JP4736755B2 - Modeling device, region extraction device, modeling method and program - Google Patents

Description

  The present invention relates to a modeling technique for a model related to a desired region in measurement data.

  In recent years, in-vivo three-dimensional images (simply referred to as “three-dimensional images”) taken by an X-ray CT (Computed Tomography) apparatus or MRI (Magnetic Resonance Imaging) apparatus have been widely used in the medical field. . As a result, information on internal organs and the like can be visually grasped, and an improvement in diagnostic accuracy can be expected. However, on the other hand, since the three-dimensional image used for diagnosis is composed of tens to hundreds of slice images, only information necessary for diagnosis is obtained from such an enormous amount of information. Is a big burden for doctors.

  Therefore, there is an increasing demand for quantitative or automatic diagnosis using a computer, and research on a computer aided diagnosis (CAD Computer-aided diagnosis) system has been actively conducted.

  In order to perform diagnosis support by a computer, it is an important issue to accurately extract information necessary for diagnosis, that is, an organ region or shape from a three-dimensional image.

  As an organ region extraction method, there is a method using model fitting (Model Fitting) for deforming a standard solid model prepared in advance based on the principle of energy minimization and finding a target contour.

  As a technique using such model fitting, a first three-dimensional model and a second three-dimensional model that are associated with each other are assumed, the first three-dimensional model is deformed by a model fitting method, and a second three-dimensional model is obtained. A technique for deforming a model in accordance with the deformation of the first three-dimensional model has been proposed (see Patent Document 1).

JP 2002-15338 A

  However, when model fitting is individually performed using a plurality of three-dimensional models for each of a plurality of different regions (organs) included in the three-dimensional image, overlapping such as contact or intersection between the plurality of three-dimensional models There is a problem that stable modeling cannot be performed.

  Therefore, the present invention has been made in view of the above problems, and it is an object of the present invention to provide a technique capable of performing stable modeling by a model fitting technique capable of preventing overlap between different regions (organs). And

In order to solve the above-described problem, the invention of claim 1 is a modeling device for modeling an organ existing in measurement data as an object, and is a standard model relating to the first object in the measurement data. Optimizing the first evaluation function using a model fitting method and means for obtaining a first standard model and a second standard model that is a standard model for the second object in the measurement data And deforming the first standard model so as to convert the first standard model, and finally creating a model relating to the first object, wherein the first evaluation function is a function of the first standard model. due the first standard model by deformation in proximity with the second standard model, characterized in that it comprises an element to which the evaluation value is degraded.
According to a second aspect of the present invention, in the modeling device according to the first aspect of the present invention, the element stores energy stored in a virtual spring set between the first standard model and the second standard model. It is expressed by the function of
According to a third aspect of the present invention, in the modeling device according to the first aspect of the invention, the evaluation value is deteriorated when the energy stored in the virtual spring is increased, and the energy stored in the virtual spring is the first Due to the deformation of the standard model, the first standard model and the second standard model approach each other, and the relative displacement between the two models increases, and the energy stored in the virtual spring is increased by the first model. Due to the deformation of the standard model, the first standard model and the second standard model are separated from each other, and the relative displacement between the two models increases.

According to a fourth aspect of the present invention, in the modeling apparatus according to any one of the first to third aspects of the present invention, the model deforming means optimizes the second evaluation function using a model fitting technique. wherein deforming the second standard model, the second evaluation function, due to the proximity of the said first standard model and the second standard model by deformation of said first standard model, the It includes an element that deteriorates the evaluation value.

Further, the invention of claim 5 is the modeling apparatus according to claim 4 of the invention, wherein the model deforming means optimizes the first evaluation function and the second evaluation function, respectively. The first standard model and the second standard model are repeatedly deformed in order, respectively.

The invention according to claim 6 is the modeling apparatus according to claim 4 or 5 , wherein the model deforming means is one standard of the first standard model and the second standard model. The modification of the model is completed, and then the deformation of the other standard model is performed.

The invention of claim 7 is the modeling apparatus according to any one of claims 1 to 6 , wherein the presence of the first object is obtained from the measurement data using a model relating to the first object. The image processing apparatus further includes extraction means for extracting a region.

The invention according to claim 8 is the modeling apparatus according to any one of claims 1 to 3 , wherein the model deforming means optimizes the second evaluation function using a model fitting technique. The second standard model is deformed to finally create a model related to the second object, and the extraction unit uses the model related to the second object to extract the second object from the measurement data. It is characterized by extracting a region where an object exists.

According to a ninth aspect of the present invention , there is provided a region extracting apparatus for extracting an organ existing in measurement data as an object, wherein the first standard model is a standard model relating to the first object in the measurement data. And a second standard model that is a standard model for the second object in the measurement data, and the first evaluation function is optimized using a model fitting technique. Model deformation means for deforming one standard model and finally creating a model relating to the first object, and using the first standard model, the existence area of the first object is determined from the measurement data. and means for extracting the first evaluation function, due to the proximity of the said first standard model and the second standard model by deformation of said first standard model, its evaluation value Includes deteriorating elements And wherein the door.

The invention of claim 10 is a modeling method in which an organ existing in measurement data is an object, and a first standard model that is a standard model related to the first object in the measurement data; Obtaining a second standard model, which is a standard model for the second object in the measurement data, and optimizing the first evaluation function using a model fitting technique. deforming the standard model, the first model for the object and a model modification step of finally created, the first evaluation function, the first standard by deformation of said first standard model Due to the approach between the model and the second standard model, an element whose evaluation value deteriorates is included.

The invention of claim 11 causes a computer, a program for executing the modeling with the object organ present in the measurement data, the computer, standard for the first object in the measurement data Obtaining a first standard model that is a model and a second standard model that is a standard model relating to a second object in the measurement data, and a first evaluation function using a model fitting technique the deforming said first standard model to optimize the model finally to execute the model deformation process to create for the first object, the first evaluation function, the first due to the proximity of the said first standard model by deformation of a standard model and the second standard model, characterized in that it comprises an element to which the evaluation value is degraded.

According to the present invention, in a model modified using the first standard model and the second standard model, Ru exacerbate evaluation value due to the approach of the first standard model and the second standard model elements Since the first standard model is deformed using the evaluation function that introduces, it is possible to prevent an excessive approach between the standard models due to the deformation of the model and to avoid duplication due to contact or intersection between the standard models. It becomes possible, and stable modeling becomes possible.

