JP7166870B2 - Image processing device and medical imaging device - Google Patents

Description

TECHNICAL FIELD Embodiments of the present invention relate to an image processing apparatus and a medical imaging apparatus.

Conventionally, there is a medical imaging apparatus that evaluates the state of adhesion between the parietal pleura of the lung and the visceral pleura from a moving image of the subject's breathing.

JP 2016-67832 A

A problem to be solved by the present invention is to provide an image processing apparatus and a medical imaging apparatus that can accurately evaluate the adhesion state.

An image processing apparatus according to an embodiment includes an acquisition unit and a first calculation unit. The acquiring unit acquires a plurality of images captured including a target region of the subject in a plurality of time phases. A first calculation unit, based on movement information between a plurality of temporal images in a plurality of pixels of an image arranged in a predetermined direction across a boundary between a first region and a second region in the image, From the classification information for classifying each of the plurality of pixels into a first class related to the first region or a second class related to a second region adjacent to the first region in a predetermined direction, the first region and the second region of the subject corresponding to the second region.

FIG. 1 is a diagram showing an example of the configuration of an X-ray CT apparatus according to the first embodiment. FIG. 2 is a flowchart illustrating an example flow of evaluation processing according to the first embodiment. FIG. 3 is a diagram for explaining an example of evaluation processing according to the first embodiment. FIG. 4 is a diagram for explaining an example of evaluation processing according to the first embodiment. FIG. 5 is a diagram for explaining an example of evaluation processing according to the first embodiment. FIG. 6 is a diagram for explaining an example of cut lines according to the first embodiment. FIG. 7 is a diagram for explaining an example of cut lines according to the first embodiment. FIG. 8 is a diagram for explaining an example of evaluation processing according to the first embodiment. FIG. 9 is a diagram for explaining an example of evaluation processing according to the first embodiment. FIG. 10 is a diagram for explaining an example of evaluation processing according to the first embodiment. FIG. 11 is a diagram for explaining an example of evaluation processing according to the first embodiment. FIG. 12 is a diagram for explaining an example of evaluation processing according to the first embodiment. FIG. 13 is a diagram for explaining an example of evaluation processing according to the first embodiment. FIG. 14 is a diagram illustrating an example of an image represented by two-dimensional image data for display according to the first embodiment. FIG. 15 is a diagram for explaining an example of processing executed by the X-ray CT apparatus according to the second modification. FIG. 16 is a diagram for explaining an example of processing executed by the X-ray CT apparatus according to the second modification. FIG. 17 is a diagram showing an example of the configuration of a system including an image processing device according to the second embodiment.

An image processing apparatus and a medical imaging apparatus according to embodiments will be described below with reference to the drawings. In addition, the contents described in one embodiment are in principle similarly applied to other embodiments.

The medical image capturing apparatus is a general term for medical image diagnostic apparatuses that capture an image of a subject and generate a medical image. For example, medical imaging devices include X-ray CT devices. In the following embodiments, a case where the technology disclosed herein is applied to an X-ray CT apparatus will be described. It is also applicable to a PET (Positron Emission Tomography) device, a SPECT (Single Photon Emission Computed Tomography) device, etc.).

(First embodiment)
FIG. 1 is a diagram showing an example of the configuration of an X-ray CT apparatus 1 according to the first embodiment. As shown in FIG. 1, the X-ray CT apparatus 1 according to the first embodiment has a gantry device 10, a bed device 20, and a console device 30. As shown in FIG. The gantry device 10, the bed device 20, and the console device 30 are communicably connected to each other.

In this embodiment, the rotation axis of the rotating frame 13 or the longitudinal direction of the top board 23 of the bed device 20 in the non-tilt state is defined as the "Z-axis direction". Further, an axial direction perpendicular to the Z-axis direction and horizontal to the floor surface is defined as an “X-axis direction”. Further, the axial direction perpendicular to the Z-axis direction and perpendicular to the floor surface is defined as the “Y-axis direction”.

The gantry device 10 is a device that irradiates an object (patient) P with X-rays, detects the X-rays that have passed through the object P, and outputs the detected X-rays to the console device 30 . The gantry device 10 includes an X-ray tube 11, an X-ray detector 12, a rotating frame 13, a control device 14, a wedge 15, a collimator 16, a DAS (Data Acquisition System) 17, and an X-ray high voltage device. 18.

The X-ray tube 11 is a vacuum tube that emits thermal electrons from a cathode (filament) toward an anode (target) by applying a high voltage from an X-ray high voltage device 18 . The X-ray tube 11 generates X-rays by colliding thermal electrons with an anode.

Wedge 15 is a filter for adjusting the dose of X-rays emitted from X-ray tube 11 . Specifically, the wedge 15 transmits the X-rays emitted from the X-ray tube 11 so that the distribution of the X-rays emitted from the X-ray tube 11 to the subject P becomes a predetermined distribution. is a filter that attenuates For example, the wedge 15 is a filter made of aluminum so as to have a predetermined target angle and a predetermined thickness. The wedge 15 is also called a wedge filter or a bow-tie filter.

The collimator 16 is a lead plate or the like for narrowing down the irradiation range of the X-rays transmitted through the wedge 15 . The collimator 16 is formed in a slit shape by combining a plurality of lead plates or the like.

The X-ray detector 12 detects X-rays emitted from the X-ray tube 11 and passing through the subject P, and outputs an electrical signal corresponding to the X-ray dose to the DAS 17 . The X-ray detector 12 has, for example, a plurality of X-ray detection element arrays in which a plurality of X-ray detection elements are arranged in the channel direction along one circular arc centering on the focal point of the X-ray tube 11 . The X-ray detector 12 has, for example, a structure in which a plurality of X-ray detection element rows each having a plurality of X-ray detection elements arranged in the channel direction are arranged in the slice direction (also called row direction or row direction).

Also, the X-ray detector 12 is, for example, an indirect conversion type detector having a grid, a scintillator array, and a photosensor array. The scintillator array has a plurality of scintillators, and the scintillator has a scintillator crystal that outputs a photon amount of light corresponding to the amount of incident X-rays. The grid has an X-ray shielding plate arranged on the surface of the scintillator array on the X-ray incident side and having a function of absorbing scattered X-rays. The photosensor array has a function of outputting an electric signal according to the amount of light from the scintillator, and has photosensors such as photomultiplier tubes (PMTs), for example. The X-ray detector 12 may be a direct conversion type detector having a semiconductor element that converts incident X-rays into electrical signals.

The X-ray high-voltage device 18 has an electric circuit such as a transformer and a rectifier, and has a high-voltage generator function to generate a high voltage to be applied to the X-ray tube 11. and an X-ray control device for controlling an output voltage according to the X-ray output. The high voltage generator may be of a transformer type or an inverter type. Note that the X-ray high-voltage device 18 may be provided on a rotating frame 13 to be described later, or may be provided on a fixed frame (not shown) side of the gantry device 10 . Note that the fixed frame is a frame that rotatably supports the rotating frame 13 .

The DAS 17 has an amplifier that amplifies the electrical signal output from each X-ray detection element of the X-ray detector 12, and an A/D converter that converts the electrical signal into a digital signal. to generate Detection data generated by the DAS 17 is transferred to the console device 30 .

