Image processing device, image processing method, and image processing program

Abstract

To provide an image processing device, an image processing method and an image processing program which are capable of efficiently detecting a feature point from a medical image.SOLUTION: An image processing device sets a first point on a medical image including a detection target, and detects a feature point of the detection target from the medical image by performing, a predetermined number of times, a binary search based on the first point and a second point different from the first point on the medical image.SELECTED DRAWING: Figure 7

Description

[Technical Field]

[0001] The present disclosure relates to an image processing device, an image processing method, and an image processing program. [Background technology]

[0002] Patent document 1 discloses a technology for acquiring a three-dimensional or multi-layer two-dimensional positioning image of an organ, locating a segment in which the organ is located in the layer direction of multiple slices contained in the positioning image based on the acquired positioning image, and performing image segmentation processing on the positioning image within the segment in which the organ is located. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2023-029225 Summary of the Invention [Problem to be solved by the invention]

[0004] Feature points, such as organ boundaries, are detected from heat maps that represent the extent of organs in medical images. However, performing feature point detection on images with a relatively large amount of data, such as high-resolution medical images, increases the amount of computation required. Therefore, it would be desirable to be able to efficiently detect feature points from medical images.

[0005] The present disclosure has been made in consideration of the above circumstances, and aims to provide an image processing device, an image processing method, and an image processing program that can efficiently detect feature points from medical images. [Means for solving the problem]

[0006] The image processing device of the present disclosure is an image processing device having at least one processor, which sets a first point on a medical image including a detection target, and detects feature points of the detection target from the medical image by performing a predetermined number of binary searches based on the first point and a second point on the medical image that is different from the first point.

[0007] In addition, the image processing method disclosed herein involves setting a first point on a medical image containing the detection target, and performing a binary search based on the first point and a second point on the medical image that is different from the first point a predetermined number of times, thereby causing a processor included in the image processing device to perform a process of detecting feature points of the detection target from the medical image.

[0008] In addition, the image processing program disclosed herein causes a processor provided in an image processing device to execute a process of detecting feature points of the detection target from the medical image by setting a first point on a medical image including the detection target and performing a binary search based on the first point and a second point on the medical image that is different from the first point a predetermined number of times. [Effects of the Invention]

[0009] According to the present disclosure, feature points can be efficiently detected from medical images. [Brief explanation of the drawings]

[0010] [Figure 1] FIG. 2 is a block diagram illustrating an example of a hardware configuration of the image processing apparatus. [Figure 2] FIG. 2 is a diagram for explaining a first medical image and a second medical image. [Figure 3] FIG. 10 is a diagram for explaining a detection model. [Figure 4] FIG. 2 is a block diagram illustrating an example of a functional configuration of the image processing apparatus. [Figure 5] FIG. 10 is a diagram for explaining a process of setting a second point based on a first point. [Figure 6] FIG. 10 is a diagram illustrating a prediction model. [Figure 7]FIG. 10 is a diagram for explaining a binary search process. [Figure 8] 10A and 10B are diagrams for explaining a method for determining feature points according to a modified example. [Figure 9] 10 is a flowchart illustrating an example of a feature point detection process. [Figure 10] FIG. 10 is a diagram for explaining a binary search process according to a modified example. DETAILED DESCRIPTION OF THE INVENTION

[0011] Hereinafter, examples of embodiments for carrying out the technology of the present disclosure will be described in detail with reference to the drawings.

[0012] First, the hardware configuration of an image processing device 10 according to this embodiment will be described with reference to Fig. 1. As shown in Fig. 1, the image processing device 10 includes a CPU (Central Processing Unit) 20, a memory 21, a display 24, an input device 25, and a network I / F (Interface) 26. Examples of the image processing device 10 include a computer such as a personal computer or a server computer.

[0013] The CPU 20 realizes various functions by executing programs stored in the storage unit 22. The CPU 20 is an example of a processor according to the disclosed technology.

[0014] The memory 21 includes a storage unit 22 and a RAM (Random Access Memory) 23. The RAM 23 is a memory for primary storage, and is, for example, a RAM such as an SRAM (Static Random Access Memory) or a DRAM (Dynamic Random Access Memory).