In particular, according to the invention described in claim 2 , since the virtual spring is set between the standard models and the model is deformed, the initial relative positional relationship between the first standard model and the second standard model is reflected. Modeling becomes possible.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

<1. First Embodiment>
<Configuration>
FIG. 1 is a diagram showing an outline of a region extraction apparatus 1A according to the first embodiment of the present invention. The region extraction device 1A is a device that extracts a region (specifically, a three-dimensional region) of a desired object from measurement data (specifically, three-dimensional measurement data (also referred to as three-dimensional measurement data)) including a plurality of regions (objects). It is. Note that the region extraction apparatus 1A is also referred to as a “modeling apparatus” because it generates a target object model using a model fitting technique.

  As shown in FIG. 1, the area extracting apparatus 1 </ b> A includes a personal computer (hereinafter simply referred to as “PC”) 2, a monitor 3, a mounting unit 5, and an operation unit 4.

  The personal computer 2 includes a control unit 20, an input / output I / F 21, and a storage unit 22.

  The input / output I / F 21 is an interface (I / F) for transmitting and receiving data between the monitor 3, the operation unit 4, the mounting unit 5 and the external network and the personal computer 2. Send and receive data with.

  The storage unit 22 is composed of, for example, a hard disk and stores a software program (hereinafter simply referred to as “program”) PG for executing region extraction.

  The control unit 20 mainly includes a CPU, a ROM 20a, a RAM 20b, and the like, and is a part that performs overall control of each unit of the personal computer 2.

  The monitor 3 is composed of, for example, an LCD, and visually outputs a display image generated by the control unit 20.

  The operation unit 4 includes a keyboard, a mouse, and the like, and transmits various signals accompanying various operations of the user (user) to the input / output I / F 21.

  The mounting unit 5 can detachably mount a storage medium such as the memory card 51. Various data or programs stored in the memory card 51 attached to the attachment unit 5 can be taken into the control unit 20 or the storage unit 22 via the input / output I / F 21.

  Next, various functions provided in the region extraction device 1A will be described.

  FIG. 2 is a block diagram showing various functions provided in the area extraction apparatus 1A.

  These various functions are realized by executing a predetermined program PG using various hardware such as a CPU in the control unit 20.

  As shown in FIG. 2, the region extraction apparatus 1A includes a model fitting unit 11, a region extraction unit 12, and an extraction region output unit 13, and receives an input three-dimensional image (“three-dimensional measurement data” or “volume data”). The object (hereinafter also referred to as “extraction target area (object)”) that the user (user) or the like wants to extract can be extracted from the user and output.

  The model fitting unit 11 has a function of creating a model of the extraction target region by deforming a standard model of the extraction target region (extraction target object) using a model fitting method.

  Here, the “standard model” is a three-dimensional model used in the model fitting method, and means a standard model of an extraction target region (for example, an organ). Further, the standard model is acquired from the model storage unit 14 in the area extraction apparatus 1A.

  The region extraction unit 12 has a function of extracting an extraction target object from the three-dimensional measurement data based on the extraction target region model created by the model fitting unit 11.

  The extraction area output unit 13 has a function of displaying and outputting the extracted target area (object) on the monitor 3.

<Process overview>
Next, each function of the region extraction apparatus 1A described above will be described in more detail. In the following description, extraction processing of an object (for example, an organ) from a three-dimensional image (volume data) acquired by an X-ray CT apparatus will be described. However, the present invention is applied to other measurement data (for example, an MRI apparatus). The present invention can also be applied to region extraction processing from acquired three-dimensional images or the like.

  Here, a three-dimensional image acquired by the X-ray CT apparatus will be described.

  FIG. 3 is a diagram showing a three-dimensional image TD1 acquired by the X-ray CT apparatus.

  As shown in FIG. 3, the three-dimensional image TD1 is composed of a plurality (for example, several tens to several hundreds) of slice images (tomographic images) SD showing a cross section of an object (human body). . Each slice image SD is an image visualized by displaying the amount of X-ray absorption (CT value) at each point (each pixel) in grayscale. In FIG. 3, a slice image PM at a predetermined position among the plurality of slice images SD constituting the three-dimensional image TD1 is also shown.

  FIG. 4 is a flowchart showing the overall operation of the area extracting apparatus 1A. FIG. 5 is a diagram showing three objects (three-dimensional objects) existing in certain volume data.

  As illustrated in FIG. 4, the region extraction device 1A performs one or more objects specified by the user (user) from the input three-dimensional image (volume data) by executing the steps S1 to S3. Extract (organ).

  First, in step S1, the individual model corresponding to the extraction target object is finally obtained by deforming the standard model arranged at the initial position in the volume data by a user (user) manual operation or the like by the model fitting method. To be created. In the model fitting executed in step S1, deformation of the model is executed such that objects (organs) existing in the volume data are in contact with each other and do not cause an overlap such as an intersection.

  In step S2, a region existing at the corresponding position of the individual model created in step S1 is extracted from the three-dimensional image as a region to be extracted (object).

  In step S <b> 3, the extracted region (object) is displayed and output on the monitor 3.

  The extraction target object is designated by the user (user) before starting the process of step S1. As a specification method, for example, a user (user) selects a desired item by operating the operation unit (for example, mouse) 4 or the like from a list of various organ items stored in the database in the storage unit 22 in advance. It is possible to adopt a mode or the like.

  In the following description, it is assumed that volume data having three objects (organs) as shown in FIG. 5 is input to the area extraction apparatus 1A, and the three objects (first object AA, second object BB). And a case where the third object CC) is extracted as the extraction target region.

<Model fitting process>
In the model fitting process of step S1, each “standard model”, which is a general (standard) three-dimensional model for each of three extraction target objects prepared in advance, is obtained from three extraction target objects ( And the like. In the present application, a standard model that is finally created after deformation by model fitting (in other words, a standard model that reflects information of an object to be extracted) is also referred to as an “individual model”.

  Here, the standard model of the extraction target object will be described in detail. FIG. 6 shows that standard models (SA, SB, SC) relating to three extraction target objects (first object AA, second object BB, and third object CC) have a standard positional relationship. It is a figure which shows a mode.

  The standard model of the extraction target object is composed of, for example, a minute polygon (for example, a triangle) and is stored in the storage unit 22 or the like. The standard model composed of polygons can express the surface shape of the model by the three-dimensional coordinates of the vertices of each polygon. In addition, of the vertices of the polygon in the standard model, all vertices or some representative vertices are also referred to as “control points”.

  In addition, each standard model SA, SB, SC related to the three extraction target objects AA, BB, CC has a plurality of “adjacent points” associated with each other. It is specified in advance from a positional relationship in which two extraction target objects generally exist (standard).

  Specifically, this will be described with reference to FIG. It is assumed that the standard model related to the three extraction target objects (first object AA, second object BB, and third object CC) has a positional relationship as shown in FIG. 6 in a predetermined cross section. In this case, in each contour (boundary) where a certain standard model and another standard model are close to each other, predetermined points on each contour are associated as adjacent points.