The rotating frame 13 is an annular frame that supports the X-ray tube 11 and the X-ray detector 12 so as to face each other and rotates the X-ray tube 11 and the X-ray detector 12 by means of a control device 14, which will be described later. In addition to the X-ray tube 11 and the X-ray detector 12, the rotating frame 13 further includes an X-ray high-voltage device 18 and a DAS 17 to support them. Detected data generated by the DAS 17 is transmitted by optical communication from a transmitter having a light emitting diode (LED) mounted on the rotating frame 13 to a photodiode mounted on a non-rotating portion (e.g., stationary frame) of the gantry 10. and transferred to the console device 30 . The method of transmitting the detection data from the rotating frame 13 to the non-rotating portion of the gantry 10 is not limited to the optical communication described above, and any method of non-contact data transmission may be employed.

The control device 14 has a processing circuit having a CPU (Central Processing Unit) and the like, and drive mechanisms such as motors and actuators. The control device 14 has a function of receiving an input signal from an input interface attached to the console device 30 or the gantry device 10 and controlling the operations of the gantry device 10 and the bed device 20 . For example, the control device 14 receives an input signal and performs control to rotate the rotating frame 13 , control to tilt the gantry device 10 , and control to operate the bed device 20 and the tabletop 23 . Note that the control for tilting the gantry device 10 is performed by the control device 14 based on the tilt angle (tilt angle) information input through the input interface attached to the gantry device 10. This is achieved by rotating the Note that the control device 14 may be provided in the gantry device 10 or may be provided in the console device 30 .

The bed device 20 is a device for placing and moving a subject P to be scanned, and includes a base 21 , a bed driving device 22 , a top plate 23 and a support frame 24 . The base 21 is a housing that supports the support frame 24 so as to be vertically movable. The bed driving device 22 is a motor or actuator that moves the table 23 on which the subject P is placed in the longitudinal direction of the table 23 . A top plate 23 provided on the upper surface of the support frame 24 is a plate on which the subject P is placed. Note that the bed drive device 22 may move the support frame 24 in the longitudinal direction of the top plate 23 in addition to the top plate 23 .

The console device 30 is a device that receives an operator's operation of the X-ray CT apparatus 1 and reconstructs CT image data using detection data collected by the gantry device 10 . The console device 30 has a memory 31, a display 32, an input interface 33, and a processing circuit 34, as shown in FIG. Memory 31, display 32, input interface 33, and processing circuitry 34 are communicatively connected to each other.

The memory 31 is implemented by, for example, a RAM (Random Access Memory), a semiconductor memory device such as a flash memory, a hard disk, an optical disk, or the like. The memory 31 stores projection data and CT image data, for example.

The display 32 displays various information. For example, the display 32 outputs a medical image (CT image) generated by the processing circuit 34, a GUI (Graphical User Interface) for accepting various operations from the operator, and the like. For example, the display 32 is a liquid crystal display or a CRT (Cathode Ray Tube) display.

The input interface 33 receives various input operations from the operator, converts the received input operations into electrical signals, and outputs the electrical signals to the processing circuit 34 . For example, the input interface 33 allows the operator to input acquisition conditions for acquiring projection data, reconstruction conditions for reconstructing CT image data, image processing conditions for generating post-processed images from CT images, and the like. accept. For example, the input interface 33 is implemented by a mouse, keyboard, trackball, switch, button, joystick, and the like.

A processing circuit 34 controls the operation of the entire X-ray CT apparatus 1 . For example, processing circuitry 34 performs system control functions 341 , preprocessing functions 342 , reconstruction processing functions 343 , and image processing functions 344 . Additionally, in this embodiment, processing circuitry 34 also performs segmentation function 345 , execution function 346 , evaluation function 347 and color allocation function 348 . The processing circuitry 34 is implemented by a processor. Segmentation function 345 is an example of a segmentation processor. The execution function 346 is an example of an execution unit and a second calculation unit. The evaluation function 347 is an example of an evaluation unit, a calculation unit, and a first calculation unit.

Here, for example, a system control function 341, a preprocessing function 342, a reconstruction processing function 343, an image processing function 344, a segmentation function 345, an execution function 346, an evaluation function 347, and a color allocation function 348, which are components of the processing circuit 34 Each function of is stored in the memory 31 in the form of a computer-executable program. The processing circuit 34 implements each function by reading each program from the memory 31 and executing each read program. In other words, the processing circuit 34 with each program read has each function shown in the processing circuit 34 of FIG.

Further, although the description has been given assuming that each function described above is realized by the single processing circuit 34, the processing circuit 34 is configured by combining a plurality of independent processors, and each processor executes a program to function. may be realized.

In addition, the term "processor" is, for example, CPU (Central Processing Unit), GPU (Graphics Processing Unit), application specific integrated circuit (ASIC), or programmable logic device (e.g., simple programmable logic Devices (Simple Programmable Logic Device: SPLD), Complex Programmable Logic Device (CPLD), Field Programmable Gate Array (FPGA), and other circuits. The processor implements its functions by reading and executing programs stored in the memory 31 . Instead of storing the program in the memory 31, the program may be directly incorporated into the circuit of the processor. In this case, the processor implements its functions by reading and executing the program embedded in the circuit. It should be noted that the configuration is not limited to a single circuit for each processor, and a single processor may be configured by combining a plurality of independent circuits to realize its function. Furthermore, multiple components may be integrated into a single processor to achieve its functionality.

The system control function 341 controls various functions of the processing circuit 34 based on input operations received from the operator via the input interface 33 . For example, the system control function 341 controls CT scans performed in the X-ray CT apparatus 1 . Also, the system control function 341 controls generation and display of CT image data in the console device 30 by controlling a preprocessing function 342 , a reconstruction processing function 343 , and an image processing function 344 . The system control function 341 that performs such display control is an example of a display control unit.

A preprocessing function 342 generates data by performing preprocessing such as logarithmic conversion processing, offset correction processing, inter-channel sensitivity correction processing, and beam hardening correction on the detection data output from the DAS 17 . Data before preprocessing (detection data) and data after preprocessing may be collectively referred to as projection data.

The reconstruction processing function 343 performs reconstruction processing using the filtered back projection method, the iterative reconstruction method, or the like on the projection data generated by the preprocessing function 342 to obtain CT image data (reconstructed image data).

The image processing function 344 converts the CT image data generated by the reconstruction processing function 343 into tomographic image data of an arbitrary cross section or three-dimensional Convert to image data. Note that the three-dimensional image data is composed of a plurality of pixels (voxels).

In addition, in this embodiment, the image processing function 344 converts CT image data of a plurality of phases into three-dimensional image data of a plurality of phases. Such three-dimensional image data in a plurality of time phases is so-called four-dimensional image data, and is obtained, for example, by CT scanning the same region including the target portion of the subject P multiple times at different time phases. .

Note that the four-dimensional image data may be generated by the reconstruction processing function 343 described above. For example, the image processing function 344 or the reconstruction processing function 343 according to this embodiment generates four-dimensional image data captured including the target region of the subject P in a plurality of time phases. That is, the four-dimensional image data is data obtained by imaging the target site of the subject P. FIG. The image processing function 344 and the reconstruction processing function 343 are examples of a generator.

