Program, moving body tracking device, radiation therapy system, and moving body tracking method

Abstract

To provide a program, a moving body tracking device, a radiation therapy system, and a moving body tracking method capable of improving the reliability of moving body tracking compared to conventional techniques.SOLUTION: A program executed by a moving body tracking calculation device 15 that performs position calculation of a tracking target B causes the moving body tracking calculation device 15 to execute: a processing procedure for performing a plurality of pieces of image recognition processing on an image in which the tracking target B appears; and a determination procedure for determining success or failure of position detection on the basis of detection positions of the tracking target B obtained by each of the plurality of pieces of image recognition processing in the processing procedure and a weighted average position obtained by weighting with the reliability of each image recognition processing.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to a radiotherapy system that irradiates an affected area such as a tumor with radiation such as charged particles or X-rays to provide treatment, a program suitable for a moving body tracking device used therein, a moving body tracking device, a radiotherapy system, and a method for tracking a moving body. [Background technology]

[0002] As an example of a radiation irradiation system and a moving object tracking device that can accurately detect a tracked object even under difficult X-ray fluoroscopy conditions, such as when the subject is thick, Patent Document 1 describes a system that performs template matching on two fluoroscopic images using a template image representing a pre-prepared marker, lists positions with high matching scores as candidate marker positions, calculates the length of a common perpendicular line for all combinations from the two candidate marker position lists, detects the position of the marker based on the matching score and the common perpendicular line, and controls the emission of a proton beam to be irradiated onto the target based on the detected marker position. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2017-209243 Summary of the Invention [Problem to be solved by the invention]

[0004] A real-time tumor-tracking radiotherapy system is known that recognizes the position of the target in real time and irradiates the therapeutic radiation in order to reduce irradiation of organs around the target by concentrating the therapeutic radiation on the target (see, for example, Patent Document 1).

[0005] The target position is recognized by an image recognition method that takes X-ray fluoroscopic images from multiple directions and recognizes the target position from those images. There are two methods for recognizing the target position: direct recognition of the target and indirect recognition of the target from tracking objects such as internal landmarks like the diaphragm or substitute markers inserted near the target.

[0006] In either case, a projected image of the target or tracked object is prepared in advance as a template image, and the two-dimensional position of the target is calculated by performing a matching process (template matching) that searches for the location in the captured image that is most similar to the template image.The three-dimensional position of the tracked object is then calculated from the two-dimensional position of the target obtained from the captured images in each direction and the geometric information of the imaging system.

[0007] When using X-ray fluoroscopy from two directions, the position where the two lines connecting the position on the X-ray measuring device where the tracked object is imaged and the X-ray generator for fluoroscopy are closest is considered to be the position of the tracked object. At the position where these two lines are closest, a line perpendicular to the two lines can be drawn. This line is called the common perpendicular line, and the midpoint of the common perpendicular line is considered to be the position of the tracked object.

[0008] When tracking a target directly, the three-dimensional position of the tumor can be recognized directly, but when tracking an internal landmark or surrogate marker, the target position can be recognized indirectly by calculating in advance the relationship between the three-dimensional position of the tracking target and the three-dimensional position of the target.

[0009] By controlling the irradiation of therapeutic radiation by recognizing the target position, it becomes possible to accurately irradiate organs that move during treatment, such as respiratory moving organs.

[0010] Based on the differences in the methods for controlling the irradiation of therapeutic radiation, two irradiation control methods have been proposed: intercept irradiation, in which irradiation is performed at a time when the target position and the radiation irradiation position according to the treatment plan coincide, and tracking irradiation, in which the radiation irradiation position is changed to match the target position.

[0011] In any irradiation control method, in order to achieve highly accurate irradiation, it is important to determine with high accuracy whether the position of the tracked object has been recognized.

[0012] Image recognition methods, such as template matching, use a template image and an X-ray fluoroscopic image to calculate a similarity measure such as Zero Means Normalized Cross Correlation (ZNCC), search for the position where the similarity measure is largest and the X-ray fluoroscopic image is most similar to the template image, and use this position as the recognition position of the tracking target.

[0013] On the other hand, when the X-ray fluoroscopy conditions are severe, such as when the subject is thick, if a structure similar to the tracking target appears near the tracking target on the fluoroscopy image, something that is not the tracking target may be mistakenly detected as the tracking target.In this case, the distance between the position on the X-ray measuring device where the tracking target is captured and the two lines connecting the fluoroscopy X-ray generator increases, which increases the length of the common perpendicular line of the tracking target and reduces the accuracy of position recognition of the tracking target.