  A predetermined point on a contour (boundary) where a certain standard model and another standard model are close to each other is associated as an adjacent point. For example, in the vicinity of the boundary where the standard model SA and the standard model SB in FIG. 6 are close to each other, any three points (points HAb1, HAb2, HAb3) on the contour (boundary) of the standard model SA are contours of the standard model SB. (Points HBa1, HBa2, and HBa3) are associated with each other on the (boundary). As a method of this association, for example, among the pixels indicating the contour (boundary) of the standard model SB, for each of three arbitrary points (adjacent points of the standard model SA) on the contour (boundary) of the standard model SA, A method can be used in which each pixel existing in the vicinity is set as an adjacent point of the standard model SB.

  The correspondence performed between the standard model SA and the standard model SB is similarly performed between the standard model SA and the standard model SC and between the standard model SB and the standard model SC. Each attached adjacent point is preset to each standard model together with its correspondence.

  In the plane shown in FIG. 6, the case where there are three sets of adjacent points of the standard model SA and the standard model SB is illustrated, but the number of sets of the adjacent points is arbitrary. Further, since each standard model is a three-dimensional model, adjacent points set between the standard models are not always set on the same plane.

  Each control point and each adjacent point of the standard model set as described above are used in model fitting described later.

  Hereinafter, the model fitting process will be described in detail with reference to FIGS.

  FIG. 7 is a flowchart showing details of the model fitting process (step S1). FIG. 8 is a diagram showing the extraction target object AA included in the three-dimensional image (measurement data), and FIG. 9 is a diagram conceptually showing how the standard model SA is moved to the initial position. FIG. 9 also shows the positional relationship between the extraction target object and the standard model when the standard model is initially arranged. FIG. 10 is a schematic diagram in which the vicinity of the region RQ (see FIG. 9) centering on the boundary point Q1 of the extraction target object AA is enlarged, and FIG. 11 is a diagram showing some slice images PQ1, PQ2, and PQ4. It is.

  First, in step S11, the standard model of the extraction target object is initially placed in the extraction target area in the measurement data. As a method for determining the initial position, for example, a method in which an operator (user) manually arranges three standard models prepared in advance at appropriate positions can be employed. Specifically, as shown in FIG. 8, when the object AA that is one of the extraction target objects exists in the three-dimensional image (measurement data) TD1, the operator (user) The standard model SA of the extraction target object AA is moved, and the standard model SA is initially arranged in the measurement data TD1 (see FIG. 9). Further, corresponding standard models (SB and SC) are initially arranged for each of the other extraction target objects (BB and CC) by the same method. Note that the initial position determination method may be automatically performed using a region expansion method or the like.

  In the next step S12, processing for determining a point corresponding to a control point of the standard model (hereinafter also referred to as “corresponding point”) is performed in each standard model. Here, a method is used in which a pixel existing closest to a control point among pixels (boundary points) indicating the contour (boundary) of the extraction target object is a corresponding point of the control point. Note that the outline (boundary) of each extraction target object is acquired by performing edge extraction processing or the like in each slice image SD.

  For example, in FIG. 10 (elevated view), the corresponding point of the control point C1 is the boundary point (pixel) Q1 that exists closest to the control point C1. Here, in FIG. 10, only the boundary points in a predetermined plane (two-dimensional space) parallel to the xz plane are shown, but actually, the control point C1 in the three-dimensional space as shown in FIG. The boundary point existing closest to is a corresponding point. In FIG. 11, the projection points in the slice images PQ1, PQ2, and PQ4 of the control point C1 are indicated by x.

  When the corresponding point of each control point in each standard model is determined, the process proceeds to step S13. Each process from step S13 to step S18 described below (also referred to as “model deformation process”) is sequentially executed for each standard model. Specifically, each process (model deformation process) from step S13 to step S18 is a loop process that is repeatedly executed until it is determined to end in a predetermined determination process described later. For example, if the model deformation process is executed for the standard model SA, the next model deformation process after the predetermined determination process is executed for a model other than the standard model SA (for example, the standard model SB or SC). In other words, in the model transformation process of the first embodiment, the standard model SA is first executed, then the standard model SB, and then the standard model SC are repeatedly executed in this order.

  Hereinafter, the model deformation process (steps S13 to S18) will be described in detail. In the description of the model deformation process, the model deformation process performed on the standard model SA will be described as an example as appropriate.

  In step S13, one arbitrary point (hereinafter also referred to as “movement target point”) (for example, control point C1) among the control points Cj of each standard model (for example, standard model SA) is minute in one direction (for example, the A1 direction). L is moved (see FIG. 10).

  Further, the three-dimensional coordinates of the adjacent points that have moved with the movement of the movement target point can be obtained by an appropriate interpolation method using the coordinate values of the respective control points Cj.

  Furthermore, in step S14, the total energy Ue of the model in which the movement target point is moved and temporarily deformed in step S13 (hereinafter also referred to as “temporary deformation model”) is calculated.

  The total energy Ue is an evaluation function related to the movement of the control point Cj performed in step S13, and can also be expressed as a model evaluation function of a temporary deformation model created by the movement of the control point Cj. This total energy Ue is used to adopt an optimal movement of the control point Cj, that is, an optimal model deformation.

  Such total energy Ue is constituted by various functions representing the deformation state of the temporary deformation model. Specifically, as shown in equation (1), external energy Fe regarding the distance between the control point Cj and the corresponding point, internal energy Ge for avoiding excessive deformation, and crossing between models are avoided. It is represented by the sum of the mutual energy He. External energy Fe, internal energy Ge, and mutual energy He will be described later.

  Next, in step S15, it is determined whether or not the movement target point has been moved in all directions. For example, if the movement direction of the movement target point in all directions in the three-dimensional space is 26 directions (direction toward 26 pixels adjacent to the movement target point (pixel)), the control point C1 is moved in all 26 directions. It is determined whether or not.

  If the creation of the temporary deformation model in which the movement target point (control point C1) is moved in all 26 directions has not been completed, the movement direction of the movement target point is changed and moved again by a small amount L, and different movement direction patterns are obtained. A temporary deformation model is created (step S13), and the total energy Ue of each temporary deformation model is calculated (step S14).

  If it is determined in step S15 that the movement in all directions has been completed, the process proceeds to step S16.