[0015] The storage unit 22 is a non-volatile memory, and is realized by, for example, at least one of a hard disk drive (HDD), a solid state drive (SSD), a flash memory, etc. The storage unit 22 serving as a storage medium stores an image processing program 30. The CPU 20 reads the image processing program 30 from the storage unit 22, expands it in the RAM 23, and executes the expanded image processing program 30.

[0016] The display 24 is a device that displays various screens, and is, for example, a liquid crystal display or an EL (Electro Luminescence) display. The input device 25 is a device for a user to input, and is, for example, at least one of a keyboard, a mouse, a microphone for voice input, a touchpad for proximity input including contact, and a camera for gesture input. The network I / F 26 is an interface for connecting to a network. The bus 27 connects the CPU 20, the memory 21, the display 24, the input device 25, and the network I / F 26.

[0017] The storage unit 22 also stores a medical image 32, a medical image 34, a detection model 36, and a prediction model 38. The medical image 32 is a two-dimensional medical image including a detection target. In this embodiment, the medical image 32 is a two-dimensional radiological image obtained by irradiating a subject with radiation such as X-rays from the front to the back of the subject, i.e., in the so-called AP (Anterior Posterior) direction. A more specific example of the medical image 32 is a scout image captured by a CT (Computed Tomography) device. The medical image 32 is an example of a first medical image according to the disclosed technology.

[0018] As an example, as shown in Fig. 2, medical image 34 includes the detection target and is an image of a partial range of medical image 32. In other words, medical image 34 is also a two-dimensional medical image that includes the detection target, and is an image that captures a narrower range than medical image 32. For example, medical image 34 is generated by cropping a partial range from medical image 32. Medical image 34 is an example of a second medical image related to the disclosed technology.

[0019] The detection model 36 is a trained model obtained in advance by machine learning. As shown in FIG. 3 as an example, the detection model 36 is a model that receives a two-dimensional medical image including a detection target as input and outputs a detection result of the detection target for the input medical image. In this embodiment, the detection model 36 outputs a heat map HM representing the range of the detection target as the detection result of the detection target. In this embodiment, the heat map HM is a region of the detection target that is filled in with a predetermined color such as red. In the heat map HM, the higher the certainty that the region is the detection target, the higher the density of the fill.

[0020] The prediction model 38 is a trained model obtained in advance by machine learning. Details of the prediction model 38 will be described later.

[0021] The image processing device 10 according to this embodiment has a function of detecting feature points of a detection target based on a medical image 32 and a medical image 34. In this embodiment, the liver is used as the detection target, and the upper end of the liver is used as the feature point of the detection target. Feature points are, for example, landmark points used to define the range of the organ to be detected. The organ range is used to define the range of 3D images captured by a CT device or the range of image processing. The detection target may be an organ other than the liver, or a region of interest other than an organ, such as a lesion region. The organs referred to here also include bones and blood vessels other than organs located in the thoracic and abdominal cavities, such as the lungs and liver.

[0022] The functional configuration of the image processing device 10 will be described with reference to Fig. 4. As shown in Fig. 4, the image processing device 10 includes an acquisition unit 40, a setting unit 42, and a detection unit 44. The CPU 20 executes the image processing program 30 to function as the acquisition unit 40, the setting unit 42, and the detection unit 44.

[0023] The acquisition unit 40 acquires the medical image 32 and the medical image 34 from the storage unit 22. The setting unit 42 sets a first point on the medical image 34 based on the detection result of the detection target on the medical image 32 acquired by the acquisition unit 40.

[0024] Specifically, first, the setting unit 42 inputs the medical image 32 to the detection model 36. As a result, the detection model 36 outputs a heat map HM as a detection result of the detection target for the medical image 32. Next, the setting unit 42 sets a first point for the medical image 32 based on the heat map HM. In this embodiment, the setting unit 42 sets a point that is set as a point that is easy to detect from the medical image 32 as the first point.

[0025] As described above, the heat map HM is filled with a higher density in areas with a higher degree of certainty that the detection target is a target. When the detection target is a liver, the closer to the center of the liver, the easier it is to detect, so the degree of certainty often increases the closer to the center of the heat map HM. Therefore, the setting unit 42 sets, as a first point, the center point of the range of the detection target represented by the heat map HM, which is the detection result of the detection target for the medical image 32.