Details of the segmentation function 345, the execution function 346, the evaluation function 347 and the color allocation function 348 will be described later.

By the way, in surgery to remove a tumor such as lung cancer, there is a case where the lung field parenchyma is removed in units of lung lobes or lung segments. In this case, it is necessary to remove the pleura from the lung parenchyma and then resect the lung parenchyma. can be done. However, if the pleura is attached to the lung parenchyma, it is necessary to remove the pleura from the lung parenchyma while burning the attachment site. Note that "adherence" means, for example, sticking between tissues. “Attachment” includes the sticking of a tissue forming one site to a tissue forming another site, and the sticking of a tissue forming the same site to another tissue. In addition, "adhesion" shall include "adhesion" and "infiltration".

Therefore, for example, immediately before the extraction surgery, an ultrasonic diagnostic apparatus displays an ultrasonic image in which the lung field and pleura of the subject are visualized in real time on the display, so that the operator can see that the pleura is attached to the lung field. It is conceivable to determine whether or not However, in this case, it is difficult for the operator to determine whether or not there is adhesion in areas where ultrasonic waves do not reach, such as the back side of the ribs and the vicinity of the subject's heart. Cardiac surgical assistance may also be required if mediastinal attachments are present. However, it is difficult for ultrasound to reach the mediastinum, and the degree of adhesion is not known.

Such a problem is not limited to the case where the pleura adheres to the lung field, but can similarly occur when adhesion occurs at other target sites. For example, such a problem can also occur when the visceral pleura adheres to the parietal pleura that constitutes the pleura.

Therefore, for example, it is conceivable to configure the image processing apparatus as described below so that the state of adhesion can be determined even in a range where ultrasonic waves do not reach. For example, a reference point is set for each of the two target parts, and if the magnitude of the difference between the movement vectors of the two reference points is equal to or less than a threshold, the image is determined to be adhered to the two target parts. It is conceivable to construct a processing unit. Hereinafter, such a medical image processing apparatus will be described as an image processing apparatus according to a comparative example.

However, the respiration volume of a subject varies greatly among individuals. Therefore, when the respiration volume of the subject is relatively small, the magnitude of the difference between the movement vectors of the two reference points may be less than or equal to the threshold even though there is no adhesion. In this case, the image processing apparatus according to the comparative example may erroneously determine adhesion.

Further, for example, when two target parts are attached at one point, the magnitude of the difference between the movement vectors of the two reference points may be larger than the threshold. For example, in the case of adhesion at one point, one target site moves relative to the other target site like a pendulum. The other target part moves. In this case, the image processing apparatus according to the comparative example may erroneously determine that there is no adhesion even though there is adhesion.

Therefore, the X-ray CT apparatus 1 according to the first embodiment performs the evaluation process described below in order to accurately evaluate the state of adhesion of one or more parts (target parts) that are subject to adhesion determination. Run. In addition, the state of attachment of the target site indicates, for example, a state in which the tissues constituting the target site are attached to each other. In the following description, the evaluation of the state of adhesion will be described as an example of the state of adhesion, but other states of adhesion such as the state of infiltration may be evaluated in a similar manner.

FIG. 2 is a flowchart illustrating an example flow of evaluation processing according to the first embodiment. For example, when the input interface 33 receives an instruction to execute the evaluation process, the evaluation process includes the system control function 341, the image processing function 344, the segmentation function 345, the execution function 346, the evaluation function 347, and the color allocation function. 348.

Step S<b>101 shown in FIG. 2 is a step corresponding to the segmentation function 345 . Step S101 is a step in which the segmentation function 345 is realized by the processing circuit 34 calling and executing a program corresponding to the segmentation function 345 from the memory 31 . Steps S 102 to S 105 and S 107 are steps corresponding to the execution function 346 . Steps S102 to S105 and S107 are steps in which the execution function 346 is realized by the processing circuit 34 calling and executing a program corresponding to the execution function 346 from the memory 31 .

Step S<b>106 is a step corresponding to the evaluation function 347 . Step S106 is a step in which the evaluation function 347 is realized by the processing circuit 34 calling and executing a program corresponding to the evaluation function 347 from the memory 31 . Step S108 is a step corresponding to the color allocation function 348. FIG. Step S108 is a step in which the color allocation function 348 is realized by the processing circuit 34 calling and executing a program corresponding to the color allocation function 348 from the memory 31 .

Step S<b>109 is a step corresponding to the image processing function 344 . Step S<b>109 is a step in which the image processing function 344 is realized by the processing circuit 34 calling and executing a program corresponding to the image processing function 344 from the memory 31 . Step S110 is a step corresponding to the system control function 341. FIG. Step S110 is a step in which the system control function 341 is realized by the processing circuit 34 calling a program corresponding to the system control function 341 from the memory 31 and executing it.

Note that, before the evaluation process is executed, a plurality of three-dimensional image data of a plurality of time phases depicting a target site to be determined for adhesion are stored in advance in the memory 31 . Also, in the following description, a case where the target site is the lung field and the site outside the lung will be described. Areas outside the lung include the pleura. However, the target site is not limited to this, and for example, the target site may be the parietal pleura and the visceral pleura. The memory 31 stores three-dimensional image data for T (T is a natural number) frames as three-dimensional image data of a plurality of time phases. In this embodiment, the three-dimensional image data of the Kth (K=1, . . . T) time phase corresponds to the three-dimensional image data of the Kth frame.

As shown in FIG. 2, the segmentation function 345 performs a known segmentation process on the three-dimensional image data of one time phase to segment the lung field and the lung field from the entire area of the three-dimensional image data. A part region is extracted (step S101).

For example, in step S<b>101 , the segmentation function 345 acquires three-dimensional image data of multiple temporal phases stored in the memory 31 . Then, the segmentation function 345 selects 3D image data of a predetermined phase from among the 3D image data of a plurality of phases.

FIG. 3 is a diagram for explaining an example of evaluation processing according to the first embodiment. For example, the segmentation function 345 selects the three-dimensional image data 40 of the first phase. A case where the three-dimensional image data 40 of the first time phase is selected as the predetermined time phase will be described below as an example. Then, the segmentation function 345 performs segmentation processing on the selected three-dimensional image data 40 of the first time phase, and extracts a lung field region 41 and a lung field region 42 . That is, the segmentation function 345 executes segmentation processing for extracting the lung field region 41 and the lung field region 42 from the image indicated by the three-dimensional image data 40 . Note that the region 42 of the portion outside the lung is an example of the first region. Also, the lung field region 41 is an example of the second region. The lung field region 41 is a region adjacent to the lung field region 42 in the normal direction of the boundary 43 . Also, the part outside the lung is an example of the first part. Also, the lung field is an example of the second part.