  In step S16, a temporary deformation model that minimizes the total energy Ue is selected from the created temporary deformation models of all patterns. In other words, an optimal temporary deformation model that minimizes the total energy Ue is selected from the temporary deformation models generated by moving a certain control point Cj in 26 directions. For example, when only the external energy Fe is considered without considering the internal energy Ge and the mutual energy He, a temporary deformation model in which the control point moves in a direction approaching the corresponding point is selected. When the internal energy Ge is considered in addition to the external energy Fe, a temporary deformation model in which the control point moves in a direction different from the direction approaching the corresponding point may be selected. Furthermore, if mutual energy He is considered in addition to external energy Fe and internal energy Ge, selection of a temporary deformation model that reflects the positional relationship between the standard model in which the deformation is performed and another standard model (optimization) ) Can be prevented, and an excessive approach between the standard models can be prevented, and an effect of avoiding duplication due to contact or intersection between the models can be obtained.

  Next, in step S17, it is determined whether or not all the control points Cj have been moved. Specifically, it is determined whether or not the movement of the minute amount L has been completed for all the control points Cj of the standard model (for example, standard model SA), and there is a control point that is not completed (also referred to as an incomplete point). Then, the movement target point is changed to the control point (incomplete point), and the operations in steps S13 to S17 are repeated to create a temporary deformation model in which the movement of all the control points Cj is completed. On the other hand, if the movement of all the control points has been completed, the process proceeds to step S18.

  In step S18, it is determined whether or not to end model deformation (creation of an individual model) for a predetermined standard model (for example, standard model SA). Specifically, when the condition is satisfied on condition that the average value of the distances between the plurality of control points and the corresponding points in the temporary deformation model in which the movement of all the control points Cj is completed is equal to or less than a predetermined value In addition, the model deformation (individual model creation) may be terminated. According to this, when each control point approaches the corresponding point to a predetermined extent, the same process can be ended. Alternatively, or in addition to this, the amount of change in the total energy Ue between the previous temporary deformation model (which has completed the movement of all control points Cj) and the current temporary deformation model is equal to or less than a predetermined amount. Whether or not it may be used as a criterion for end determination. According to this, a process can be complete | finished when the total energy does not change so much even if a control point is moved.

  Thereafter, the process proceeds to step S19 in principle.

  In step S19, the deformation target of the model is changed, and thereafter, the process from step S13 to step S18 is re-executed for the changed standard model.

  For example, when the processing target in step S13 to step S18 is the standard model SA, the processing target is changed to the standard model SB, and then the processing from step S13 to step S18 is executed for the standard model SB. If the target of processing up to that time is the standard model SB, the processing target is changed to the standard model SC, and then the processing from step S13 to step S18 is executed for the standard model SC. Further, when the processing target up to that time is the standard model SC, the processing target is changed to the standard model SA again, and thereafter, the processing from step S13 to step S18 is executed for the standard model SA.

  As described above, the model transformation process target is changed in the order of standard model SA → standard model SB → standard model SC → standard model SA →, and so on, and sequentially (cyclically) for each standard model. A model transformation process is executed.

  If it is determined in step S18 that the model deformation (individual model creation) for the predetermined standard model is to be completed while the above-described operation is repeated, the predetermined standard model finally obtained is determined. The temporary deformation model is determined as the individual model of the extraction target object corresponding to the predetermined standard model, the model deformation (creation of the individual model) for the predetermined standard model is terminated, and the process proceeds to step S22.

  In step S22, it is determined whether or not model deformation (creation of individual models) has been completed for all three standard models. When it is determined that the model deformation for all the standard models has not been completed yet, the process proceeds to step S19 and the above-described repetitive processing is continued. Specifically, among the three standard models, the above-described repetitive processing (steps S13 to S18 and the like) may be sequentially performed on a standard model that has not been deformed.

  On the other hand, when it is determined that the model deformation has been completed for all three standard models, the model fitting process (step S1) is terminated.

  In this embodiment, it is assumed that corresponding points are corrected in each iteration. Specifically, after step S19, the processes of steps S20 and S21 are performed.

  In step S20, it is determined whether the unit processing loop is executed a predetermined number of times W (for example, 10 times) for each standard model with the processing executed in steps S13 to S19 described above as a unit processing loop. If the predetermined number of times W has been executed, the process proceeds to step S21.

  In step S21, the corresponding points determined in step S12 are updated. Specifically, the corresponding point is updated with the pixel (boundary point) closest to each control point having moved W a predetermined number of times by the above processing loop as a new corresponding point for each control point, and the step is again performed. The processes of S13 to S21 are repeated. According to such “update of corresponding point”, it is possible to optimize the corresponding point even when the nearest boundary point of the control point changes with the movement of the control point. Note that after the corresponding point is updated, the counter variable for the standard model is reset.

  In this way, every time the temporary deformation model that completes the movement of all control points is created W a predetermined number of times, the corresponding point update operation is performed.

  In such a model fitting process (step S1), model deformation considering each other's shape change is performed by sequentially deforming a plurality of different standard models. When attention is paid to a certain standard model, an individual model corresponding to an extraction target object (extraction target region) is created by continuously deforming a small amount L for each control point. Further, as will be described later, by considering the mutual energy He, it is possible to create an individual model in consideration of the relative distance between the standard models.

  Here, the external energy Fe, the internal energy Ge, and the mutual energy He constituting the total energy Ue will be described. FIG. 12 is a schematic diagram in which control points are connected by virtual springs. FIG. 13 is a schematic diagram in which adjacent points corresponding to each standard model are connected by a virtual spring. FIG. 14 is a diagram illustrating a state in which the shape of each standard model is changed by the model deformation process.

  The external energy Fe is expressed by Equation (2) using the square of the distance between each control point Cj and the corresponding point Qj corresponding to each control point Cj in one target standard model (for example, standard model SA). It is expressed in

  Here, α is a constant, Nt is the number of control points, and | Cj−Qj | is the distance between the control point Cj and the corresponding point Qj.

  Since the temporary deformation model in which the external energy Fe is very large, that is, the temporary deformation model in which the distance between the control point Cj and the corresponding point Qj is larger than that before the movement, the total energy Ue becomes large. It becomes difficult to be employed in the above-described step S16 (step of employing movement of the control point that minimizes the total energy Ue).

  Further, for example, as shown in FIG. 12, the internal energy Ge is a virtual spring (also referred to as “control spring”) SPR (SPR1, SPR1,) between the control points Cj in one target standard model (eg, standard model SA). If it is assumed that they are connected by SPR2, SPR3,...

  Here, β is a constant, Kr is a spring constant of each control spring, Nr is the number of control springs, and Wr is a displacement amount of each control spring. The subscript u indicates that it relates to the u-th control spring.