[0014] In Patent Document 1, a method is reported in which, in order to improve the recognition accuracy of the three-dimensional position of a tracked object, a matching score value and a common perpendicular line are calculated, a value representing the accuracy of this detection and a value representing the correlation are weighted, and the position of the tracked object is detected based on the weighted result. However, even when this method is adopted, the detection accuracy decreases on the epipolar line where the length of the common perpendicular line is small, making it difficult to determine false detections.

[0015] Furthermore, when X-ray fluoroscopic images are taken frequently at regular intervals (for example, 30 Hz), the interval between images is sufficiently short compared to the respiratory cycle, so it is possible to determine false detection from the amount of displacement from the recognized position of the tracked target in the fluoroscopic image taken at the previous timing, the moving speed, etc. However, when the interval between X-ray fluoroscopic images is long, such as 1-2 Hz, it is difficult to accurately determine false detection from the recognition result of the tracked target taken at the previous timing.

[0016] The present invention provides a program, a moving body tracking device, a radiotherapy system, and a moving body tracking method that can improve the reliability of moving body tracking compared to conventional methods. [Means for solving the problem]

[0017] The present invention includes a number of means for solving the above-mentioned problems. One example is a program executed by a computing system that calculates the position of a tracked object, which causes the computing system to execute a processing procedure for performing a number of image recognition methods on an image showing the tracked object, and a determination procedure for determining whether position detection was successful or not based on the detected position of the tracked object in each of the multiple image recognition methods in the processing procedure and a weighted average position weighted by the reliability of each of the image recognition methods. [Effects of the Invention]

[0018] According to the present invention, it is possible to improve the reliability of moving object tracking compared to the prior art. Problems, configurations, and effects other than those described above will become apparent from the following description of the embodiments. [Brief explanation of the drawings]

[0019] [Figure 1] 1 is a diagram illustrating the overall configuration of an X-ray therapy system equipped with a moving body tracking device according to an embodiment. [Figure 2] FIG. 10 is a flowchart showing the recognition position of a landmark and the determination of whether the recognition is successful or not in the moving object tracking device of the embodiment. [Figure 4] 10A and 10B are diagrams illustrating a method for calculating the reliability of an image recognition method in the moving object tracking device of the embodiment. [Figure 5] 10A and 10B are diagrams illustrating a method for calculating the reliability of an image recognition method in the moving object tracking device of the embodiment. [Figure 3] 10A and 10B are diagrams illustrating a method for calculating the reliability of an image recognition method in the moving object tracking device of the embodiment. [Figure 6] FIG. 10 is an explanatory diagram regarding the success or failure determination of landmark recognition in the moving object tracking device of the embodiment. [Figure 7]FIG. 10 is an explanatory diagram regarding the success or failure determination of landmark recognition in the moving object tracking device of the embodiment. [Figure 8] FIG. 10 is an explanatory diagram regarding the success or failure determination of landmark recognition in the moving object tracking device of the embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0020] Embodiments of the program, moving body tracking device, radiotherapy system, and moving body tracking method of the present invention will be described with reference to FIGS.

[0021] In the drawings used in this specification, identical or corresponding components are denoted by the same or similar reference numerals, and repeated description of these components may be omitted.

[0022] In the following examples, X-rays are used as radiation, but the radiation used is not limited to X-rays, and particle beams such as protons, helium, and carbon can also be used.

[0023] Furthermore, we will explain a form in which the position of a moving object is tracked and the radiation irradiation position is changed relative to the target, but the present invention can also be applied to a form in which the position of a moving object is tracked and radiation is irradiated only when the target is located within a specified area.

[0024] First, the overall configuration of an X-ray therapy system of the present invention will be described with reference to Fig. 1. Fig. 1 is a diagram showing the overall configuration of an X-ray therapy system equipped with a dynamic body tracking device of an embodiment.

[0025] The X-ray therapy system 10 shown in Fig. 1 is a radiotherapy device having a group of devices for irradiating a tumor present in a subject, a patient A, with an X-ray particle beam as a target. The X-ray therapy system 10 shown in Fig. 1 includes an X-ray irradiation device, X-ray detectors 12A and 12B, X-ray tubes 13A and 13B, an infrared camera 14, a moving subject tracking calculation device 15, a main control device 16, a communication device 17, a storage device 18, an input device 19, and a fluoroscopic X-ray imaging device 20.