Then, the execution function 346 sets a one-dimensional region of interest (ROI) 44 on the boundary 43 between the lung field region 41 and the lung field region 42 (step S102). As a specific example, the execution function 346 aligns the longitudinal direction of the region of interest 44 with the normal direction of the boundary 43 and the region of interest 44 includes a portion of the boundary 43 . A region 44 is set. Thus, the region of interest 44 is set such that the longitudinal direction of the one-dimensional region of interest 44 and the boundary 43 are perpendicular to each other. The region of interest 44 set in this manner includes a plurality of pixels 45 arranged one-dimensionally. That is, the plurality of pixels 45 of the image represented by the three-dimensional image data 40 are aligned in the normal direction of the boundary 43 across the boundary 43 between the lung field region 41 and the lung field region 42 . Here, the normal direction of the boundary 43 is an example of a predetermined direction. Note that in the present embodiment, the execution function 346 sets the region of interest 44 such that the central portion of the region of interest 44 in the longitudinal direction is positioned at the boundary 43 . However, execution function 346 may set region of interest 44 such that the longitudinal central portion of region of interest 44 does not lie on boundary 43 .

As shown in FIG. 3, for example, when the length of the region of interest 44 in the longitudinal direction is 40 mm, the region of interest 44 includes a plurality of pixels 45 arranged over the length of 40 mm. . A case where the region of interest 44 has a length of 40 mm will be described below, but the length is not limited to 40 mm and may be other values.

The execution function 346 then redefines the region of interest 44 by removing a predetermined region centered on the boundary 43 from the entire region of the region of interest 44 (step S103). For example, in step S103, as shown in FIG. 3, a region of 10 mm in the longitudinal direction of the region of interest 44 centered on the boundary 43 is removed from the entire region of the region of interest 44 defined in step S102. The area left without being processed is used as a new area of interest 44 .

As shown in FIG. 3, the redefined region of interest 44 will include 14 pixels 45 . A case where the number of pixels 45 included in the region of interest 44 is 14 will be described below, but the number of pixels 45 included in the region of interest 44 is not limited to this. Pixels 45 included in the redefined region of interest 44 are used for various processes after step S104.

Note that the execution function 346 may omit the process of redefining the region of interest 44 in step S103. In this case, a plurality of pixels 45 within the region of interest 44 set in step S102 are used for various processes after step S104.

Then, the execution function 346 executes each process of steps S104 and S105 by executing part of the clustering process for clustering the plurality of pixels 45 into a plurality of clusters. A cluster is also called a class. In this embodiment, the execution function 346 executes part of the graph cut (graph cut processing, graph cut method), which is an example of clustering processing.

FIG. 4 is a diagram for explaining an example of evaluation processing according to the first embodiment. As a specific example, as shown in FIG. 4, the execution function 346 generates a graph 50 having nodes 51 to 66 (step S104). The graph 50 is, for example, a graph in which each of the 14 pixels 45 shown in FIG. In FIG. 4, eight nodes 54 to 61 out of the 14 nodes 51 to 64 are omitted.

For example, node 51 corresponds to the first pixel 45 from the left in FIG. The same is true for the other nodes 52-64. That is, in FIG. 4, among the plurality of 14 nodes 51 to 64 arranged from left to right, the j-th node from the left (j=1, 2, . . . 14) is the j-th node from the left in FIG. corresponds to the pixel 45 of . In the following description, the j-th pixel 45 from the left in FIG. 3 is referred to as the "j-th pixel", and the j-th node from the left among the plurality of nodes 51 to 64 in FIG. 4 is referred to as the "j-th node". There is

A node 65 corresponds to the region 42 of the lung field located outside the boundary 43 . A node 66 corresponds to the lung field region 41 located inside the boundary 43 .

Further, each of the plurality of nodes 51 to 64 and the node 65 are connected by lines (line segments). Similarly, each of the plurality of nodes 51 to 64 and node 66 are also connected by lines. Two adjacent nodes are also connected by a line. Here, two adjacent nodes refer to, for example, the n (n=1, 2, . . . 13)-th node and the (n+1)-th node.

At step S104, execution function 346 sets a weight for each line. First, the weight P(l _{j — 0} ) set to the line connecting each of the plurality of nodes 51 to 64 and the node 65 will be described. A weight P(l j _{— 0} ) is set to the line connecting the jth node and node 65 . Note that the weight P(l _{j — 0} ) is an example of a value related to the cost when each of the plurality of pixels 45 is classified into a first cluster and a second cluster which will be described later. For example, the weight P(l _{j — 0} ) is given by the following equation (1).

Here, in Expression (1), p(l _j ) is, for example, a value equal to or greater than "0" and equal to or less than "1". A specific example of p(l _j ) will be described. As described above, node 65 corresponds to region 42 of the lung field. For this reason, a relatively high value is assigned to the line connecting each of the nodes 51 to 57 corresponding to the seven pixels 45 located in the region 42 outside the lung shown in FIG. l _{j — 0} ), execution function 346 does the following: That is, the execution function 346 substitutes "0.001" for p(l _j ) in equation (1).

On the other hand, a relatively low value is the weight P(l _{j_0} ), execution function 346 does the following: That is, the execution function 346 substitutes "0.999" for p(l _j ) in equation (1).

Next, the weight P(l _j — 1) set to the line connecting each of the plurality of nodes 51 to 64 and the node 66 will be described. A weight P(l _j — 1 ) is set to the line connecting the jth node and node 66 . Note that the weight P(l _j — 1 ) is an example of a value related to the cost when each of the plurality of pixels 45 is classified into a first cluster (to be described later) and a second cluster (to be described later). For example, the weight P(l _j — 1) is given by Equation (2) below.

A specific example of p(l _j ) in equation (2) will now be described. As described above, node 66 corresponds to lung field region 41 . For this reason, a relatively high value is assigned to the weight P(l _{j_1} ), the execution function 346 performs the following processing. That is, the execution function 346 substitutes "0.999" for p(l _j ) in equation (2).

On the other hand, a relatively low value is the weight P(l _j — 1) for the line connecting each of the nodes 51 to 57 corresponding to the seven pixels 45 not located in the lung field region 41 shown in FIG. , the execution function 346 performs the following processing. That is, the execution function 346 substitutes "0.001" for p(l _j ) in equation (2).

From the above, when the execution function 346 calculates the weight P(l _{j_0} ) and the weight P(l _{j_1} ) for the j-th node, p(l _j ) assign a common value to

Next, the weight G(n, n+1) set for the line connecting two adjacent nodes will be described. A weight G(n,n+1) is set to the line connecting the nth node and the (n+1)th node. For example, the weight G(n, n+1) is represented by the following formula (3).

Here, in equation (3), “N” refers to the number of pixels 45 included in the redefined region of interest 44 . As a specific example, the value of "N" is "14" in this embodiment.
Also, g(n, n+1) in the formula (3) is represented by the following formula (4).

Here, in Equation (4), X _{n_i} indicates the motion vector (movement vector) of the n-th pixel. Specifically, X _{n_i} is the 3D image of the (i+1)th frame from the position of the nth pixel in the 3D image data of the i (i=1, 2, . . . T−1)th frame. Points to the vector pointing to the location of the nth pixel in the data.

