by the camera 7 (step ST75), and the information processing apparatus 10 ends the shooting range specifying process.

[0103] As described above, in the fourth embodiment, first, feature points are detected from the RGB camera image using the first detection model 22A that emphasizes the frame rate. When the detection accuracy of the feature points detected using the first detection model 22A is good, the shooting range is specified using the feature points detected using the first detection model 22A that emphasizes the frame rate. When the detection accuracy of the feature points detected using the first detection model 22A is poor, feature points are detected from the RGB camera image using the third detection model 22C that emphasizes accuracy. When the detection accuracy of the feature points detected using the third detection model 22C is good, the shooting range is specified using the feature points detected using the third detection model 22C. When the detection accuracy of the feature points detected using the third detection model 22C is poor, feature points are detected from the NIR camera image using the fourth detection model 22D that emphasizes accuracy. Therefore, when setting the shooting range, feature points can be appropriately detected according to the detection accuracy of the feature points. As a result, the shooting range at the time of scouting shooting can be appropriately specified using the detected feature points.

[0104] Next, a fifth embodiment of the present disclosure will be described. Since the hardware configuration of the information processing apparatus according to the fifth embodiment is the same as that of the third embodiment described above, detailed description thereof will be omitted here. FIG. 14 is a diagram showing a functional configuration of the information processing apparatus according to the fifth embodiment. In FIG. 14, the same components as those in FIG. 10 are given the same reference numerals, and detailed description thereof will be omitted. The information processing apparatus 10B according to the fifth embodiment is different from the third embodiment in that it includes a motion detection unit 25 that detects the motion of the subject H.

[0105] The motion detection unit 25 detects the motion of the subject H. Specifically, it detects the two-dimensional motion of the feature points detected by the feature point detection unit 23 between temporally adjacent frames in the camera image. In the fifth embodiment, the selection unit 22 first selects the first detection model 22A that emphasizes the frame rate, and the feature point detection unit 23 detects feature points from the RGB camera image using the first detection model 22A. As the feature points for detecting the motion of the subject H, those corresponding to the photographed part may be used. For example, when the photographed part is the head, the nose, both eyes, or both ears may be used, and when the photographed part is the chest, both shoulders and the bases of the left and right feet may be used. Note that as the motion, a representative value of the motion of all 17 feature points may be obtained.

[0106] When the detected motion is equal to or greater than a predetermined threshold, the motion detection unit 25 determines that the motion is large, and when it is less than the threshold, it determines that the motion is small. The threshold for determining the magnitude of the motion is an example of a criterion.

[0107] In the fifth embodiment, when the motion detection unit 25 determines that the motion of the subject H is small, the selection unit 22 selects the third detection model 22C that emphasizes accuracy in detecting feature points from the RGB camera image and the fourth detection model 22D that emphasizes accuracy in detecting feature points from the NIR camera image instead of the first detection model 22A. Instead of the fourth detection model 22D, the second detection model 22B that emphasizes the frame rate may be selected. The feature point detection unit 23 detects feature points using both the third detection model 22C and the fourth detection model 22D selected by the selection unit 22. The shooting range specifying unit 24 specifies the shooting range using the feature points detected by the detection model with higher detection accuracy among the third detection model 22C and the fourth detection model 22D. On the other hand, when the motion detection unit 25 determines that the motion of the subject H is large, the feature point detection unit 23 continues to use the first detection model 22A used for detecting the feature points for detecting the motion to detect the feature points. The shooting range specifying unit 24 specifies the shooting range using the feature points detected using the first detection model 22A.

[0108] Next, the processing performed in the fifth embodiment will be described. FIGS. 15 and 16 are flowcharts showing the processing performed in the fifth embodiment. For example, when an instruction to start shooting is issued from the input device 15, the processing is started, and the camera control unit 21 starts shooting by the camera 7 to acquire a camera image including an RGB camera image and an NIR camera image (step ST81). Next, the selection unit 22 selects the first detection model 22A that emphasizes the frame rate (step ST82), and the feature point detection unit 23 detects feature points from the RGB camera image using the detection model selected by the selection unit 22 (step ST83).