[0026] Among these, the X-ray detectors 12A and 12B and the X-ray tubes 13A and 13B constitute one or more sets of fluoroscopic devices that capture fluoroscopic images of the tracked object B. Note that, although the treatment system of this embodiment is shown as an example in which two sets of fluoroscopic devices are provided, the number of fluoroscopic devices may be one set or three or more sets, and is not particularly limited.

[0027] The X-ray detectors 12A and 12B, the X-ray tubes 13A and 13B, and the object tracking calculation device 15 constitute an object tracking device.

[0028] The X-ray irradiation device includes a gantry 110 , an irradiation nozzle 111 for emitting therapeutic radiation, and a bed 112 .

[0029] When an extraction start signal is output from the main controller 16, the linear accelerator in the irradiation nozzle 111 accelerates the electron beam, and the accelerated electron beam is irradiated onto the tungsten to generate X-rays.

[0030] The irradiation nozzle 111 is provided with a collimator (hereinafter referred to as a multi-leaf collimator (MLC)) consisting of multiple plate-shaped shields (hereinafter referred to as leaves) arranged on the left and right, and the generated X-rays are shaped into a desired distribution by arbitrarily changing the position of each leaf based on instructions from the main control device 16. Furthermore, the irradiation nozzle 111 is provided with a dose monitor that measures the X-ray irradiation dose, and the detected measurement value is output to the radiation irradiation control device and used for control during radiation irradiation.

[0031] The irradiation nozzle 111 is mounted on the gantry 110, and receives instructions from the main control device 16 to set the gantry rotation angle, making it possible to irradiate X-rays onto a patient A on a bed 112 from any angle. The center of rotation of this gantry 110 is called the isocenter. The irradiation nozzle 111 also has a gimbal mechanism that swings the irradiation nozzle 111 laterally (x and y directions) to track respiratory-moving organs.

[0032] The main control device 16 is connected to the gantry 110, irradiation nozzle 111, bed 112, moving object tracking calculation device 15, communication device 17, memory device 18, input device 19, fluoroscopic X-ray imaging device 20, etc., and controls the equipment inside the irradiation nozzle 111, the gantry 110, bed 112, and fluoroscopic X-ray imaging device 20.

[0033] The communication device 17 is connected to a data server 172 via a network, and obtains irradiation parameters (gantry angle, irradiation dose, leaf position information, etc.) created by a treatment planning device 171 via the network before irradiation from the data server 172, and stores the irradiation parameters in a storage device 18.

[0034] The input device 19 is connected to the main control device 16, receives input signals from a medical professional operating the X-ray therapy system 10, and transmits various control signals to the main control device 16. When the main control device 16 receives an instruction to start radiation irradiation via the input device 19, it starts irradiation based on the irradiation parameters stored in the storage device 18.

[0035] The X-ray therapy system 10 includes an X-ray tube 13A that irradiates X-rays onto a patient A from a first direction, an X-ray detector 12A that detects the two-dimensional dose distribution of the X-rays irradiated from the X-ray tube 13A and transmitted through the patient A, and a signal processing circuit (not shown).

[0036] The X-ray detector 12A has a plurality of two-dimensionally arranged detecting elements (more specifically, semiconductor elements that convert radiation into electric charges, for example), and outputs analog signals from the detecting elements. The signal processing circuit processes the analog signals from the X-ray detector 12A to generate X-ray fluoroscopic image data and transmits it to the tumor position recognition device 150.

[0037] It should be noted that the X-ray fluoroscopy device captures images at a frequency sufficient to confirm the movement of the tracking target B (for example, about 1-2 Hz).

[0038] On the other hand, the infrared camera 14 captures images at a frequency (for example, about 60 Hz) that allows the movement of the body surface caused by the infrared markers C to be captured with sufficient accuracy.

[0039] Similarly, the X-ray therapy system 10 includes an X-ray tube 13B that irradiates X-rays onto patient A from a second direction (in this embodiment, a direction perpendicular to the first direction), an X-ray detector 12B that detects a two-dimensional dose distribution of the X-rays irradiated from this X-ray tube 13B and transmitted through patient A, and a signal processing circuit (not shown). The X-ray detector 12B has a plurality of detecting elements arranged two-dimensionally and outputs analog signals from these detecting elements. The signal processing circuit processes the analog signals from the X-ray detector 12B to generate X-ray fluoroscopic image data and transmits it to the tumor position recognition device 150. Note that imaging by the X-ray tube 13B is performed synchronously with imaging by the X-ray tube 13A.