[0026] Then, the setting unit 42 sets a first point at a position on the medical image 34 that corresponds to the first point on the medical image 32. Specifically, the setting unit 42 aligns the medical images 32 and 34, and sets the first point at a position on the medical image 34 that corresponds to the first point on the medical image 32.

[0027] Furthermore, the setting unit 42 sets the second point in the medical image 34 at a point between the first point and the second point and the feature point, based on the anatomical positions of the detection target and the feature point of the detection target, and the relative positional relationship between the first point and the feature point. As shown in FIG. 5 , a first point P1 based on the heat map HM is set within the liver LV of the detection target, and the unknown feature point T of the detection target is a point at the upper end of the liver LV. In this case, based on the anatomical positions of the liver LV and the feature point T, it is considered that the feature point T exists at least above the first point P1. Therefore, in this embodiment, the setting unit 42 sets the second point P2 at a point at the upper end of the medical image 34 on the opposite side of the first point P1 with respect to the feature point T, based on the relative positional relationship between the first point P1 and the feature point T in the medical image 34.

[0028] The detection unit 44 detects the feature point T from the medical image 34 by performing a binary search a predetermined number of times based on the first point P1 and the second point P2 set on the medical image 34 by the setting unit 42. A specific example of the binary search performed by the detection unit 44 will be described below.

[0029] The detection unit 44 performs a first prediction process that predicts the relative positional relationship between the first point P1 or the second point P2 and the feature point T by binary classification. In this embodiment, the upper end of the liver LV is used as the feature point T, and therefore, a case will be described in which the vertical positional relationship of the medical image 34 is used as the relative positional relationship between the first point P1 or the second point P2 and the feature point T. Note that, for example, when the left end or right end of the liver LV is used as the feature point T, the horizontal positional relationship of the medical image 34 may be used as the relative positional relationship between the first point P1 or the second point P2 and the feature point T.

[0030] The detection unit 44 uses a prediction model 38 for the first prediction process. As shown in FIG. 6 as an example, the prediction model 38 is a trained model that receives position information of a point on a medical image 34 as input, predicts the position of an unknown feature point T on the medical image 34, and outputs a binary value representing whether the point represented by the input position information is above the feature point T as a prediction result. In this embodiment, an example will be described in which coordinates of an axis along the up-down direction with the upper left point or center point of the medical image 34 as the origin (hereinafter referred to as "y coordinates") are used as the position information of the point on the medical image 34. Furthermore, the prediction model 38 outputs a binary value of True or False. Specifically, the prediction model 38 outputs True if it predicts that the point represented by the input position information is above the feature point T in the medical image 34, and outputs False if it predicts that the point represented by the input position information is not above the feature point T.

[0031] Then, the detection unit 44 performs a first movement process to move the first point P1 or the second point P2 to a temporary feature point between the first point P1 and the second point P2 based on the prediction result of the first prediction process.

[0032] As shown in STEP 1 of FIG. 7, for example, the detection unit 44 detects the y-coordinate y of the first point P1. p1 is input to the prediction model 38. Note that y t represents the y coordinate of feature point T, and y pnrepresents the y-coordinate of point Pn (n = 1, 2, . . . ). In this case, the prediction model 38 outputs False as the prediction result. As a result, the detection unit 44 determines that feature point T is above the first point P1 and moves the first point P1 to a provisional feature point between the first point P1 and the second point P2. As shown in STEP 2 of FIG. 7 , in this embodiment, the detection unit 44 moves the first point P1 to a provisional feature point P3 that bisects the distance between the first point P1 and the second point P2. Note that the detection unit 44 may determine as the provisional feature point P3 any of two points that divide the distance between the first point P1 and the second point P2 into thirds, or any of three points that divide the distance between the first point P1 and the second point P2 into fourths. The binary search in this embodiment refers to dividing the distance between two points in the vertical direction and is not limited to bisecting the distance.