[0109] Subsequently, the motion detection unit 25 detects the motion of the subject H using the feature points detected by the feature point detection unit 23 (step ST84), and determines whether the motion is large (step ST85). If the motion of the subject H is small and step ST85 is negated, the selection unit 22 first selects the third detection model 22C that emphasizes the accuracy of detecting feature points from the RGB camera image instead of the first detection model 22A (step ST86). Then, the feature point detection unit 23 detects feature points from the RGB camera image (step ST87). Subsequently, the selection unit 22 selects the fourth detection model 22D that emphasizes the accuracy of detecting feature points from the NIR camera image instead of the third detection model 22C (step ST88). Then, the feature point detection unit 23 detects feature points from the NIR camera image (step ST89).

[0110] Note that the processing of steps ST86 and ST87 and the processing of steps ST88 and ST89 may be performed in parallel, or the processing of steps ST88 and ST89 may be performed before the processing of steps ST86 and ST87.

[0111] Next, the shooting range specifying unit 24 compares the detection accuracy of feature points by the third detection model 22C (referred to as R3) with the detection accuracy of feature points by the fourth detection model 22D (referred to as R4) (step ST90). Specifically, the shooting range specifying unit 24 compares the representative value of the probabilities for the 17 feature points derived by the third detection model 22C with the representative value of the probabilities for the 17 feature points derived by the fourth detection model 22D. As the representative value, an average value, a median value, a weighted average value according to the shooting part, etc. can be used, but it is not limited thereto.

[0112] When the detection accuracy R3 of feature points by the third detection model 22C is equal to or higher than the detection accuracy R4 of feature points by the fourth detection model 22D (R3≥R4), the shooting range specifying unit 24 determines to use the feature points detected using the third detection model 22C for specifying the shooting range in order to specify the shooting range (step ST91). When the detection accuracy R3 of feature points by the third detection model 22C is less than the detection accuracy R4 of feature points by the fourth detection model 22D (R3<R4), the shooting range specifying unit 24 determines to use the feature points detected using the fourth detection model 22D for specifying the shooting range in order to specify the shooting range (step ST92).

[0113] Subsequently, the shooting range specifying unit 24 determines the detection accuracy of the feature points (step ST93). When step ST85 is affirmed, the process proceeds to step ST93, and the shooting range specifying unit 24 determines the detection accuracy for the feature points detected in step ST83.

[0114] When it is determined that the detection accuracy is good, the shooting range specifying unit 24 specifies the shooting range at the time of scouting shooting based on the feature points (step ST94), and based on the specified shooting range, the scan start line and the scan end line at the time of scouting shooting are drawn on the schema displayed on the display 14 (line drawing; step ST95). On the other hand, when it is determined that the detection accuracy is bad, the shooting range specifying unit 24 performs a warning display (step ST96). When the warning display is performed, the information processing apparatus 10 ends the shooting range specifying process. In this case, as described above, the operator manually sets the shooting range of the scouting shooting.

[0115] Subsequently, it is determined whether or not an instruction to start scouting shooting has been given by the operator (step ST97). If step ST97 is negated, the process returns to step ST81, and the processes after step ST81 are repeated. If step ST97 is affirmed, the camera control unit 21 stops the shooting by the camera 7 (step ST98), and the information processing apparatus 10B ends the shooting range specifying process.

[0116] As described above, in the fifth embodiment, first, feature points are detected from the RGB camera image by the first detection model 22A that emphasizes the frame rate. When the detection accuracy of the feature points detected using the first detection model 22A is good, the shooting range is specified using the feature points detected using the first detection model 22A. When the detection accuracy of the feature points detected using the first detection model 22A is bad, feature points are detected from the RGB camera image using the third detection model 22C that emphasizes accuracy, and feature points are detected from the NIR camera image using the fourth detection model 22D that emphasizes accuracy. Then, the shooting range is specified using the feature points with better detection accuracy. Therefore, when setting the shooting range, feature points can be appropriately detected using a detection model according to the detection accuracy of the feature points. As a result, the shooting range at the time of scouting shooting can be appropriately specified using the detected feature points.

[0117] In each of the above embodiments, the NIR camera 35 is provided in the camera 7, but the present invention is not limited to this. Instead of the NIR camera 35, a night vision camera may be used. Alternatively, instead of the NIR camera 35, a camera having a higher ISO sensitivity than the RGB camera 31 may be used.