[0040] Here, the X-ray emission intervals of the X-ray tubes 13A and 13B, as well as imaging parameters such as tube voltage and tube current, are set by the X-ray fluoroscopic imaging device 20 via the input device 19.

[0041] The fluoroscopic X-ray imaging device 20 controls the emission of transmission X-rays based on a signal from the main control device 16 .

[0042] The moving object tracking calculation device 15 is a device that executes the program described below and determines the position of the tracking target B, and is composed of a tumor position recognition device 150, a tumor position prediction device 151, a displacement amount calculation device 152, an input device 153 that provides input information to each device, and a display device 154 that displays the output from each device.

[0043] Tumor position recognition device 150 recognizes the position of tracking target B in real time from images captured using X-ray tubes 13A and 13B that capture X-ray images of tracking target B in patient A on bed 112 from multiple directions, and calculates the tumor position. This tumor position recognition device 150 has tumor position recognition unit 150a and recognition success / failure determination unit 150b.

[0044] The tumor position recognition unit 150a calculates the landmark position for each perspective image using a plurality of different image recognition methods using the same template image. This tumor position recognition unit 150a preferably executes a processing procedure (process) that performs a plurality of image recognition methods on the image showing the tracking target B.

[0045] The recognition success / failure determination unit 150b determines the success / failure of landmark recognition based on the positional relationship between the weighted average position, which is the landmark position calculated by the tumor position recognition unit 150a weighted using the reliability of each image recognition method, and the landmark position obtained by each image recognition method.

[0046] This recognition success / failure determination unit 150b is preferably the entity that executes a determination procedure (determination step) that determines the success / failure of position detection based on the detected position of the tracked target B in each of the multiple image recognition processing methods in the processing procedure and a weighted average position weighted by the reliability of each image recognition processing method.

[0047] In addition, steps S201 to S204 in FIG. 2, which will be described later, correspond to the processing procedure, and steps S205 to S207 correspond to the determination procedure.

[0048] The tumor position prediction device 151 recognizes the position information of the body surface from an image captured by an infrared camera 14 of an infrared marker C placed on the body surface of patient A, and predicts the tumor position from the body surface position using a correlation model between the body surface and tumor position.

[0049] The displacement amount calculation device 152 has a tracking success / failure determination unit 152a and a displacement amount calculation unit 152b, and calculates the amount of displacement from the previous image capture timing of the tracked object B based on the tumor position recognition result of the tumor position recognition device 150 and the tumor position prediction result of the tumor position prediction device 151.

[0050] The main control device 16, the moving object tracking calculation device 15, and the fluoroscopic X-ray image capturing device 20 may each have a central processing unit (CPU) and a memory connected to this CPU, or these devices may be configured as an integrated computer, and are not particularly limited thereto.

[0051] The control processes for the operations to be executed may be integrated into one program, or may be divided into multiple programs, or may be a combination of these.

[0052] Some or all of the programs stored in each device may be implemented using dedicated hardware or may be modularized. Furthermore, various programs may be installed in each device from a program distribution server or external storage media, or may be used to update existing devices.

[0053] Furthermore, each device may be an independent device connected via a wired or wireless network, or two or more devices may be integrated together.

[0054] Next, the process from recognizing landmark positions in X-ray fluoroscopic images to determining whether the recognition has been successful, which is carried out by the tumor position recognition device 150 during dynamic tracking irradiation, a feature of the present invention, will be described with reference to Figures 2 to 8. Figure 2 is a flow diagram showing the recognized positions of landmarks in the moving object tracking device of the embodiment and determining whether the recognition has been successful, Figures 3 to 5 are diagrams explaining a method for calculating the reliability of the image recognition method in the moving object tracking device of the embodiment, and Figures 6 to 8 are explanatory diagrams relating to the determination of whether the landmark recognition has been successful in the moving object tracking device of the embodiment.

[0055] Figure 2 shows a flow diagram for calculating landmark positions from X-ray fluoroscopic images and determining whether the landmark positions have been recognized successfully during dynamic tracking irradiation. Here, we will explain an example in which the diaphragm is set as a landmark and tracked.

[0056] First, the operator sets a template image or a group of template images for tracking the diaphragm corresponding to patient A in tumor position recognition unit 150a (step S201).