The same is true for X _{(n+1)_i} . Specifically, X _{(n+1)_i} indicates the motion vector of the (n+1)-th pixel from the left in FIG. For example, X _{(n+1)_i} is a vector from the position of the (n+1)-th pixel in the 3-dimensional image data of the i-th frame to the position of the (n+1)-th pixel in the 3-dimensional image data of the (i+1)-th frame. point to

In this embodiment, the execution function 346 traces the position of each of the plurality of pixels 45 from the 3D image data of the first frame to the 3D image data of the Tth frame to obtain the motion vector X _{n_i} and the motion vector X _{(n+1)_i} can be calculated. In this way, the execution function 346 calculates the motion vector X _{n_i} and the motion vector X _{(n+1)_i} between the images of the plurality of time phases in the plurality of pixels 45 . Note that the motion vector _{Xn_i} and the motion vector X _{(n+1)_i} are examples of movement information.

FIG. 5 is a diagram for explaining an example of evaluation processing according to the first embodiment. As shown in FIG. 5 and equation (4), the execution function _{346 calculates the} difference vector (X Calculate the magnitude of _{n_i} - X _{(n+1)_i} ). Then, the execution function 346 calculates the sum of the magnitudes of the plurality of calculated difference vectors (X n — i _{−X (n+1)} — i ) for all three-dimensional images adjacent on the time axis, as shown in Equation (4). g(n, n+1) is calculated by dividing by the number of data intervals (T−1).

If two adjacent pixels (nth pixel and (n+1)th pixel) 45 move in the same way, the value of g(n,n+1) will be relatively small. In such a case, it is assumed that one pixel 45 of the two adjacent pixels 45 is a pixel located in the lung field region 41 and the other pixel 45 is a pixel located in the lung field region 42 . Conceivable. Alternatively, two adjacent pixels 45 are both considered to be pixels located in the lung field region 41 or the lung field region 42 . When the value of g(n, n+1) is relatively small and one pixel 45 is located in the lung field region 41 and the other pixel 45 is located in the lung field region 42, the subject P It is considered that the lung field and the part outside the lung are adhered to each other.

On the other hand, if the motions of two adjacent pixels 45 differ relatively greatly, the value of g(n, n+1) will be relatively large. In such a case, one pixel 45 of the two adjacent pixels 45 is a pixel located in the lung field region 41, and the other pixel 45 is a pixel located in the lung field region 42. , it is thought that there is no adhesion between the lung field and the lung field.

Note that g(r, r+1) in equation (3) is the same as g(n, n+1).

Here, in a general graph cutting process (graph cutting method), a graph is cut by each of a plurality of lines (hereinafter referred to as cut lines) in order to finally classify a plurality of pixels into a plurality of clusters. , the energy is calculated for each cutline. Then, in general graph cut processing, clustering is performed in which each pixel is classified into each cluster based on the cutline with the minimum energy among the energies calculated for each cutline.

On the other hand, in the present embodiment, although the execution function 346 performs part of the graph cutting process, it does not perform clustering for finally classifying each pixel 45 into each cluster. The execution function ₃₄₆ calculates the _energy E (L ₁ , L ₂ , .

Note that “γ” in Equation (5) is a coefficient with a positive value.

Also, Lc (c=1, 2, . be done. On the other hand, Lc is set to "1" when the cut line intersects the line connecting the c-th node and the node 66.

Also, "w" in equation (5) is set to "c" when the cut line intersects the line connecting the c-th node and the (c+1)-th node. If the cut line does not intersect any line connecting two adjacent nodes among the plurality of nodes 51 to 64, a positive integer other than "1" to "N-1" is set to a predetermined value (for example, "-1" which is a negative integer).

Also, "E _d (L ₁ , L ₂ , . . . , L _N )" in Equation (5) is represented by Equation (6) below.

“E _d ( _{L 1} , L ₂ , . With such energy, there is a high possibility that the boundary between the lung field region and the lung field region based on the cut line is the actual boundary between the lung field region and the lung field region of the subject P. becomes smaller.

Also, "E _s (w)" in Equation (5) is represented by Equation (7) below.

"Es(w)", also referred to as a smoothing term, refers to the energy that accounts for boundary smoothness independently of the data terms described above. For example, only the data term considers only the CT value, so it is susceptible to noise and may not have smooth boundaries. However, the presence of the smoothing term smooths the boundaries. In "E _s (w)", the value of "E _s (-1)" in which "w" is set to the above-described predetermined value, for example, -1 becomes "0".

Further, when it is considered that the lung field and the lung field of the subject P are adhered to each other, the value of g(n, n+1) becomes relatively small as described above. Then, as the value of g(n, n+1) decreases, the value of the smoothing term increases. On the other hand, when it is considered that the lung field and the lung field of the subject P are not adhered to each other, the value of g(n, n+1) is relatively large and the value of the smoothing term is small.

6 and 7 are diagrams for explaining an example of cut lines according to the first embodiment. As shown in FIG. 6, for example, execution function 346 cuts graph 50 through cut line 70 . The cut line 70 is a cluster (first cluster) to which the first pixel and the second pixel belong to the cluster (first cluster) corresponding to the lung field region 42, and a cluster (second cluster ) to classify the 14 pixels 45 so that the 3rd to 14th pixels belong to the . Note that the first cluster is also called a first class. The second cluster is also referred to as the second class.

Also, as shown in FIG. 7, the execution function 346 cuts the graph 50 along cut lines 71 . The cut line 71 is a line for classifying the 14 pixels 45 so that the 1st to 14th pixels belong to the second cluster. The execution function 346 also cuts the graph 50 along the cut line 72 . The cut line 72 is a line for classifying the 14 pixels 45 so that the 1st to 14th pixels belong to the first cluster. As shown in FIG. 7, the cut lines 71 and 72 do not intersect any line connecting two adjacent nodes among the plurality of nodes 51-64.

In this embodiment, the execution function 346 has 13 cutlines that intersect the line connecting the c-th node and the (c+1)th node, and the above-described two cutlines 71 and 72, for a total of 15 cutlines. cut the graph 50 at each of the cut lines.

And the execution function 346 calculates the energy E ₍ L1, L2,..., _LN , w) by Formula (5) for _every cutline. As described above, in this embodiment, the execution function 346 calculates the energy E (L ₁ , L ₂ , . . . , L _N , w) for each cutline, but each pixel 45 is Do not perform clustering to classify into

In steps S104 and S105, the execution function 346 performs clustering processing for classifying the plurality of pixels 45 into a plurality of clusters based on the respective motion vectors of the plurality of pixels 45 within the region of interest 44 including the boundary 43 of the target part. Do some processing. For example, as part of the clustering process, the execution function 346 calculates motion vectors for each of the plurality of pixels 45, and based on the calculated plurality of motion vectors, in the plurality of pixels 45, two adjacent pixels 45, a process of calculating the magnitude of the difference vector (X n — i _{−X (n+1)} — i ) between the two motion vectors of the two pixels 45 is executed. The two adjacent pixels 45 here are pixels adjacent to each other in the normal direction of the boundary 43 .

Further, in steps S104 and S105, the execution function 346 performs the first cluster related to the lung field region 42 or the lung field region 41 based on the motion vector _{Xn_i} and the motion vector X _{(n+1)_i} . _Calculate the energy E(L ₁ , L ₂ , . Energy E (L ₁ , L ₂ , . . . , L _N , w) is an example of classification information. Also, the energy E (L ₁ , _{L 2} , . be.