[0033] The detector 44 also detects the y-coordinate y of the second point P2. p2 may be input to the prediction model 38. In this case, the prediction model 38 outputs True as the prediction result. In this case, the detection unit 44 may determine that the feature point T is below the second point P2, and move the second point P2 to a tentative feature point P3 between the first point P1 and the second point P2.

[0034] Next, the detection unit 44 performs a second prediction process that predicts the relative positional relationship between the tentative feature point P3 and the feature point T by binary classification. The detection unit 44 uses the prediction model 38 for the second prediction process, as in the first prediction process. Then, based on the prediction result of the second prediction process, the detection unit 44 performs a second movement process that moves one of two points, the unmoved point and the tentative feature point P3, from the immediately preceding movement process based on the first point P1 or the second point P2, to a new tentative feature point. When repeating the second prediction process and the second movement process a predetermined number of times, the detection unit 44 sets a new tentative feature point between the tentative feature point P3 and one of two points, the unmoved point and the tentative feature point P3, from the immediately preceding second movement process. In the second movement process, as in the first movement process, the detection unit 44 sets the new tentative feature point as a point that bisects the distance between the two points.

[0035] As shown in STEP 2 of FIG. 7, the detection unit 44 detects the y-coordinate y p3 is input to the prediction model 38. The prediction model 38 outputs True as the prediction result. As a result, the detection unit 44 determines that the feature point T is below the tentative feature point P3, and moves the first point P1 to a new tentative feature point P4 between the first point P1 and the tentative feature point P3.

[0036] 7, the detection unit 44 determines that feature point T is below tentative feature point P4, and sets a new tentative feature point between the first point P1 and tentative feature point P4. This new tentative feature point is used in the next step, STEP 4 (not shown). In this way, the detection unit 44 repeats the process of setting a new tentative feature point between the tentative feature point and one of the two points from the previous step, based on the predicted hierarchical relationship between the tentative feature point and feature point T, a predetermined number of times.

[0037] The number of times that the detection unit 44 repeats the second prediction process and the second movement process may be set in advance as a fixed value. Furthermore, for example, the detection unit 44 may determine that convergence has occurred when the amount of movement of the tentative feature points becomes equal to or less than a threshold, and may terminate the second prediction process and the second movement process. In this case, the number of times that the detection unit 44 repeats the second prediction process and the second movement process is not fixed.

[0038] The detection unit 44 detects the final tentative feature point obtained by repeating the second prediction process and the second movement process a predetermined number of times as the feature point T to be detected.

[0039] The detection unit 44 may determine the feature point T to be detected based on the history of the tentative feature points. As an example, as shown in FIG. 8, the detection unit 44 may detect, as the feature point T, the position with the highest probability in a probability distribution obtained by superimposing the individual probability distributions of the history of the tentative feature points. In the example of FIG. 8, points Z0 to Z3 represent the history of the tentative feature points, curves C0 to C3 represent the probability distribution corresponding to points Z0 to Z3, and curve C4 represents the probability distribution obtained by superimposing the probability distributions represented by curves C0 to C3. In this example, the detection unit 44 determines, as the feature point T, the position corresponding to the vertex of curve C4 in the medical image 34.

[0040] Next, the operation of the image processing device 10 will be described with reference to Fig. 9. The CPU 20 executes the image processing program 30, thereby executing the feature point detection process shown in Fig. 9. The feature point detection process shown in Fig. 9 is executed, for example, when a command to start execution is input by the user.

[0041] 9, the acquisition unit 40 acquires the medical image 32 and the medical image 34 from the storage unit 22. In step S12, the setting unit 42 sets a first point P1 on the medical image 34 based on the detection result of the detection target on the medical image 32 acquired in step S10, as described above.

[0042] In step S14, as described above, the setting unit 42 sets the second point P2 to a point on the medical image 34 where the first point P1 and the second point P1 sandwich the feature point T, based on the anatomical positions of the detection target and the feature point T of the detection target, and the relative positional relationship between the first point P1 and the feature point.

[0043] In step S16, as described above, the detection unit 44 performs a first prediction process of predicting, by binary classification, the relative positional relationship between the first point P1 or the second point P2 and the feature point T. In step S18, as described above, the detection unit 44 performs a first movement process of moving the first point P1 or the second point P2 to a temporary feature point P3 between the first point P1 and the second point P2 based on the prediction result of the first prediction process.