[0057] The template image is preferably an image of the diaphragm and its surroundings cut out from a pseudo-fluoroscopic X-ray image created by arranging a 3D image of patient A in the same virtual space as the imaging system that generates the fluoroscopic X-ray image, using the CT image data or 4DCT image data used when creating the treatment plan for patient A, and projecting the image, but an image of the diaphragm and its surroundings cut out from a fluoroscopic X-ray image acquired before treatment may also be used. Furthermore, when performing continuous rotational irradiation such as VMAT, a group of template images created from pseudo-fluoroscopic X-ray images created at multiple exposure angles is set.

[0058] Next, the image recognition method to be used for detecting the diaphragm position is selected for the tumor position recognition unit 150a (step S202). Here, the case where template matching using an edge-enhanced image and template matching using a normal image are selected will be described as an example, but template matching using a feature-enhanced image obtained by phase-only correlation or a Gabor filter may also be used.

[0059] Next, the tumor position recognition unit 150a performs preprocessing of the fluoroscopic image and template image required to implement the selected image recognition method (step S203). For edge enhancement, a Sobel filter capable of simultaneously implementing flattening and edge enhancement is applied, and a sliding window process using a Sobol kernel in the body axis direction (y direction) is performed to create edge-enhanced images of the template image and the X-ray fluoroscopic image.

[0060] Next, as a matching process, the tumor position recognition unit 150a calculates a similarity measure such as normalized cross-correlation (ZNCC) for the contour-enhanced image and the normal image, searches for the position where the similarity measure is largest and the X-ray fluoroscopic image is most similar to the template image, and sets this position as the recognition position of the tracking target (step S204).

[0061] Next, the recognition success / failure determination unit 150b sets and acquires a reliability for the image recognition method selected in step S202 (step S205). The reliability may be set by the operator when selecting the method in step S202, or a fixed value may be set in advance for each method. In this way, the processing procedure allows the reliability to be set or calculated.

[0062] Here, an example will be described in which the reliability is calculated using the similarity scale map used when calculating the landmark recognition position in step S204 for an image showing a recognition position 300 on an image captured by the X-ray tube 13A-X-ray detector 12A and recognition position candidate points 301, 302, 303, and 304 on an image captured by the X-ray tube 13A-X-ray detector 12A as shown in FIG.

[0063] Figures 4 and 5 show the different similarity scale distributions depending on the detection method.

[0064] Figure 4 shows the results of image recognition method A, such as template matching processing using raw images. Since the similarity scale distribution is generally smooth, it is expected that the recognized position will fluctuate with a slight difference in the similarity scale. In other words, since the score distribution is gentle, the reliability w of the recognized detection position 1A a is considered to be low.

[0065] In contrast, Figure 5 shows the results of image recognition method B, such as template matching processing using an edge-enhanced image. Because the similarity scale map is steep, it is assumed that there is little possibility that the recognized position will fluctuate due to differences in the similarity scale distribution. In other words, the reliability w of the detected position 1B by image recognition method B, which has a steep score distribution of the similarity scale, is b is considered to be high.

[0066] In this way, in order to set the reliability of the image recognition method, the processing procedure focuses on the average gradient information of the similarity scale map and calculates the reliability based on the gradient information of the similarity scale distribution calculated by the image recognition method.

[0067] The recognition success / failure determination unit 150b extracts the recognition position candidates having the top N similarity scale values ​​from the similarity scale map calculated in step S204, and calculates the average gradient information of the similarity scale map, which is expressed by the following equation (1).

[0068]

number

[0069] where r i is the distance from the position where the similarity scale distribution is maximum to each recognition position candidate, and is the coordinate (x k,max ,y k,max ) is expressed by the following equation (2).

[0070]

number

[0071] Therefore, by using the value of equation (1), the average gradient information of the similarity scale distribution can be set as the reliability of the image recognition method.

[0072] Next, the recognition success / failure determination unit 150b calculates a weighted average position based on the reliability calculated in step S205 and the diaphragm tracking position calculated in step S204 (recognition position 1A in FIG. 4 and recognition position 1B in FIG. 5) (step S206).

[0073] This weighted average position can be calculated by a general method. In this step, the recognition position 1A in FIG. 4 is calculated by x1 and its reliability is calculated by w a ,The recognition position 1B in Figure 5 is x2 and its reliability is w b When we assume that (w a x1+w b x2) / (w a +w b ) can be calculated.