The evaluation function 347 then calculates an index indicating the state of adhesion (step S106). A specific example will be given for explanation. The evaluation function 347 calculates the probability p as an index indicating the state of adhesion, for example, using the following equation (8).

The denominator of the term on the right side of equation (8) represents the sum of 15 exp(-E(L ₁ , L ₂ , . . . , L _N , w)) for 15 cutlines.

Also, exp(-E _{( 0} , . , L _N , w), L ₁ =L ₂ = . . . =L _N =0, and w=−1.

Also, exp(-E _{( 1} , . , L _N , w), L ₁ =L ₂ = . . . =L _N =1, and w=−1.

The probability p is a small value between "0" and "1". The probability p is, for example, the probability that a plurality of pixels 45 belong to one cluster (first cluster or second cluster) when clustering the plurality of pixels 45 within the region of interest 44 . In this way, the probability p is a value indicating the possibility that the part outside the lung is adhered to the lung field, which is the target part. As the value of probability p approaches “1”, the possibility of adhesion increases. Note that the probability p does not indicate the degree of adhesion.

As described above, in step S106, as shown in equation (8), the evaluation function 347 calculates the region 42 of the region outside the lung from the energy E (L ₁ , L ₂ , . . . , L _N , w). and the lung field of the subject P corresponding to the region 41 of the lung field.

Also, in step _S106 , the evaluation function 347 calculates energy E ₍ L ₁ , L ₂ , . . . , L _N , w) to calculate the index. That is, the evaluation function 347 calculates an index based on the magnitude of the calculated difference vector (X n — i _{−X (n+1)} — i ).

8 to 13 are diagrams for explaining an example of evaluation processing according to the first embodiment. 8 to 13 show a region of interest 44 that is set to include part of a boundary 43 between a lung field region 41 and a lung field region 42 . 8 to 13 show the case where the process of step S103 for redefining the region of interest 44 is omitted. That is, FIGS. 8 to 13 show the case where one one-dimensional region of interest 44 is set.

The three-dimensional image data 40 shown in FIG. 8 is imaged data including the lung field and the lung field part of the subject P whose respiration volume is relatively large. FIG. 9 shows a plurality of arrows indicating the orientations (directions) and magnitudes of motion vectors of a plurality of pixels 45 within the region of interest 44 shown in FIG. In FIG. 9, the orientations of a plurality of motion vectors are unified in approximately two directions. Also, in FIG. 9, the magnitudes of the plurality of motion vectors are relatively large. This is because the respiration volume of the subject P is relatively large.

On the other hand, the three-dimensional image data 40 shown in FIG. 10 is imaged data including the pulmonary and pulmonary regions of the subject P whose respiratory volume is relatively small. FIG. 11 shows a plurality of arrows indicating the direction and magnitude of motion vectors of a plurality of pixels 45 within the region of interest 44 shown in FIG. In FIG. 11, similar to FIG. 9, the orientations of a plurality of motion vectors are unified in approximately two directions. However, in FIG. 11, the magnitudes of the motion vectors are relatively small. This is because the respiration volume of the subject P is relatively small.

Further, the three-dimensional image data 40 shown in FIG. 12 is imaged data including the lung field and the part of the lung field that are adhered at one point. FIG. 13 shows a plurality of arrows indicating the direction and magnitude of motion vectors of a plurality of pixels 45 within the region of interest 44 shown in FIG. In FIG. 13, the orientations and magnitudes of the motion vectors are non-uniform. This is because the lung field and the lung field of the subject P are fused together at one point, and the lung field moves relative to the lung field like a pendulum, for example.

In this embodiment, the evaluation function 347 calculates the probability p indicating the possibility of adhesion based on the similarity of motion vectors of a plurality of pixels 45 . For example, when the orientations and magnitudes of a plurality of motion vectors are uniform, the evaluation function 347 considers that the lung field region 41 and the lung field region 42 are moving while adhering to each other. A probability p close to "1" is calculated. For example, the evaluation function 347 calculates a probability p close to "1" when the directions of a plurality of motion vectors are substantially the same (substantially one direction) and the magnitudes are substantially the same.

In addition, the evaluation function 347 determines that the lung field region 41 and the lung field region 42 are moving independently of each other when the orientations of a plurality of motion vectors having substantially the same magnitude are substantially in two directions. Therefore, the probability p close to "0" is calculated. For example, when the respiration volume of the subject P is relatively large as shown in FIGS. 9 and 10, and when the respiration volume of the subject P is relatively small as shown in FIGS. 11 and 12, the evaluation function 347 computes the probability p that is close to '0'. Therefore, according to the first embodiment, the state of adhesion can be evaluated with high accuracy not only when the respiration volume is relatively large, but also when it is relatively small.

Also, the evaluation function 347 calculates a probability p close to "1" when the directions and magnitudes of a plurality of motion vectors are uneven. For example, as shown in FIGS. 12 and 13, when the lung field and the lung field part of the subject P are fused at one point, the directions and magnitudes of the plurality of motion vectors are uneven, Since it is considered appropriate for more than one pixel 45 to belong to one cluster, the evaluation function 347 calculates a probability p close to "1". Therefore, according to the first embodiment, even if the lung field and the part of the lung field of the subject P are adhered at one point, the adhesion state can be evaluated with high accuracy.

Thus, in step S106, the evaluation function 347 evaluates the state of adhesion of the target site within the region of interest 44 based on the result of a part of the clustering process. In addition, the evaluation function 347 uses the probability p that the plurality of pixels 45 belong to one cluster when clustering the plurality of pixels 45 based on the result of a part of the clustering processing as an index indicating the adhesion state. calculate. Note that the evaluation function 347 determines the probability (1-p ) may be calculated as an index indicating the state of adhesion. That is, when the evaluation function 347 classifies the plurality of pixels 45 into two clusters (the first cluster and the second cluster described above), based on the probability p that the plurality of pixels 45 belong to the same cluster, An index may be calculated to assess the state of adhesion.

Also, in step S106, the evaluation function 347 _{calculates weights P(lj_0 )} , weights P( _{lj_1} ) and energies E _{( L1} , _L2 , . . . , L _N , w), the probability p is calculated, and the index is calculated based on the probability p.

Also, in step _{S106 ,} the evaluation function 347 calculates energy E(L ₁ , L ₂ , . . . , L _N ,w) to calculate the index.

The evaluation function 347 then associates the 14 pixels 45 within the region of interest 44 with the probability p. For example, the evaluation function 347 generates correspondence information in which identification information for identifying each pixel 45 is associated with the probability p, and stores the generated correspondence information in the memory 31 .

The execution function 346 then determines whether or not the region of interest 44 has been set over the entire boundary 43 (step S107). If it is determined that the region of interest 44 has not been set over the entire portion (step S107: No), the execution function 346 returns to step S102 and sets the region of interest 44 out of the entire portion of the boundary 43. A region of interest 44 is set in the portion. Then, each process of steps S103 to S106 is repeatedly executed until the execution function 346 determines that the region of interest 44 has been set over the entire boundary 43 .

On the other hand, if the execution function 346 determines that the region of interest 44 has been set over the entire area (step S107: Yes), the color allocation function 348 assigns A color corresponding to the probability p is assigned to it (step S108).