[0044] In step S20, the detection unit 44 performs the second prediction process, as described above, to predict, by binary classification, the relative positional relationship between the tentative feature point P3 and the feature point T. In step S22, the detection unit 44 performs the second movement process, as described above, to move one of two points, the unmoved point in the immediately preceding movement process based on the first point P1 or the second point P2 and the tentative feature point P3, to a new tentative feature point based on the prediction result of the second prediction process.

[0045] In step S24, the detection unit 44 determines whether a predetermined termination condition is satisfied. An example of the termination condition is when steps S20 and S22 have been repeatedly executed a predetermined number of times. Another example of the termination condition is when the amount of movement of the provisional feature point due to the second movement process executed immediately before in step S22 becomes equal to or less than a threshold. If the determination in step S24 is negative, the process returns to step S20; if the determination is positive, the process proceeds to step S26.

[0046] In step S26, the detection unit 44 detects the final provisional feature points obtained by repeating the second prediction process in step S20 and the second movement process in step S22 a predetermined number of times as the feature points T to be detected. When the process of step S26 ends, the feature point detection process ends. The feature points T detected by the above processes are used, for example, to determine the imaging range of a three-dimensional medical image.

[0047] As described above, according to this embodiment, feature points can be efficiently detected from medical images.

[0048] In the above embodiment, the medical image 34 is an image of a partial range of the medical image 32, but the disclosed technology is not limited to this. For example, the medical image 34 may be an image capturing the same range as the medical image 32 but with a higher resolution than the medical image 32. In this case, for example, the CPU 20 may use an image obtained by a medical imaging device, such as a scout image, as the medical image 34, and an image with a lower resolution of the medical image 34 as the medical image 32. In this case, the amount of calculation required for the detection process of the detection target for the medical image 32 can be reduced. Alternatively, the medical image 32 may be an image with a lower resolution of the scout image, and the medical image 34 may be an image of a partial range including the detection target of the scout image.

[0049] Furthermore, in the above embodiment, a case has been described in which the center point of the range of the detection target represented by the detection result of the detection target in the medical image 32 is used as the first point, but the disclosed technology is not limited to this aspect. For example, a point on the edge of the range of the detection target represented by the detection result of the detection target in the medical image 32 may be used as the first point. In this case, an organ that is easier to detect at the edge than at the center may be used as the detection target. An example of an organ that is easier to detect at the edge than at the center is an organ that has a certain level of contrast with surrounding organs.

[0050] Alternatively, the first point may be a point whose brightness is equal to or greater than a threshold within the range of the detection target represented by the detection result of the detection target in the medical image 32. A point whose brightness is equal to or greater than a threshold is, for example, a point corresponding to a lesion, a cyst, a calcification point, or the like.

[0051] Although the above embodiment describes a case where a one-way positional relationship in the vertical direction is used in the binary search, the disclosed technology is not limited to this aspect. A two-way positional relationship in the vertical and horizontal directions may also be used in the binary search. In this case, as shown in FIG. 10 as an example, the setting unit 42 sets a point at the upper left corner of the medical image 34 as the second point P2 based on the anatomical positions of the detection target and the feature point T of the detection target, and the relative positional relationship between the first point P1 and the feature point T. Then, in this case, the detection unit 44 moves the first point P1 or the second point P2 to a provisional feature point P3 between the first point P1 and the second point P2 based on the vertical and horizontal positional relationship between the first point P1 or the second point P2 and the feature point T. Furthermore, in this case, the detection unit 44 updates the provisional feature point P3 to a provisional feature point P4 based on the vertical and horizontal positional relationship between the provisional feature point P3 and the feature point T. In this way, the detection unit 44 can narrow down the range on the two-dimensional plane where the feature point T exists. Note that the two numbers in parentheses in Fig. 10 represent the x-coordinate and y-coordinate of the nearby point.

[0052] Furthermore, in the above embodiment, the following various processors can be used as the hardware structure of a processing unit that executes various processes, such as each functional unit of the image processing device 10. As described above, the various processors include a CPU, which is a general-purpose processor that executes software (programs) and functions as various processing units, as well as dedicated electrical circuits that are processors having a circuit configuration specifically designed to execute specific processes, such as a programmable logic device (PLD), a processor whose circuit configuration can be changed after manufacture, such as an FPGA (Field Programmable Gate Array), and an ASIC (Application Specific Integrated Circuit).