[0074] Next, the recognition success / failure determination unit 150b determines whether the diaphragm has been recognized based on the weighted average position and the recognized position of the diaphragm (landmark recognized position) by each tracking method, taking into account the reliability of the image recognition method (step S207). Here, in this determination procedure, the success / failure can be determined based on whether or not a detected position by one or more of the image recognition methods exists within a threshold distance from the weighted average position.

[0075] Hereinafter, an example will be described in which the success / failure determination is performed using a threshold distance ΔR from the weighted average position. This threshold distance ΔR is assumed to be determined in advance at the manufacturing stage, but may also be set by the user.

[0076] In the determination procedure, if a detected position by one image recognition method exists within a threshold distance ΔR from the weighted average position, it is determined that the position detection was successful, and the detected position can be considered as the position of the tracked object B on the image.

[0077] For example, as shown in Figure 6, the reliability w a <Reliability w b If there is a difference in the reliability of the two image recognition methods, the weighted average position 1C1 is set to a position close to the detected position 1B1 by the image recognition method B on the other side, and only the detected position 1B1 by the image recognition method B is within the threshold distance ΔR from the weighted average position 1C1, and the detected position 1A1 by the image recognition method A is not within the threshold distance ΔR from the weighted average position 1C1. In this case, the detection is deemed successful, and the detected position 1B1 by the image recognition method B, which is close to the weighted average position 1C1, is set as the recognized position on the image captured by the X-ray tube 13A-X-ray detector 12A.

[0078] In addition, in the judgment procedure, if there are detected positions using multiple image recognition processing methods within a threshold range from the weighted average position, it is judged that the position detection was successful, and any of the detected positions closest to the weighted average position, the weighted average position itself, or the detected position using the most reliable image recognition processing method can be considered to be the position of the tracked object B on the image.

[0079] For example, as shown in Figure 7, the reliability of two image recognition methods A and B is a ,w b If the difference between the weighted average position 1C2 and the detected positions 1A2 and 1B2 by both methods is small and exist within the range of the threshold distance ΔR from the weighted average position 1C2, the weighted average position 1C2 will be close to both of the detected positions 1A2 and 1B2. In this case, the detection is deemed successful, and the detected position 1A2 by image recognition method A, which is close to the weighted average position 1C2, can be set as the recognized position on the image captured by the X-ray tube 13A and X-ray detector 12A. However, the weighted average position 1C2 itself may be used as the recognized position, or the detected positions 1A2 and 1B2 by the image recognition method with the highest reliability may be used as the recognized position.

[0080] Furthermore, as shown in Figure 8, the reliability w a And reliability w b Even if the difference in reliability between the two image recognition methods A and B is small, if the detected positions 1A3 and 1B3 of both methods are not within the range of the threshold distance ΔR from the weighted average position 1C3, recognition will fail.

[0081] In step S207, the recognition success / failure determination unit 150b can execute an output procedure for outputting a display signal to display the weighted average positions 1C1, 1C2, and 1C3, the detected positions 1A and 1B of the tracking target B in each process, the threshold distance, and the success / failure determination results 401, 402, and 403 on the display device 154. In this case, only the success / failure determination results 401, 402, and 403 may be displayed, or any of the images shown in Figs. 6 to 8 may be displayed.

[0082] The processing shown in Figure 2 above is performed on both the X-ray fluoroscopic image captured by the X-ray tube 13A-X-ray detector 12A pair and the X-ray fluoroscopic image captured by the X-ray tube 13B-X-ray detector 12B pair, and tracking is determined to be successful only if recognition is successful in both pairs.

[0083] In this case, a common perpendicular to two lines connecting the positions on the X-ray detectors 12A and 12B where the tracking target is imaged and the X-ray generator for fluoroscopy is calculated, and the midpoint of the common perpendicular is calculated as the position of the diaphragm.

[0084] Next, the relationship (offset vector) between the three-dimensional position of the diaphragm and the three-dimensional position of the target, which has been calculated in advance, is read, the coordinates of the target position are calculated, and the coordinates of the target position are transmitted together with information on successful recognition to the displacement amount calculation device 152. On the other hand, if recognition of the diaphragm fails in either pair, it is determined that tracking has failed, and information on the recognition failure is transmitted to the displacement amount calculation device 152.

[0085] The tracking success / failure determination unit 152a of the displacement amount calculation device 152 determines that tracking has failed when it receives information of recognition failure from the tumor position recognition device 150, or when the deviation between the predicted position of the tumor calculated by the tumor position prediction device 151 and the tumor position calculated by the tumor position recognition device 150 is greater than a set threshold.