For example, the color allocation function 348 obtains correspondence information stored in the memory 31 . Then, the color allocation function 348 allocates a color to each pixel according to the correspondence relationship between the identification information indicated by the correspondence information and the probability p. Specifically, the color allocation function 348 allocates a color according to the probability p corresponding to the pixel indicated by the identification information. For example, the color assignment function 348 may assign a color closer to red to a pixel as the probability p approaches "1", and assign a color closer to blue as the probability p approaches "0". Also, the color allocation function 348 may allocate a predetermined color to pixels that are not associated with a probability p. Here, the predetermined color is, for example, a color other than the color assigned to the pixel according to the probability p.

Then, the image processing function 344 generates two-dimensional image data for display based on the color-assigned three-dimensional image data (step S109). For example, the image processing function 344 performs surface rendering on three-dimensional image data to generate surface rendering image data as two-dimensional image data for display.

Then, the system control function 341 causes the display 32 to display an image indicated by the two-dimensional image data for display (step S110), and ends the evaluation process. FIG. 14 is a diagram illustrating an example of an image represented by two-dimensional image data for display according to the first embodiment. For example, as shown in FIG. 14, the system control function 341 causes the display 32 to display the surface rendering image 80 indicated by the surface rendering image data. In this way, the system control function 341 causes the display 32 to display the evaluation results evaluated by the evaluation function 347 . As a result, the operator who has confirmed the surface rendering image 80 can easily grasp which part of the target site is likely to be adhered.

The X-ray CT apparatus 1 according to the first embodiment has been described above. According to the X-ray CT apparatus 1, as described above, the state of adhesion can be evaluated with high accuracy.

(First Modification of First Embodiment)
Although the case where the one-dimensional region of interest 44 is set has been described in the above-described first embodiment, the region of interest 44 may be, for example, two-dimensional or three-dimensional instead of one-dimensional. Therefore, such a modified example will be described as a first modified example of the first embodiment. In the first modification, the X-ray CT apparatus 1 uses pixels in the two-dimensional or three-dimensional region of interest 44 to perform the same processing as in the first embodiment.

For example, in the first modified example, in the two-dimensional region of interest 44 , a pixel row composed of a plurality of pixels arranged one-dimensionally along the normal direction of the boundary 43 is arranged in the normal direction of the boundary 43 . are lined up in a direction perpendicular to the Therefore, in the first modification, the X-ray CT apparatus 1 performs the same processing as that performed on the plurality of pixels 45 arranged one-dimensionally in the first embodiment, on each of the plurality of pixel columns. against

Even when the three-dimensional region of interest 44 is set, the execution function 346 performs the same processing as when the two-dimensional region of interest 44 is set. Therefore, according to the first modified example, the same effects as those of the first embodiment can be obtained.

(Second Modification of First Embodiment)
Also, in the first embodiment and the first modified example described above, the case where the execution function 346 executes a part of the clustering process has been described, but the entire clustering process may be executed. In this case, the execution function _{346 selects} the minimum energy E ₍ L ₁ , L ₂ , . . . , L _N , w) is used to classify each pixel 45 into each cluster. That is, the execution function 346 executes all clustering processing.

Thus, execution function 346 selects the first cluster or the first cluster based on energy E(L ₁ , L ₂ , . . . , L _N , w) based on motion vector Xn_i and motion vector X _{(n+1)_i} . A clustering process is performed that classifies each of the plurality of pixels 45 into two clusters. That is, the execution function 346 executes clustering processing based on the motion vector Xn_i and the motion vector X _{(n+1)_i} .

When the X-ray CT apparatus 1 executes all the clustering processes, the clustering process results may be used to calculate the reliability indicating the likelihood of the probability p. Therefore, such a modified example will be described as a second modified example of the first embodiment.

15 and 16 are diagrams for explaining an example of processing executed by the X-ray CT apparatus according to the second modification. FIG. 15 shows a lung field region 41 and a lung field region 42 extracted from the three-dimensional image data 40 by the segmentation function 345 . FIG. 15 also shows a region 82 in which the pixels 45 classified into the first cluster by the execution function 346 are located, and a region 81 in which the pixels 45 classified into the second cluster are located.

Then, in the second modification, the evaluation function 347 calculates the degree of matching between the boundary 43 between the lung field region 41 and the lung field region 42 and the boundary 83 between the regions 81 and 82 with the probability p It is calculated as a reliability that indicates the certainty of Note that the evaluation function 347 uses a known technique as a method of calculating the degree of matching.

Here, when the degree of matching between the boundary 43 and the boundary 83 is high, the degree of matching between the clustering processing result and the segmentation processing result is also high. The higher the degree of matching between the result of the clustering process and the result of the segmentation process, the higher the reliability indicating the likelihood of the result of the clustering process. As the reliability indicating the certainty of the result of the clustering process increases, the reliability indicating the certainty of the probability p also increases.

Therefore, the evaluation function 347 according to the second modification calculates the degree of matching between the boundary 43 and the boundary 83 as the degree of reliability indicating the likelihood of the probability p.

Also, assume that execution function 346 classifies pixels 45 such that all pixels 45 within region of interest 44 belong to the second cluster. In this case, as shown in FIG. 16, there is a region 84 where the pixels 45 belonging to the second cluster are located, but there is no region where the pixels 45 belonging to the first cluster are located. Therefore, there is no boundary between the area 84 and the area where the pixels 45 belonging to the first cluster are located. In this way, when there is no boundary, the evaluation function 347 calculates a predetermined value, for example "0", as the degree of reliability indicating the likelihood of the probability p. Similarly, evaluation function 347 also calculates a predetermined value, e.g., "0", when execution function 346 classifies pixels 45 within region of interest 44 such that all pixels 45 belong to the first cluster. do.

In this way, the evaluation function 347 calculates the reliability of the adhesion state evaluation result based on the segmentation processing result and the clustering processing result.

Then, the system control function 341 according to the second modification causes the display 32 to display the calculated reliability.

The X-ray CT apparatus 1 according to the second modification has been described above. According to the second modification, it is possible to quantitatively display the reliability indicating the probability of the probability p. Therefore, according to the second modified example, it is possible for the operator to grasp the likelihood of the probability p more reliably.

(Second embodiment)
Here, the functions of the X-ray CT apparatus 1 according to the first embodiment, the first modification, or the second modification are provided to an image processing apparatus connected to the X-ray CT apparatus 1 via a network. be able to. Such an embodiment will be described as a second embodiment with reference to FIG.

FIG. 17 is a diagram showing an example of the configuration of a system including an image processing device according to the second embodiment. The system shown in the example of FIG. 17 has an X-ray CT device 600 , an image storage device 700 , an image display device 800 and an image processing device 900 . The X-ray CT apparatus 600, the image archiving apparatus 700, the image display apparatus 800, and the image processing apparatus 900 are directly or indirectly connected by an in-hospital LAN (Local Area Network) 500 installed in the hospital, for example. can communicate with each other. For example, when PACS (Picture Archiving and Communication System) is introduced, the devices 600 to 900 mutually transmit and receive images and the like according to the DICOM (Digital Imaging and Communications in Medicine) standard.