[0053] A single processing unit may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, multiple processing units may be configured with a single processor.

[0054] Examples of configuring multiple processing units with a single processor include, first, a form in which one processor is configured with a combination of one or more CPUs and software, and this processor functions as multiple processing units, as typified by computers such as client and server. Second, a form in which a processor is used to realize the functions of the entire system including multiple processing units with a single IC (Integrated Circuit) chip, as typified by systems on chips (SoCs). In this way, various processing units are configured using one or more of the above-mentioned various processors as a hardware structure.

[0055] Furthermore, more specifically, the hardware structure of these various processors can be an electric circuit that combines circuit elements such as semiconductor elements.

[0056] In the above embodiment, the image processing program 30 is pre-stored (installed) in the storage unit 22, but the disclosed technology is not limited to this. The image processing program 30 may be provided in a form recorded on a recording medium such as a CD-ROM (Compact Disc Read Only Memory), a DVD-ROM (Digital Versatile Disc Read Only Memory), or a USB (Universal Serial Bus) memory. The image processing program 30 may also be downloaded from an external device via a network.

[0057] The following additional notes are provided regarding the above-described embodiments. (Appendix 1) An image processing device comprising at least one processor, The processor: A first point is set on a medical image including the detection target; A binary search is performed a predetermined number of times based on the first point and a second point on the medical image that is different from the first point, thereby detecting the feature point of the detection target from the medical image. Image processing device.

[0058] (Appendix 2) The processor: Based on a detection result of the detection target for a first medical image including the detection target, the first point is set for a second medical image which includes the detection target and is an image with a higher resolution than the first medical image or an image of a partial range of the first medical image; Detecting the feature points from the second medical image 2. The image processing device according to claim 1.

[0059] (Appendix 3) The detection result of the detection target is a heat map showing the range of the detection target. 3. The image processing device according to claim 2.

[0060] (Appendix 4) The first point is a point that is set as a point that is easy to detect from the medical image. 4. An image processing device according to any one of claims 1 to 3.

[0061] (Appendix 5) The first point is the center point of the range of the detection object represented by the detection result of the detection object in the medical image. 5. The image processing device according to claim 4.

[0062] (Appendix 6) The first point is a point on the edge of the range of the detection object represented by the detection result of the detection object in the medical image. 5. The image processing device according to claim 4.

[0063] (Appendix 7) The detection target is an organ that has a certain level of contrast with the surrounding organs. 7. The image processing device according to claim 6.

[0064] (Appendix 8) The first point is a point whose brightness within the range of the detection target represented by the detection result of the detection target in the medical image is equal to or greater than a threshold value. 5. The image processing device according to claim 4.

[0065] (Appendix 9) The processor: In the medical image, the second point is set to a point on either side of the feature point between the first point and the second point based on the anatomical positions of the detection target and the feature point and the relative positional relationship between the first point and the feature point. 9. An image processing device according to any one of claims 1 to 8.

[0066] (Appendix 10) The processor: performing a first prediction process for predicting a relative positional relationship between the first point or the second point and the feature point by binary classification; performing a first movement process of moving the first point or the second point to a tentative feature point between the first point and the second point based on a prediction result of the first prediction process; performing a second prediction process of predicting a relative positional relationship between the tentative feature points and the feature points by binary classification; performing a second movement process based on a prediction result of the second prediction process to move one of two points, the unmoved point obtained by the immediately preceding movement process based on the first point or the second point, and the tentative feature point, to a new tentative feature point; When the second prediction process and the second movement process are repeated a predetermined number of times, the new tentative feature point is set between the tentative feature point and one of the two points in the immediately preceding second movement process. 10. An image processing device according to any one of claims 1 to 9.

[0067] (Appendix 11) The processor: The second prediction process and the second movement process are repeated a predetermined number of times to finally obtain the provisional feature points, which are then detected as the feature points of the detection target. 11. The image processing device of claim 10.