[0086] When the tracking success / failure determination unit 152a determines that tracking is successful, the displacement amount calculation unit 152b of the displacement amount calculation device 152 calculates the amount of displacement from the position of the tracked object in the previous frame and outputs the calculated amount of displacement to the main control unit 16.

[0087] The X-ray therapy system 10 has a preset number of tracking failures, and if tracking fails consecutively for more than the preset number of times, the radiation therapy device stops irradiation. After the radiation therapy device stops, the operator can instruct the body surface position-tumor correlation model to be updated using body surface position data and tumor position data from previous successful tracking attempts.

[0088] Therefore, the recognition success / failure determination unit 150b of the tumor position recognition device 150 can appropriately exclude imaging data for which recognition failed from the "recognition success" condition in the image recognition process for recognizing landmark positions. As a result, a body surface position-tumor correlation model can be constructed using only high-quality tracking position data, making it possible to improve the accuracy of the body surface position-tumor correlation model. This can contribute to improving the reliability of moving object tracking.

[0089] Next, the effects of this embodiment will be described.

[0090] The program executed by the moving object tracking calculation device 15 that performs position calculation of the tracking object B in the above-mentioned embodiment causes the moving object tracking calculation device 15 to execute a processing procedure for performing a plurality of image recognition processing methods on an image showing the tracking object B, and a determination procedure for determining whether the position detection was successful or not based on the detected position of the tracking object B in each of the processing of the plurality of image recognition processing methods in the processing procedure and a weighted average position weighted by the reliability of each image recognition processing method.

[0091] Therefore, even under strict conditions, such as when X-ray fluoroscopic images are taken infrequently, the success or failure of position detection can be determined with high accuracy, thereby improving the reliability of moving object tracking.

[0092] Furthermore, since the reliability is set or calculated in the processing procedure, the reliability can be set more appropriately, thereby realizing tracking with higher accuracy.

[0093] Furthermore, in the processing procedure, the reliability is calculated based on the gradient information of the similarity scale distribution calculated by the image recognition processing method, thereby making it possible to obtain a more accurate reliability for each image recognition processing method, thereby achieving even more accurate tracking.

[0094] In addition, in the judgment procedure, the success or failure is determined based on whether or not the detected positions 1A, 1B by one or more of the image recognition processing methods are present within a threshold distance from the weighted average position, making it possible to determine the detected position with high accuracy as the recognized position.

[0095] Furthermore, in the judgment procedure, if a detected position using one image recognition processing method exists within a threshold range from the weighted average position, it is judged that the position detection has been successful, and the detected position is regarded as the position of the tracked object B on the image.In addition, in the judgment procedure, if detected positions using multiple image recognition processing methods exist within a threshold range from the weighted average position, it is judged that the position detection has been successful, and one of the detected position closest to the weighted average position, the weighted average position itself, or the detected position using the image recognition processing method with the highest reliability is regarded as the position of the tracked object B on the image, thereby making it possible to determine a more appropriate detected position as the recognized position.

[0096] Furthermore, by further executing an output procedure that outputs a display signal to display the weighted average position, the detected position of the tracking target B in each process, the threshold distance, and the success / failure determination results 401, 402, 403 on the display device 154, an operator such as a doctor can understand the tracking situation, thereby further improving the stability of treatment.

[0097] <Other> The present invention is not limited to the above-described embodiment, and various modifications and applications are possible. The above-described embodiment has been described in detail to clearly explain the present invention, and the present invention is not necessarily limited to having all of the described configurations.

[0098] For example, the present invention may be embodied in the following manner.

[0099] (1) A program to be executed by a computing system that calculates the position of a tracked object, the program causing the computing system to execute a processing procedure that performs a plurality of image recognition processing methods on an image showing the tracked object, and a judgment procedure that judges whether the position detection was successful or not based on the detected position of the tracked object in each of the processing of the plurality of image recognition processing methods in the processing procedure and a weighted average position weighted by the reliability of each of the image recognition processing methods.

[0100] (2) In the program according to (1), the processing step includes setting or calculating the reliability.

[0101] (3) In the program described in (2), the processing procedure calculates the reliability based on gradient information of the similarity scale distribution calculated by the image recognition processing method.

[0102] (4) In the program described in any one of (1) to (3), the judgment procedure judges success or failure based on whether the detected position in one or more of the image recognition processing methods is within a threshold distance from the weighted average position.