[0118] In each of the above embodiments, the NIR camera 35 is a stereo camera, but the present invention is not limited to this. Only one NIR camera may be used.

[0119] In each of the above embodiments, the RGB camera 31 is provided in the camera 7, but the present invention is not limited to this. Instead of the RGB camera 31, a camera capable of capturing a monochrome image may be used.

[0120] In the above embodiment, the RGB camera 31 and the NIR camera 35 are provided in the camera 7, but the present invention is not limited to this. The RGB camera 31 and the NIR camera 35 may be provided separately.

[0121] In each of the above embodiments, the information processing apparatus according to the present disclosure is applied to the CT apparatus, but the present invention is not limited to this. The information processing apparatus according to the present disclosure may be applied to an MRI apparatus or the like as long as it is a photographing apparatus that acquires a scout image for setting a photographing range before the main photographing.

[0122] In each of the above embodiments, the information processing apparatus includes the photographing control unit 20, but the present invention is not limited to this. The photographing control unit 20 may be provided separately from the information processing apparatus.

[0123] In each of the above embodiments, for the detection model that emphasizes the frame rate, a plurality of detection models having different processing speeds for feature point detection may be used. Also, for the detection model that emphasizes accuracy, a plurality of detection models having different accuracies for feature point detection may be used.

[0124] Also, in the above-described fourth and fifth embodiments, the second detection model 22B that emphasizes frame rate and the fourth detection model 22D that emphasizes accuracy are used as the detection models for detecting feature points from the NIR camera images, but the present invention is not limited to this. The information processing apparatus 10B may have only one of the second detection model 22B that emphasizes frame rate and the fourth detection model 22D that emphasizes accuracy. Further, instead of the second detection model 22B and the fourth detection model 22D, one detection model capable of detecting feature points from the NIR camera images with a certain degree of frame rate and a certain degree of accuracy may be used.

[0125] Also, in the above embodiments, for example, as the hardware structure of a processing unit (Processing Unit) that executes various processes such as the shooting control unit 20, the camera control unit 21, the selection unit 22, the feature point detection unit 23, the shooting range specifying unit 24, and the motion detection unit 25, the following various processors (Processor) can be used. In addition to the CPU, which is a general-purpose processor that executes software (program) and functions as various processing units as described above, the above various processors include a programmable logic device (PLD), which is a processor whose circuit configuration can be changed after manufacturing, such as an FPGA (Field Programmable Gate Array), and a dedicated electric circuit, which is a processor having a circuit configuration designed specifically to execute specific processes, such as an ASIC (Application Specific Integrated Circuit).

[0126] One processing unit may be constituted by one of these various processors, or may be constituted by a combination of two or more processors of the same type or different types (for example, a combination of a plurality of FPGAs or a combination of a CPU and an FPGA). Further, a plurality of processing units may be constituted by one processor.

[0127] Examples of configuring a plurality of processing units with a single processor include: First, as represented by computers such as clients and servers, one processor is configured by a combination of one or more CPUs and software, and this processor functions as a plurality of processing units. Second, as represented by a System On Chip (SoC), there is a form in which a processor that realizes the functions of the entire system including a plurality of processing units with a single IC (Integrated Circuit) chip is used. Thus, various processing units are configured using one or more of the above various processors as a hardware structure.

[0128] Furthermore, as a hardware structure of these various processors, more specifically, circuitry combining circuit elements such as semiconductor elements can be used.