[0068] (Appendix 12) The processor: determining feature points of the detection target based on a history of the tentative feature points; 11. The image processing device of claim 10.

[0069] (Appendix 13) A first point is set on a medical image including the detection target; A binary search is performed a predetermined number of times based on the first point and a second point on the medical image that is different from the first point, thereby detecting the feature point of the detection target from the medical image. An image processing method in which processing is performed by a processor provided in an image processing device.

[0070] (Appendix 14) A first point is set on a medical image including the detection target; A binary search is performed a predetermined number of times based on the first point and a second point on the medical image that is different from the first point, thereby detecting the feature point of the detection target from the medical image. An image processing program for causing a processor included in the image processing device to execute the processing. [Explanation of symbols]

[0071] 10 Image processing device 20 CPU 21 Memory 22 Memory section 23 RAM 24 displays 25 Input Devices 26 Network I / F 27 Bus 30 Image Processing Programs 32, 34 Medical Imaging 36 Detection Models 38 Predictive Models 40 Acquisition Department 42 Setting section 44 Detector HM Heatmap LV liver P1 First point P2 Second point P3, P4 provisional feature points T feature points

Claims

1. An image processing device comprising at least one processor, The processor: setting a first point on a medical image including the detection target; A binary search is performed a predetermined number of times based on the first point and a second point on the medical image that is different from the first point, thereby detecting the feature point of the detection target from the medical image. Image processing device.

2. The processor: based on a detection result of the detection target for a first medical image including the detection target, setting the first point for a second medical image which includes the detection target and is an image with a higher resolution than the first medical image or an image of a partial range of the first medical image; Detecting the feature points from the second medical image The image processing device according to claim 1 .

3. The detection result of the detection target is a heat map showing the range of the detection target. The image processing device according to claim 1 .

4. The first point is a point that is set as a point that is easy to detect from the medical image. The image processing device according to any one of claims 1 to 3.

5. The first point is the center point of the range of the detection object represented by the detection result of the detection object in the medical image. The image processing device according to claim 4 .

6. The first point is a point on an edge of a range of the detection target represented by a detection result of the detection target in the medical image. The image processing device according to claim 4 .

7. The detection target is an organ that has a certain level of contrast with the surrounding organs. The image processing device according to claim 6 .

8. The first point is a point within the range of the detection target represented by the detection result of the detection target in the medical image, the brightness of which is equal to or greater than a threshold value. The image processing device according to claim 4 .

9. The processor: In the medical image, the second point is set to a point on either side of the feature point between the first point and the second point based on the anatomical positions of the detection target and the feature point and the relative positional relationship between the first point and the feature point. The image processing device according to any one of claims 1 to 3.

10. The processor: performing a first prediction process of predicting a relative positional relationship between the first point or the second point and the feature point by binary classification; performing a first movement process of moving the first point or the second point to a tentative feature point between the first point and the second point based on a prediction result of the first prediction process; performing a second prediction process for predicting a relative positional relationship between the tentative feature points and the feature points by binary classification; performing a second movement process based on a prediction result of the second prediction process to move one of two points, the unmoved point obtained by the immediately preceding movement process based on the first point or the second point, and the tentative feature point, to a new tentative feature point; When the second prediction process and the second movement process are repeated a predetermined number of times, the new tentative feature point is set between the tentative feature point and one of the two points in the immediately preceding second movement process. The image processing device according to any one of claims 1 to 3.

11. The processor: The second prediction process and the second movement process are repeated a predetermined number of times to detect the final provisional feature points as the feature points of the detection target. The image processing device according to claim 10.

12. The processor: determining feature points of the detection target based on a history of the tentative feature points; The image processing device according to claim 10.

13. setting a first point on a medical image including the detection target; A binary search is performed a predetermined number of times based on the first point and a second point on the medical image that is different from the first point, thereby detecting the feature point of the detection target from the medical image. An image processing method in which processing is performed by a processor provided in an image processing device.

14. setting a first point on a medical image including the detection target; A binary search is performed a predetermined number of times based on the first point and a second point on the medical image that is different from the first point, thereby detecting the feature point of the detection target from the medical image. An image processing program for causing a processor included in the image processing device to execute the processing.

Images