  According to Equation (3), excessive movement of each control point Cj is expressed as an increase in energy stored in the control spring SPR. For example, if one control point Cz moves to a certain point Vz and the relative displacement with other control points increases, the control springs SPR1, SPR2 and SPR3 store energy due to the extension of each control spring. The internal energy Ge and thus the total energy Ue increases.

  In such a temporary deformation model with excessive deformation, the internal energy Ge increases and the total energy Ue also increases. Therefore, in the above-described step S16 (step of adopting the movement of the control point that minimizes the total energy Ue). It becomes difficult to be adopted.

  In other words, the temporary deformation model that reduces the internal energy Ge and decreases the total energy Ue is selected in step S16, thereby obtaining an effect of preventing excessive deformation due to movement of each control point Cj. it can.

  That is, by introducing such internal energy Ge, the control point Cj can be moved without impairing the shape of the standard model, that is, the shape of each polygon constituting the standard model.

  The mutual energy He is an evaluation value (model evaluation value) calculated based on the positional relationship (reciprocal relationship) between the standard models, and is composed of elements that deteriorate (become large) due to the proximity of the standard models. The

  Specifically, if the standard models related to the three extraction target objects are in the positional relationship as shown in FIG. 6, the mutual energy He (H1) related to the standard model SA is set to the other two standard models (SB, SC). ), And the mutual energy conceived between the standard model SC and the standard model SC is represented as Hab and the mutual energy conceived between the standard model SC and the standard model SC as Hac. The

  Here, as shown in FIG. 13, it is assumed that each adjacent point HAb of the standard model SA and each corresponding adjacent point HBa in the standard model SB are connected by a virtual spring (also referred to as “adjacent spring”) SABn. Then, the mutual energy Hab conceived between the standard model SA and the standard model SB is expressed as shown in Expression (5).

  However, γab is a constant, Kab is a spring constant of each adjacent spring, Nab is the number of adjacent springs, and Wab is a displacement amount of each adjacent spring. The subscript n indicates that it relates to the nth adjacent spring.

  Similarly, assuming that each adjacent point HAc of the standard model SA and each corresponding adjacent point HCa in the standard model SC are connected by a virtual spring (adjacent spring) SACm, the standard model SA and the standard model SC The mutual energy Hac conceived in between is expressed as shown in Equation (6).

  However, γac represents a constant, Kac represents a spring constant of each adjacent spring, Nac represents the number of adjacent springs, and Wac represents a displacement amount of each adjacent spring. The subscript m indicates that it relates to the m-th adjacent spring.

  According to the above formulas (5) and (6), excessive approach between standard models is expressed as an increase in energy stored in adjacent springs (SABn, SACm).

  For example, as shown in FIG. 14, if the standard model SA and the standard model SB approach each other and the relative displacement between the two models increases due to repeated model deformation processing (steps S13 to S18), The adjacent springs SABn existing between the models store energy due to the contraction of the adjacent springs SABn, and the mutual energy He and thus the total energy Ue increase.

  Such a temporary deformation model that greatly increases the mutual energy He increases the total energy Ue, and thus is adopted in the above-described step S16 (step of adopting the movement of the control point that minimizes the total energy Ue). It becomes difficult to be done.

  That is, by introducing such mutual energy He, a temporary deformation model that reduces the mutual energy He and decreases the total energy Ue is selected in step S16, and the standard models are excessively close to each other. It is possible to obtain an effect of avoiding duplication due to contact between standard models or intersection between standard models.

  In addition, when the mutual energy He is expressed by the energy stored in the virtual spring provided between the standard models as described above, model deformation processing reflecting the initial relative positional relationship of the standard models is also possible.

  Specifically, as shown in FIG. 14, if the standard model SA and the standard model SC are separated by the model deformation process and the relative displacement between the two models increases, the adjacent spring SACm existing between the two models. The energy due to the extension of each adjacent spring SACm is stored, and the mutual energy He and consequently the total energy Ue increases. Therefore, as described above, in step S16, the movement of the control point that minimizes the total energy Ue is employed. Therefore, such a deformation that separates the standard model SA from the standard model SC is difficult to be employed. That is, when the mutual energy He is expressed by the energy stored in the virtual spring, it is possible to obtain an effect of preventing excessive deformation in which the standard models are extremely close to or away from the initial relative position.

  In the above description, the mutual energy He of the standard model SA has been described. However, the mutual energy He (H2) of the standard model SB and the mutual energy He (H3) of the standard model SC can be similarly considered.

  Specifically, the mutual energy He (H2) of the standard model SB is considered with the other two standard models (SA, SC), and the mutual energy considered with the standard model SC is expressed as Hbc. Then, using the mutual energy Hab conceived between the standard model SA and the standard model SA, it is expressed as shown in Equation (7).

  Here, assuming that each adjacent point HBc of the standard model SB and each corresponding adjacent point HCb in the standard model SC are connected by a virtual spring (adjacent spring) SBCq, the standard model SB is connected between the standard model SB and the standard model SC. The mutual energy Hbc envisioned by is expressed as in equation (8).

  However, γbc is a constant, Kbc is the spring constant of each adjacent spring, Nbc is the number of adjacent springs, and Wbc is the amount of displacement of each adjacent spring. The subscript q indicates that it relates to the qth adjacent spring.

  In addition, the mutual energy He (H3) of the standard model SC is considered with the other two standard models (SA, SB), and the mutual energy Hac and the standard model SB are considered between the standard model SA and the standard model SC. The mutual energy that is conceived between and is expressed as in equation (9) using Hbc.

  The total energy Ue for each standard model is expressed by the sum of the external energy Fe for each standard model, the internal energy Ge for each standard model, and the mutual energy He for each standard model. For example, the total energy Ue related to the standard model SB is expressed by the sum of the external energy Fe of the standard model SB, the internal energy Ge of the standard model SB, and the mutual energy He (H2) of the standard model SB. The total energy Ue related to the standard model SC is expressed by the sum of the external energy Fe of the standard model SC, the internal energy Ge of the standard model SC, and the mutual energy He (H3) of the standard model SC.

  According to the above-described processing (step S1), temporary deformation models (optimally deforming each standard model by repeatedly creating a temporary deformation model that minimizes the total energy Ue of each standard model in each standard model) It is also possible to acquire an individual model for each object to be extracted.

  In step S2, each region (existing region) present at the corresponding position of each individual model is extracted and output as an extraction target region.

  As described above, in the model fitting using a plurality of standard models, the region extraction device 1A deforms each standard model so as to optimize the evaluation function that introduces an element that deteriorates due to the proximity of the standard models. Thereby, it is possible to prevent the standard models from being excessively approached due to the deformation of the model, and to avoid duplication due to contact or intersection between the standard models. For this reason, it is possible to stably extract an appropriate extraction target region.