The X-ray CT apparatus 600 is the X-ray CT apparatus 1 according to the first embodiment, the first modified example, or the second modified example. The X-ray CT apparatus 600 transmits four-dimensional image data (three-dimensional image data for T frames) to the image processing apparatus 900 .

The image storage device 700 is a database that stores two-dimensional image data for display generated by the X-ray CT device 600 and the image processing device 900 .

The image processing apparatus 900 is a workstation and has the functions of the X-ray CT apparatus 1 according to the first embodiment, the first modified example, or the second modified example. The image processing apparatus 900 uses T frames of three-dimensional image data transmitted from the X-ray CT apparatus 600 to perform the X-ray CT apparatus according to the first embodiment, the first modification, or the second modification. 1 performs the same processing as the processing performed by 1.

The image processing device 900 has an input interface 901 , a display 902 , a memory 903 and a processing circuit 904 .

The input interface 901 receives input operations of various instructions and various information from the operator. Specifically, the input interface 901 is connected to the processing circuit 904 , converts an input operation received from the operator into an electrical signal, and outputs the electrical signal to the processing circuit 904 . For example, the input interface 901 includes a trackball, a switch button, a mouse, a keyboard, a touch pad that performs input operations by touching an operation surface, a touch screen that integrates a display screen and a touch pad, and a non-optical sensor using an optical sensor. It is realized by a contact input interface, a voice input interface, and the like. In this specification, the input interface 901 is not limited to having physical operation parts such as a mouse and a keyboard. For example, the input interface 901 also includes an electrical signal processing circuit that receives an electrical signal corresponding to an input operation from an external input device provided separately from the apparatus and outputs the electrical signal to the processing circuit 904. .

The display 902 displays various information and various images. Specifically, the display 902 is connected to the processing circuit 904, converts various information and image data sent from the processing circuit 904 into electrical signals for display, and outputs the electrical signals. For example, the display 902 is realized by a liquid crystal monitor, a CRT (Cathode Ray Tube) monitor, a touch panel, or the like.

The memory 903 stores various data. Specifically, the memory 903 stores various images. For example, the memory 903 is implemented by a RAM (Random Access Memory), a semiconductor memory device such as a flash memory, a hard disk, an optical disk, or the like.

A processing circuit 904 performs overall control of the image processing apparatus 900 . For example, when receiving T frames of three-dimensional image data transmitted from the X-ray CT apparatus 600 , the processing circuit 904 stores the received three-dimensional image data in the memory 903 . Processing circuitry 904 is implemented by a processor. The processing circuitry 904 has an acquisition function 904a, a segmentation function 904b, an execution function 904c, an evaluation function 904d, a color allocation function 904e, a display control function 904f and an image processing function 904g. The acquisition function 904a is an example of an acquisition unit. The segmentation function 904b is an example of a segmentation processor. The execution function 904c is an example of an execution unit and a second calculation unit. The evaluation function 904d is an example of an evaluation unit, a calculation unit, and a first calculation unit. The display control function 904f is an example of a display control unit. The image processing function 904g is an example of a generator.

Here, for example, each function of the acquisition function 904a, the segmentation function 904b, the execution function 904c, the evaluation function 904d, the color allocation function 904e, the display control function 904f, and the image processing function 904g, which are the components of the processing circuit 904, are executed by a computer. It is stored in the memory 903 in the form of an executable program. The processing circuit 904 implements each function by reading each program from the memory 903 and executing each read program. In other words, the processing circuit 904 with each program read has each function shown in the processing circuit 904 of FIG.

The acquisition function 904 a acquires three-dimensional image data for T frames stored in the memory 903 . That is, the acquisition function 904a acquires a plurality of three-dimensional image data captured including the target region of the subject P at a plurality of time phases.

The segmentation function 904b executes the same processing as the processing of step S101 described above. The execution function 904c executes the same processes as those of steps S102 to S105 and S107 described above. The evaluation function 904d executes the same processing as the processing of step S106 described above.

The color allocation function 904e executes processing similar to the processing of step S108 described above. The image processing function 904g executes the same processing as the processing of step S109 described above. The display control function 904f executes the same processing as the processing of step S110 described above.

The image processing apparatus 900 according to the second embodiment has been described above. According to the image processing apparatus 900, similarly to the first embodiment, the first modified example, and the second modified example, it is possible to accurately evaluate the state of adhesion.

According to at least one embodiment or modified example described above, the state of adhesion can be evaluated with high accuracy.

While several embodiments of the invention have been described, these embodiments have been presented by way of example and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, replacements, and modifications can be made without departing from the scope of the invention. These embodiments and their modifications are included in the scope and spirit of the invention, as well as the scope of the invention described in the claims and equivalents thereof.

900 Image processing device 904a Acquisition function 904d Evaluation function

Claims (8)

an acquisition unit that acquires a plurality of images captured including a target region of a subject in a plurality of time phases;
Based on the movement information between the plurality of time phase images in the plurality of pixels of the image arranged in a predetermined direction across the boundary between the first region and the second region in the image, the first From classification information for classifying each of the plurality of pixels into a first class related to an area or a second class related to a second area adjacent to the first area in the predetermined direction, the first a first calculator that calculates an index indicating a state of adhesion at the boundary between a first portion of the subject corresponding to the region and a second portion of the subject corresponding to the second region;
An image processing device comprising: The first calculation unit calculates the index based on a probability that the plurality of pixels belong to the same class when classifying the plurality of pixels into the first class and the second class. 2. The image processing apparatus according to claim 1, wherein said state of adhesion is evaluated. The first calculation unit calculates the probability using a value associated with a cost for classifying each of the plurality of pixels into the first class or the second class based on the movement information, 3. The image processing apparatus according to claim 2, wherein said index is calculated based on probability. A motion vector of each of the plurality of pixels is calculated as the movement information, and based on the plurality of calculated motion vectors, for each two pixels adjacent to each other in the predetermined direction in the plurality of pixels, 2 pixels of the two pixels are calculated. Further comprising a second calculation unit for calculating the magnitude of the difference vector of the two motion vectors,
4. The image processing device according to claim 1, wherein said first calculator calculates said index based on the magnitude of said difference vector. Further comprising a segmentation processing unit that performs segmentation processing for extracting the first region and the second region from the image,
The second calculation unit performs clustering processing for classifying each of the plurality of pixels into the first class or the second class based on the movement information,
The first calculation unit calculates the reliability of the indicator based on the result of the segmentation process and the result of the clustering process.
The image processing apparatus according to claim 4. The image processing apparatus according to any one of claims 1 to 5, further comprising a display control section that displays the index on a display section. 6. The image processing apparatus according to claim 5, further comprising a display control section for displaying said reliability on a display section. a generation unit that generates a plurality of images captured including a target region of a subject in a plurality of time phases;
Based on the movement information between the plurality of time phase images in the plurality of pixels of the image arranged in a predetermined direction across the boundary between the first region and the second region on the image, the first from the classification information for classifying each of the plurality of pixels into a first class related to a region of or a second class related to a second region adjacent to the first region in the predetermined direction, the first a calculation unit that calculates an index indicating a state of adhesion at the boundary between a first part of the subject corresponding to the area of and a second part of the subject corresponding to the second area;
A medical imaging device comprising:

Images