[0103] (5) In the program described in (4), in the judgment procedure, if the detected position by one of the image recognition processing methods is within a threshold range from the weighted average position, it is judged that the position detection is successful, and the detected position is regarded as the position of the tracked object on the image.

[0104] (6) In the program described in (4), in the judgment procedure, if there are multiple detected positions by the image recognition processing method within a threshold range from the weighted average position, it is judged that the position detection was successful, and one of the detected position closest to the weighted average position, the weighted average position itself, or the detected position by the image recognition processing method with the highest reliability is considered to be the position of the tracked object on the image.

[0105] (7) In the program described in any one of (4) to (6), an output procedure is further executed to output a display signal to display the weighted average position, the detected position of the tracked object in each process, the threshold distance, and the success / failure determination result on a display device. [Explanation of symbols]

[0106] 1A, 1A1, 1A2, 1A3, 1B, 1B1, 1B2, 1B3...Detection position 1C1,1C2,1C3…Weighted average position 10...X-ray therapy system (radiation therapy system) 12A, 12B...X-ray detector (radiation detector for fluoroscopy, fluoroscopy device, moving body tracking device) 13A, 13B...X-ray tube (radiation generator for fluoroscopy, fluoroscopy device, moving body tracking device) 14...Infrared camera 15...Moving object tracking calculation device (calculation system, moving object tracking device) 16...Main control unit 17...Communication equipment 18...Storage device 19...Input device 20...Fluoroscopic X-ray imaging device 110...Gantry (irradiation device) 111...Irradiation nozzle (irradiation device) 112...Bed (irradiation device) 150...Tumor location recognition device 150a...Tumor location recognition unit 150b...Recognition success / failure determination unit 151...Tumor location prediction device 152...Displacement calculation device 152a...Tracking success / failure determination unit 152b...Displacement amount calculation unit 153...Input device 154...Display device 171...Treatment planning device 172...Data Server 300...Recognition position on the captured image of the X-ray tube 13A and the X-ray detector 12A 301, 302, 303, 304...Candidate points for recognition positions on the image captured by the X-ray tube 13A and the X-ray detector 12A 401, 402, 403...Success / failure judgment result A...Patient B...Tracked object C...Infrared marker

Claims

1. A program to be executed by a calculation system that calculates the position of a tracked object, A processing procedure for performing a plurality of image recognition processes on an image showing a tracking target; a determination procedure for determining whether position detection has been successful or not based on the detected position of the tracked object in each of the plurality of image recognition processing methods in the processing procedure and a weighted average position weighted by the reliability of each of the image recognition processing methods. program.

2. 2. The program according to claim 1, In the processing procedure, the reliability is set or calculated. program.

3. 3. The program according to claim 2, In the processing procedure, the reliability is calculated based on gradient information of the similarity scale distribution calculated by the image recognition processing method. program.

4. 2. The program according to claim 1, In the determination step, success or failure is determined based on whether or not the detected position by one or more of the image recognition processing methods exists within a threshold distance from the weighted average position. program.

5. 5. The program according to claim 4, In the determination step, if the detected position by one of the image recognition methods is within a threshold range from the weighted average position, it is determined that the position detection has been successful, and the detected position is regarded as the position of the tracked object on the image. program.

6. 5. The program according to claim 4, In the determination step, if the detected positions by the image recognition processing method are present within a threshold range from the weighted average position, it is determined that the position detection has been successful, and one of the detected position closest to the weighted average position, the weighted average position itself, and the detected position by the image recognition processing method with the highest reliability is regarded as the position of the tracked object on the image. program.

7. 5. The program according to claim 4, An output step is further executed to output a display signal to display the weighted average position, the detected position of the tracked object in each process, the threshold distance, and the success / failure determination result on a display device. program.

8. one or more sets of fluoroscopy devices each including a fluoroscopy radiation detector and a fluoroscopy radiation generator, for capturing a fluoroscopic image of the tracked object; a calculation system that executes the program according to any one of claims 1 to 7 and determines the position of the tracked object. Motion tracking device.

9. The moving object tracking device according to claim 8 ; an irradiation device that emits therapeutic radiation; Radiation therapy system.

10. A method for detecting the position of a tracked object, comprising: a processing step of performing a plurality of image recognition processes on an image showing the tracking target; a determining step of determining whether position detection has been successful or not based on the detected position of the tracked object in each of the processes of the plurality of image recognition processes in the processing step and a weighted average position weighted by the reliability of each of the image recognition processes. A method for detecting the location of a tracked object.

Images