[0129] Hereinafter, the appended claims of the present disclosure will be described. (Appended Claim 1) Comprising at least one processor, The processor, Obtains at least one of a first camera image generated by video-capturing a subject on a bed with a first camera, and a second camera image generated by video-capturing the subject with a second camera having a higher shooting sensitivity than the first camera, Selects at least one detection model from a plurality of detection models including a first detection model constructed to detect a plurality of feature points on the subject included in the first camera image, and a second detection model constructed to detect the plurality of feature points on the subject included in the second camera image, Detects the plurality of feature points on the subject included in the first camera image or the second camera image using the selected detection model, An information processing apparatus that specifies the shooting range of the subject based on the plurality of feature points. (Appended Claim 2) When the brightness of the environment where the bed is installed is equal to or higher than a reference, the processor selects the first detection model, detects the plurality of feature points on the subject included in the first camera image, and when the brightness is lower than the reference, selects the second detection model and detects the plurality of feature points on the subject included in the second camera image. The information processing apparatus according to appended claim 1. (Appended claim 3) The processor acquires the first camera image and determines the brightness based on the luminance information derived from the first camera image. The information processing apparatus according to appended claim 2. (Appended claim 4) The processor acquires the first camera image and determines the brightness based on the noise included in the first camera image. The information processing apparatus according to appended claim 2. (Appended claim 5) The processor acquires the first camera image, detects the feature points from the first camera image using the first detection model, and determines the brightness based on the detection accuracy of the feature points. The information processing apparatus according to appended claim 2. (Appended claim 6) The processor determines the brightness by a sensor that detects the brightness of the environment. The information processing apparatus according to appended claim 2. (Appended claim 7) The processor acquires the first camera image, selects the first detection model and detects the feature points from the first camera image, determines whether the brightness is equal to or higher than a reference based on the first camera image. When the brightness is equal to or higher than the reference, the shooting range is specified based on the feature points detected using the first detection model. When the brightness is lower than the reference, the second detection model is selected, and the shooting range is specified based on the feature points detected using the second detection model. The information processing apparatus according to any one of appended claims 2 to 6. (Appended claim 8) The first detection model and the second detection model are models that emphasize the frame rate at the time of detecting the feature points. The plurality of detection models further includes a third detection model constructed to detect the plurality of feature points on the subject included in the first camera image, which emphasizes the accuracy at the time of detecting the feature points, and a fourth detection model constructed to detect the plurality of feature points on the subject included in the second camera image, which emphasizes the accuracy at the time of detecting the feature points. The information processing apparatus according to claim 1, wherein the processor selects the detection model according to the brightness of the environment where the bed is installed and the imaging part of the subject. (Claim 9) The information processing apparatus according to claim 8, wherein when the brightness is equal to or higher than a reference, the processor selects either the first detection model or the third detection model to detect the plurality of feature points on the subject included in the first camera image, and when the brightness is less than the reference, the processor selects either the second detection model or the fourth detection model to detect the plurality of feature points on the subject included in the second camera image. (Claim 10) The information processing apparatus according to claim 9, wherein the processor selects either one of the first detection model and the third detection model, and either one of the second detection model and the fourth detection model according to the imaging part of the subject. (Claim 11) The information processing apparatus according to any one of claims 1 to 10, wherein the processor determines the detection accuracy of the feature points, and when the detection accuracy is equal to or higher than a reference, specifies the imaging range based on the feature points, and when the detection accuracy is lower than the reference, issues a warning. (Claim 12) The first detection model is a model that emphasizes the frame rate at the time of detecting the feature points. The plurality of detection models further includes a third detection model constructed to detect the plurality of feature points on the subject included in the first camera image, which emphasizes the accuracy at the time of detecting the feature points. The processor acquires the first camera image and the second camera image, selects the first detection model, and detects the feature points from the first camera image. Determine the detection accuracy of the feature points detected using the first detection model, When the detection accuracy is equal to or higher than a first criterion, specify the shooting range based on the feature points detected using the first detection model, When the detection accuracy is lower than the first criterion, select the third detection model and detect the feature points from the first camera image using the third detection model, Determine the detection accuracy of the feature points detected using the third detection model, When the detection accuracy is equal to or higher than a second criterion, specify the shooting range based on the feature points detected using the third detection model, When the detection accuracy is lower than the second criterion, select the second detection model