<2. Second Embodiment>
Next, a second embodiment will be described. The configuration of the region extraction device 1B in the second embodiment is the same as that described in the first embodiment, and elements having the same functions as those in the first embodiment are denoted by the same reference numerals. Description is omitted.

  In the first embodiment, when three extraction target objects (regions) are extracted, the three standard models are sequentially and repeatedly deformed in the model fitting process (step S1), whereby individual models relating to the respective extraction target objects. However, in this embodiment, optimization of the evaluation function of one standard model (deformation of the standard model) is completed one by one in the model fitting process (step S1B), and each individual model is obtained. The case of acquiring is illustrated.

  Specifically, it is assumed that volume data having three objects (organs) as shown in FIG. 5 is input to the area extracting apparatus 1B, and the three objects (first object AA, second object BB). And the third object CC) as an extraction target region, the case where the individual model is acquired in this order, first, the object AA, then the object BB, and the object CC will be described.

  FIG. 15 is a flowchart showing details of the model fitting process (step S1B) executed by the region extracting apparatus 1B. FIG. 16 is a diagram illustrating an operation state executed by the region extraction device 1B.

  First, in steps S31 and S32, processing similar to that in steps S11 and S12 is executed, and the initial position of each standard model related to each extraction target object and the corresponding point of each control point of each standard model are determined.

  In steps S33 to S40, processing similar to that in steps S13 to S18, S20, and S21 described above is executed. Here, in the present embodiment, as described above, the individual model GA of the extraction target object AA is first created. Therefore, each processing of the steps S33 to S40 (hereinafter also referred to as “individual model creation processing”) is performed as the standard model SA. Done about.

  Further, in the individual model creation process of the extraction target object AA, the total energy Ue calculated in step S14 does not include the mutual energy He conceived with other standard models, and is expressed by Expression (10). Is simply expressed as the sum of external energy Fe and internal energy Ge. That is, in the individual model creation process of the extraction target object AA that is performed first, the mutual energy He based on the positional relationship with other standard models is not considered.

  Next, when the individual model GA for the extraction target object AA is created (state STA1 in FIG. 16), the individual model creation process for the standard model SA is terminated (step S38), and the process proceeds to step S41.

  In step S41, processing similar to that in step S22 is performed, and it is determined whether or not the creation of individual models has been completed for all three standard models.

  If it is determined in step S41 that the individual model creation process has not been completed for all three standard models, the process proceeds to step S42, and the processing target of the individual model creation process is changed. That is, the processing target of the individual model creation process is changed from the standard model SA to the standard model SB.

  Then, an individual model creation process (steps S33 to S40) is executed for the standard model SB. The total energy Ue used in the individual model creation process of the standard model SB is the mutual energy based on the positional relationship between the individual model GA of the extraction target object AA that has already been created and the standard model SB as shown in the state STA2 of FIG. Calculated taking into account the Hab. That is, the total energy Ue is represented by the sum of the external energy Fe, the internal energy Ge, and the mutual energy Hab as shown in the equation (11).

  Here, the mutual energy Hab based on the positional relationship between the individual model GA and the standard model SB is calculated based on the correspondence between each adjacent point of the individual model GA that has already been created and each adjacent point of the standard model SB. .

  When the individual model GB for the standard model SB is created, the individual model creation process is executed for the standard model SC. The total energy Ue used in the individual model creation process of the standard model SC is based on the individual model GA and the individual model GB already created as shown in the state STA3 in FIG. It is calculated in consideration of the energy Hac and Hbc. That is, the total energy Ue is expressed as in Expression (12).

  Here, the mutual energy Hac (Hab) is calculated from the correspondence between each adjacent point of the individual model GA (GB) already created and each adjacent point of the standard model SC.

  Then, when the individual model GC for the standard model SC is created, the individual model creation for all three standard models has been completed, so the model fitting process (step S1B) is terminated (step S41).

  According to the region extraction apparatus 1B including the above-described processing (step S1B), optimization of the evaluation function of the standard model (deformation of the standard model) is completed one by one, and individual models are created one by one. In other words, the created individual model, in other words, the individual model can be created sequentially with the positional relationship with the already extracted object as a constraint condition, and stable region extraction is possible.

<3. Modification>
Although the embodiments of the present invention have been described above, the present invention is not limited to the contents described above.

  For example, in the first embodiment, all three objects (regions) existing in the input volume data are extracted. However, the present invention is not limited to this. Specifically, the above-described model fitting process may be used even when one object is extracted from a plurality of objects existing in the input volume data.

  Specifically, for example, only the object AA is extracted from the volume data having three objects (the first object AA, the second object BB, and the third object CC) as shown in FIG. Let us consider a case where it is not necessary to extract the other two objects. Even in such a case, the above-described model fitting process (step S1) is performed in the same manner as in the above-described first embodiment, that is, in the same manner as when each of the three objects is extracted. Then, the object AA is extracted using the individual model for the object AA acquired by the model fitting process.

  According to this, it is possible to perform model fitting in which the positional relationship with an object not to be extracted is also taken into consideration as a constraint condition, and stable region extraction is possible.

  In the above modification, the case where only the object AA is extracted has been illustrated. However, the present invention is not limited to this, and 1 to (M−1) objects are extracted from the M objects existing in the input volume data. May be. For example, only the object BB (or only the object CC) may be extracted from the three objects AA, BB, and CC. Furthermore, the object BB and the object CC may be extracted from the three objects AA, BB, and CC.

  Similarly, in the second embodiment, all three objects (regions) existing in the input volume data are extracted. However, the present invention is not limited to this, and M existing in the input volume data is extracted. One to (M−1) objects may be extracted from each object. For example, out of the three objects AA, BB, and CC, only the object BB may be extracted, or only the object CC may be extracted. Furthermore, the object BB and the object CC may be extracted from the three objects AA, BB, and CC.

  In the second embodiment, three objects (the first object AA, the second object BB, and the third object CC) are used as extraction target areas, first the object AA, then the object BB, and the object CC. The case where individual models are acquired in this order is illustrated, but the present invention is not limited to this. Specifically, because the quality of the extraction results depends on the quality of the initial position accuracy of the standard model, a standard model that can determine the exact initial position among multiple objects to be extracted (standard with high initial position accuracy) Model) may be specified, and individual models may be acquired in order from the specified standard model.

  The following methods TQ1 and TQ2 can be used as a method for determining whether or not the initial position accuracy is good (specific).

◎ TQ1 (characteristic shape):
In each standard model corresponding to a plurality of extraction target objects, creation of individual models may be executed in order from a standard model having a characteristic shape compared to other standard models.

  FIG. 17 is a diagram showing three objects (three-dimensional objects) existing in certain volume data.

  For example, as illustrated in FIG. 17, a case is assumed in which the extraction target object BB1 has a unique feature point IP (for example, a point at the tip of a sharpened portion). Further, it is assumed that the feature point IP is a feature point of a portion related to the most characteristic shape in the three extraction target objects AA1, BB1, and CC1. In other words, it is assumed that the extraction target object BB1 has the most characteristic shape among the three extraction target objects AA1, BB1, and CC1.

  In this case, predetermined image processing is first performed to extract feature points IP from the three-dimensional image (volume data). Then, the feature point IS corresponding to the feature point IP in the standard model SB1 corresponding to the extraction target object BB1 (that is, the corresponding feature point IS in the standard model SB1) is moved to the feature point IP extracted from the three-dimensional image. Then, the standard model is initially arranged, and the evaluation function is optimized (deformation of the standard model) from the standard model SB1.

  Thereafter, the evaluation function of the standard model (SA1) for the remaining one extraction target object (for example, AA1) is optimized, and the standard model (SC1) for the last extraction target object (for example, CC1) is evaluated. Perform function optimization.

  In this way, among the standard models related to a plurality of extraction target objects, a standard model having a characteristic shape compared to other standard models is identified, and based on the characteristic shape of the extraction target object The individual model may be sequentially acquired from the standard model by determining the initial position of the standard model.

  Since the initial position of the standard model SB1 is determined based on the characteristic shape of the object to be extracted, it can be considered that the initial position accuracy is relatively higher than the other standard models SA1 and SC1.

  Thus, when determining the initial position of the standard model based on the characteristic shape of the extraction target object, compared to determining the initial position by other initial position determination methods (for example, manual operation by the user), It can be considered that the initial position accuracy is high. Then, by optimizing the evaluation function in order from the standard model with relatively high initial position accuracy (deformation of the standard model), deformation of the standard model with the positional relationship with the accurately extracted object as a constraint condition Will be possible sequentially.

◎ TQ2 (a priori knowledge):
Each standard model placed at the initial position in the volume data by a user's (user) manual operation, etc., and an existence probability map created based on a priori knowledge about the same type of object as each extraction target object The individual models may be created in order from the standard model having the highest existence probability by superimposing them.

  Hereinafter, the quality determination (specification) method of the initial position accuracy using the existence probability map created based on a priori knowledge will be described in detail. Here, assuming that the extraction target object exists in the actual volume data at the position of high existence probability in the existence probability map, when the standard model is initially placed near the high existence probability position, the standard It is assumed that the model is located near the actual object to be extracted, that is, the accuracy of the initial position of the standard model is high.

  FIG. 18 is a diagram showing the presence positions of a plurality of objects (organs) JB1 to JB5 of the same type obtained in advance, and FIG. 19 is a diagram showing a presence probability map MPa of the object of that type. The existence probability map MPa indicates the existence probability of an object at each position. In FIG. 19, the existence probability is expressed by the density of points, and the existence probability of a relatively black portion is higher than that of a relatively white portion. FIG. 20 is a diagram illustrating the existence probability maps MPa, MPb, and MPc for the three extraction target objects, and FIG. 21 is a diagram illustrating a state in which the standard model and the existence probability map for the three extraction target objects are superimposed. It is.

  First, the existence probability map will be described.

  The existence probability map MPa shown in FIG. 19 is obtained based on a priori data regarding the existence positions (see FIG. 18) of a plurality of objects of the same type. Specifically, in the existence probability map MPa of FIG. 19, as can be seen from comparison with FIG. 18, the existence probability is high at a position where a large number of objects among the plurality of objects JB1 to JB5 of the same type exist, The existence probability is low at a position where only a small number of objects among the plurality of objects JB1 to JB5 exist. Here, for simplification, five objects are illustrated in FIG. 18, but actually, it is preferable to generate the existence probability map MPa for the object of that type based on a larger number of objects. Further, the existence probability map MPa is shown in a planar manner in FIG. 19 and the like, but actually, a three-dimensional map (stereoscopic map) showing the existence probability for each point (pixel) in the three-dimensional space. ).

  Then, the existence probability maps obtained as described above are obtained for each of a plurality of extraction target objects, and each of the existence probability maps is obtained with the initial position accuracy of each standard model arranged at the initial position in the volume data. Used to determine (specify) the quality of

  For example, a case where an individual model is acquired using three objects (first object AA2, second object BB2, and third object CC2) as extraction target regions will be described.

  In this case, as shown in FIG. 20, the existence probability maps MPa, MPb, and MPc in the three-dimensional space are acquired for each of the three objects.

  After that, as shown in FIG. 21, each standard model SA2, SB2, SC2 arranged in the volume data by manual operation of the user (user) etc. is superposed on each existence probability map MPa, MPb, MPc. Then, for each standard model SA2, SB2, SC2, a value obtained by dividing the sum of the pixel existence probabilities by the number of pixels (that is, an average value of the existence probabilities) is calculated.

  As described above, in this modification, it is assumed that the extraction target object exists in the actual volume data at the position of the high existence probability in the existence probability map, and each standard placed initially is based on the assumption. The standard model (for example, SB2) having the highest average probability of existence among the models SA2, SB2, and SC2 is determined to have the highest initial position accuracy of the standard model among the three standard models SA2, SB2, and SC2. Is done.

  Then, optimization of the evaluation function is executed from the standard model (for example, SB2) having the highest initial position accuracy, and then evaluation functions of other standard models (for example, SA2, SC2) are optimized.

  As described above, it is possible to execute the determination of the initial position accuracy using the existence probability map, and sequentially generate individual models from the standard model determined to have the highest initial position accuracy.

  As described above, by identifying a standard model with high initial position accuracy based on the initial position of each standard model and executing optimization of the evaluation function (transformation of the standard model) from the standard model, The standard model can be sequentially deformed with the positional relationship with the extracted object as a constraint.

  In the above embodiment, the mutual energy He is expressed by the energy stored in the virtual spring provided between the standard models, but is not limited to this. Specifically, energy due to repulsion generated between each adjacent point of a certain standard model and each corresponding adjacent point of another standard model may be virtually considered.

  FIG. 22 is a schematic diagram showing the distance between adjacent points corresponding to each standard model.

  For example, as shown in FIG. 22, if the distance between each adjacent point HA of the standard model SA and each corresponding adjacent point (HB and HC) of the other standard model is represented by RABn and RACm, respectively, The energy He is expressed as in Expression (13).