and detect the feature points from the second camera image using the second detection model, The information processing apparatus according to claim 1, wherein the shooting range is specified based on the feature points detected using the second detection model. (Claim 13) The processor determines the detection accuracy of the feature points detected using the first detection model when the detection accuracy is equal to or higher than the first criterion, the detection accuracy of the feature points detected using the third detection model when the detection accuracy is equal to or higher than the second criterion, or the detection accuracy of the feature points detected using the second detection model, When the detection accuracy is equal to or higher than a third criterion, specify the shooting range based on the feature points detected using the first detection model when the detection accuracy is equal to or higher than the first criterion, the feature points detected using the third detection model when the detection accuracy is equal to or higher than the second criterion, or the feature points detected using the second detection model, The information processing apparatus according to claim 12, wherein a warning is issued when the detection accuracy is lower than the third criterion. (Claim 14) The first detection model is a model that emphasizes the frame rate at the time of detecting the feature points, The plurality of detection models further includes a third detection model constructed to detect the plurality of feature points on the subject included in the first camera image, and emphasizing the accuracy at the time of detection of the feature points. The processor acquires the first camera image and the second camera image, selects the first detection model, and detects the feature points from the first camera image. Detect the movement of the subject based on the first camera image. When the movement of the subject is equal to or greater than a first criterion, the shooting range is specified based on the feature points detected using the first detection model. When the movement of the subject is less than the first criterion, the second detection model and the third detection model are selected. The third detection model is used to detect the feature points from the first camera image. The second detection model is used to detect the feature points from the second camera image. Compare the detection accuracy of the feature points detected using the third detection model with the detection accuracy of the feature points detected using the second detection model. When the detection accuracy of the feature points detected using the third detection model is higher, the shooting range is specified based on the feature points detected using the third detection model. The information processing apparatus according to appended claim 1, wherein when the detection accuracy of the feature points detected using the second detection model is higher, the shooting range is specified based on the feature points detected using the second detection model. (Appended claim 15) The processor determines the detection accuracy of the feature points detected using the first detection model, the detection accuracy of the feature points detected using the third detection model, or the detection accuracy of the feature points detected using the second detection model when the movement is equal to or greater than the first criterion. When the detection accuracy is equal to or higher than a second criterion, if the movement is equal to or higher than the first criterion, the shooting range is specified based on the feature points detected using the first detection model, the feature points detected using the third detection model, or the feature points detected using the second detection model. An information processing apparatus according to appended claim 14 that issues a warning when the detection accuracy is lower than the second criterion. (Appended claim 16) The information processing apparatus according to any one of appended claims 1 to 15, wherein the processor derives a moving range of the bed based on the shooting range. (Appended claim 17) The information processing apparatus according to appended claim 16, wherein the processor displays a human body image simulating a human body on a display, and draws a moving start line and a moving end line of the bed based on the moving range of the bed on the human body image. (Appended claim 18) The information processing apparatus according to any one of appended claims 1 to 17, wherein the shooting range is a shooting range when shooting a positioning image acquired before actual shooting of the subject. (Appended claim 19) A computer acquires at least one of a first camera image generated by video shooting a subject on a bed with a first camera, and a second camera image generated by video shooting the subject with a second camera having a higher shooting sensitivity than the first camera. Select at least one detection model from a plurality of detection models including a first detection model constructed to detect a plurality of feature points on the subject included in the first camera image, and a second detection model constructed to detect the plurality of feature points on the subject included in the second camera image. Detect the plurality of feature points on the subject included in the first camera image or the second camera image using the selected detection model. An information processing method for specifying a shooting range of a subject based on the plurality of feature points. (Appended claim 20) A procedure for acquiring at least one of a first camera image generated by video-recording a subject on a bed with a first camera, and a second camera image generated by video-recording the subject with a second camera having a higher shooting sensitivity than the first camera, A procedure for selecting at least one detection model from a plurality of detection models including a first detection model constructed to detect a plurality of feature points on the subject included in the first camera image, and a second detection model constructed to detect the plurality of feature points on the subject included in the second camera image, A procedure for detecting the plurality of feature points on the subject included in the first camera image or the second camera image using the selected detection model, An information processing program for causing a computer to execute a procedure for specifying the shooting range of the subject based on the plurality of feature points.

Explanation of Signs

[0130] 1 CT apparatus 2 Gantry 3 Bed 3A Bed section 3B Base 3C Drive section 4 Console 5 Aperture 7 Camera 10, 10A, 10B Information processing apparatus 11 CPU 12 Information processing program 13 Storage 14 Display 15 Input device 16 Memory 17 Network I / F 18 Bus 20 Shooting control section 21 Camera control section 22 Selection section 22A First detection model 22B Second detection model 22C Third detection model 22D Fourth detection model 23 Feature point detection unit 24 Shooting range specification unit 25 Motion detection unit 31 RGB camera 32 RGB sensor 33 Base 35 NIR camera 36, 37 NIR sensors 38 NIR projector 40 Schema 41 Start line 42 End line H Subject

Claims

1. At least one processor; The processor, At least one of a first camera image generated by taking a moving image of a subject lying on a bed with a first camera and a second camera image generated by taking a moving image of the subject with a second camera having a higher imaging sensitivity than the first camera is obtained; selecting at least one detection model from a plurality of detection models including a first detection model constructed to detect a plurality of feature points on the subject included in the first camera image and a second detection model constructed to detect the plurality of feature points on the subject included in the second camera image; detecting the plurality of feature points on the subject included in the first camera image or the second camera image using the selected detection model; An information processing device that identifies a shooting range of the subject based on the plurality of feature points.

2. 2. The information processing device according to claim 1, wherein the processor selects the first detection model when the brightness of the environment in which the bed is placed is equal to or higher than a reference value, and detects the plurality of feature points on the subject included in the first camera image, and selects the second detection model when the brightness is less than the reference value, and detects the plurality of feature points on the subject included in the second camera image.

3. The information processing device of claim 2 , wherein the processor acquires the first camera image and determines the brightness based on luminance information derived from the first camera image.

4. The information processing apparatus according to claim 2 , wherein the processor acquires the first camera image and determines the brightness based on noise contained in the first camera image.

5. The information processing apparatus according to claim 2 , wherein the processor acquires the first camera image, detects the feature points from the first camera image using the first detection model, and determines the brightness based on detection accuracy of the feature points.

6. The information processing apparatus according to claim 2 , wherein the processor determines the brightness using a sensor that detects the brightness of the environment.

7. 7. The information processing device of claim 2, wherein the processor acquires the first camera image, selects the first detection model, detects the feature points from the first camera image, determines whether the brightness is equal to or greater than a reference value based on the first camera image, and if the brightness is equal to or greater than the reference value, identifies the shooting range based on the feature points detected using the first detection model, and if the brightness is less than the reference value, selects the second detection model and identifies the shooting range based on the feature points detected using the second detection model.

8. the first detection model and the second detection model are models that prioritize a frame rate when detecting the feature points, the plurality of detection models further include a third detection model that is constructed to detect the plurality of feature points on the subject included in the first camera image and that emphasizes accuracy in detecting the feature points, and a fourth detection model that is constructed to detect the plurality of feature points on the subject included in the second camera image and that emphasizes accuracy in detecting the feature points; The information processing apparatus according to claim 1 , wherein the processor selects the detection model according to brightness of an environment in which the bed is placed and an imaging region of the subject.

9. 9. The information processing device of claim 8, wherein the processor selects either the first detection model or the third detection model when the brightness is equal to or greater than a reference value, and detects the plurality of feature points on the subject included in the first camera image, and selects either the second detection model or the fourth detection model when the brightness is less than the reference value, and detects the plurality of feature points on the subject included in the second camera image.

10. The information processing device according to claim 9 , wherein the processor selects either the first detection model or the third detection model, and either the second detection model or the fourth detection model, depending on an imaging part of the subject.

11. The information processing device according to claim 1 , wherein the processor determines detection accuracy of the feature points, and if the detection accuracy is equal to or higher than a reference level, identifies the shooting range based on the feature points, and if the detection accuracy is lower than the reference level, issues a warning.

12. the first detection model is a model that places importance on a frame rate at the time of detecting the feature points, the plurality of detection models further includes a third detection model that is constructed to detect the plurality of feature points on the subject included in the first camera image and that emphasizes accuracy in detecting the feature points; the processor acquires the first camera image and the second camera image, selects the first detection model, and detects the feature points from the first camera image; determining a detection accuracy of the feature points detected using the first detection model; If the detection accuracy is equal to or greater than a first standard, the imaging range is identified based on the feature points detected using the first detection model; If the detection accuracy is lower than the first standard, selecting the third detection model, and detecting the feature points from the first camera image using the third detection model; determining a detection accuracy of the feature points detected using the third detection model; If the detection accuracy is equal to or greater than a second standard, the imaging range is identified based on the feature points detected using the third detection model; If the detection accuracy is lower than the second standard, selecting the second detection model, and detecting the feature points from the second camera image using the second detection model; The information processing apparatus according to claim 1 , wherein the shooting range is identified based on the feature points detected using the second detection model.