  However, Kg represents the coefficient of restitution, Nab represents the number of sets of adjacent points with the model SB, and Nac represents the number of sets of adjacent points with the model SC. The subscripts n and m indicate that they relate to the nth and mth adjacent points, respectively.

  Moreover, in the said embodiment, although the energy by the virtual spring between the control points Cj was considered as internal energy Ge, it is not limited to this. FIG. 23 is a diagram illustrating angles formed by straight lines connecting the control points. For example, as shown in FIG. 23, a function that maintains an angle θr formed by a straight line connecting the control points may be the internal energy Ge.

  Moreover, in the said embodiment, although each standard model (SA, SB, SC) was acquired from the model storage part 14 in 1 A of area | region extraction apparatuses (1B), it is not limited to this. For example, you may make it acquire via networks, such as LAN and the internet, from the model storage part provided in the exterior of area | region extraction apparatus 1A (1B).

It is a figure which shows the outline | summary of the area | region extraction apparatus which concerns on 1st Embodiment of this invention. It is a block diagram which shows the various functions of an area | region extraction apparatus. It is a figure which shows the three-dimensional image acquired by X-ray CT apparatus, and the slice image of a predetermined position. It is a flowchart which shows the whole operation | movement of an area | region extraction apparatus. It is a figure which shows three objects which exist in a certain volume data in two dimensions. It is a figure which shows a mode that the standard model regarding three extraction target objects generally exists in two dimensions. It is a flowchart which shows the model fitting process of the area | region extraction apparatus which concerns on 1st Embodiment. It is a figure which shows the extraction target object contained in a three-dimensional image (measurement data). It is a figure which shows notionally a mode that a standard model is moved to an initial position. It is the schematic diagram which expanded and displayed the area | region RQ (refer FIG. 9) centering on the boundary point Q1 of an extraction target object. It is a figure which shows the slice images PQ1, PQ2, and PQ4. It is the schematic diagram which connected between control points with the virtual spring. It is the schematic diagram which connected between the adjacent points which each standard model respond | corresponds with the virtual spring. It is a figure which shows a mode that the shape of each standard model changed by the model deformation process. It is a flowchart which shows the model fitting process of the area | region extraction apparatus which concerns on 2nd Embodiment. It is a figure which shows the operation state performed with the area | region extraction apparatus which concerns on 2nd Embodiment. It is a figure which shows three objects (three-dimensional object) which exist in a certain volume data. It is a figure which superimposes and shows the existing position of the several object (organ) of the same kind calculated | required beforehand. It is a figure which shows the presence probability map of an extraction target object. It is a figure which shows the existence probability map regarding each of three extraction object. It is a figure which shows a mode that the standard model and presence probability map regarding three extraction object were piled up. It is a schematic diagram which shows the distance between the adjacent points to which each standard model respond | corresponds. It is a figure which shows the angle | corner which the straight line which connects each control point makes.

Explanation of symbols

AA First object Cj Control point RQ region SA Standard model of first object SD Slice image TD1 Measurement data SPR Virtual spring

Claims (11)

A modeling device for modeling an organ existing in measurement data as an object ,
Means for obtaining a first standard model that is a standard model related to the first object in the measurement data and a second standard model that is a standard model related to the second object in the measurement data When,
A model deforming means for deforming the first standard model so as to optimize the first evaluation function using a model fitting technique, and finally creating a model relating to the first object;
With
Said first evaluation function, include the first due to the proximity of the said first standard model by deformation of a standard model and the second standard model, elements the evaluation value is degraded A featured modeling device.   The modeling device according to claim 1,
  The modeling device is characterized in that the element is represented by a function of energy stored in a virtual spring set between the first standard model and the second standard model.   The modeling device according to claim 2, wherein
  The evaluation value deteriorates as the energy stored in the virtual spring increases,
  The energy stored in the virtual spring increases as the first standard model and the second standard model approach due to the deformation of the first standard model, and the relative displacement between both models increases.
  The energy stored in the virtual spring increases when the first standard model and the second standard model are separated by the deformation of the first standard model and the relative displacement between the two models increases. A featured modeling device.   The modeling device according to any one of claims 1 to 3,
  The model deforming means deforms the second standard model so as to optimize the second evaluation function using a model fitting technique,
  The second evaluation function includes an element whose evaluation value deteriorates due to the approach between the first standard model and the second standard model due to deformation of the second standard model. Modeling equipment.   The modeling device according to claim 4, wherein
  The model deforming means sequentially and repeatedly deforms the first standard model and the second standard model in order to optimize the first evaluation function and the second evaluation function, respectively. Modeling device characterized by   The modeling device according to claim 4 or 5,
  The model deforming means completes the deformation of one of the first standard model and the second standard model, and then deforms the other standard model. .   The modeling device according to any one of claims 1 to 6,
  Extraction means for extracting an existence area of the first object from the measurement data using a model relating to the first object;
A modeling apparatus further comprising:   The modeling device according to any one of claims 1 to 3,
  The model deforming means deforms the second standard model so as to optimize the second evaluation function using a model fitting method, and finally creates a model related to the second object,
  The modeling apparatus characterized in that the extraction means extracts a region where the second object exists from the measurement data using a model related to the second object.   A region extraction device for extracting an organ existing in measurement data as a target,
  Means for obtaining a first standard model that is a standard model related to the first object in the measurement data and a second standard model that is a standard model related to the second object in the measurement data When,
  A model deforming means for deforming the first standard model so as to optimize the first evaluation function using a model fitting technique, and finally creating a model relating to the first object;
  Means for extracting an existing region of the first object from the measurement data using the first standard model;
With
  The first evaluation function includes an element whose evaluation value deteriorates due to the approach between the first standard model and the second standard model due to deformation of the first standard model. An area extraction device.   A modeling method using an organ existing in measurement data as an object,
  A step of obtaining a first standard model that is a standard model related to a first object in the measurement data and a second standard model that is a standard model related to the second object in the measurement data. When,
  A model deformation step of deforming the first standard model so as to optimize the first evaluation function using a model fitting technique, and finally creating a model relating to the first object;
With
  The first evaluation function includes an element whose evaluation value deteriorates due to the approach between the first standard model and the second standard model due to deformation of the first standard model. Modeling method.   A program for causing a computer to execute modeling on an organ existing in measurement data,
  In the computer,
  A step of obtaining a first standard model that is a standard model related to a first object in the measurement data and a second standard model that is a standard model related to the second object in the measurement data. When,
  A model deformation step of deforming the first standard model so as to optimize the first evaluation function using a model fitting technique, and finally creating a model relating to the first object;
And execute
  The first evaluation function includes an element whose evaluation value deteriorates due to the approach between the first standard model and the second standard model due to deformation of the first standard model. Program.