13. the processor determines a detection accuracy of the feature points detected using the first detection model when the detection accuracy is equal to or greater than the first standard, a detection accuracy of the feature points detected using the third detection model when the detection accuracy is equal to or greater than the second standard, or a detection accuracy of the feature points detected using the second detection model; When the detection accuracy is equal to or greater than a third criterion, identifying the shooting range based on the feature points detected using the first detection model when the detection accuracy is equal to or greater than the first criterion, the feature points detected using the third detection model when the detection accuracy is equal to or greater than the second criterion, or the feature points detected using the second detection model; The information processing apparatus according to claim 12 , further comprising: a warning being issued when the detection accuracy is lower than the third standard.

14. the first detection model is a model that places importance on a frame rate at the time of detecting the feature points, the plurality of detection models further includes a third detection model that is constructed to detect the plurality of feature points on the subject included in the first camera image and that emphasizes accuracy in detecting the feature points; The processor acquires the first camera image and the second camera image, selects the first detection model, and detects the feature points from the first camera image; Detecting a movement of the subject based on the first camera image; When the motion of the subject is equal to or greater than a first criterion, the imaging range is identified based on the feature points detected using the first detection model; selecting the second detection model and the third detection model if the subject's movement is less than the first criterion; Detecting the feature points from the first camera image using the third detection model; Detecting the feature points from the second camera image using the second detection model; comparing a detection accuracy of the feature points detected using the third detection model with a detection accuracy of the feature points detected using the second detection model; When the detection accuracy of the feature points detected using the third detection model is higher, the shooting range is specified based on the feature points detected using the third detection model; The information processing device according to claim 1 , wherein, when a detection accuracy of the feature points detected using the second detection model is higher, the shooting range is identified based on the feature points detected using the second detection model.

15. the processor determines, when the movement is equal to or greater than the first criterion, a detection accuracy of the feature points detected using the first detection model, a detection accuracy of the feature points detected using the third detection model, or a detection accuracy of the feature points detected using the second detection model; When the detection accuracy is equal to or greater than a second criterion, identifying the shooting range based on the feature points detected using the first detection model, the feature points detected using the third detection model, or the feature points detected using the second detection model when the movement is equal to or greater than the first criterion; The information processing apparatus according to claim 14 , further comprising: a warning being issued when the detection accuracy is lower than the second standard.

16. The information processing apparatus according to claim 1 , wherein the processor derives a movement range of the bed based on the imaging range.

17. The information processing device according to claim 16 , wherein the processor displays a human body image simulating a human body on a display, and draws a movement start line and a movement end line of the bed based on a movement range of the bed on the human body image.

18. The information processing apparatus according to claim 1 , wherein the imaging range is a range for imaging a positioning image acquired before performing actual imaging of the subject.

19. A computer acquires at least one of a first camera image generated by taking a moving image of a subject lying on a bed with a first camera and a second camera image generated by taking a moving image of the subject with a second camera having a higher imaging sensitivity than the first camera; selecting at least one detection model from a plurality of detection models including a first detection model constructed to detect a plurality of feature points on the subject included in the first camera image and a second detection model constructed to detect the plurality of feature points on the subject included in the second camera image; detecting the plurality of feature points on the subject included in the first camera image or the second camera image using the selected detection model; An information processing method for identifying a shooting range of the subject based on the plurality of feature points.

20. acquiring at least one of a first camera image generated by taking a moving image of a subject lying on a bed with a first camera and a second camera image generated by taking a moving image of the subject with a second camera having a higher imaging sensitivity than the first camera; selecting at least one detection model from a plurality of detection models including a first detection model constructed to detect a plurality of feature points on the subject included in the first camera image and a second detection model constructed to detect the plurality of feature points on the subject included in the second camera image; detecting the plurality of feature points on the subject included in the first camera image or the second camera image using the selected detection model; and a step of specifying a photographing range of the subject based on the plurality of feature points.

Images