JP2016073541A - Ultrasonic image processor, program and ultrasonic image processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To appropriately set a cut line on an ultrasonic image.SOLUTION: A temporary control point group (intermediate points P1, P2, ...) is set to a cross-sectional image 50. A candidate path (for example, a candidate path L1) passing through the temporary control point (for example, the intermediate point P1) selected from the temporary control point group and connecting an end point Pa of one side 60a and an end point Pb of the other side 60b is generated. A candidate path group (candidate paths L1, L2, ...) is generated by sequentially making the temporary control points being selection objects different. An evaluation value (for example, the sum total of pixel values) is calculated on the basis of the pixel value array on the candidate path for each candidate path. The best path is selected from the candidate path group on the basis of the evaluation value. The best route is used as a cut line for extracting a target tissue image or the temporary control point selected at the time of generating the best path is used as the control point for determining the cut line.SELECTED DRAWING: Figure 5

Description

  The present invention relates to an ultrasonic image processing apparatus, and more particularly to a technique for setting a cut line for separating a target tissue image from other tissue images in an ultrasonic image.

  In the medical field, three-dimensional ultrasonic diagnosis is spreading. For example, in obstetrics, ultrasonic waves are transmitted / received to / from a three-dimensional space including a fetus in the mother body, thereby acquiring volume data. In general, a three-dimensional region of interest (3D-ROI) is set for volume data, and a three-dimensional image of the fetus is formed by rendering processing on the data in the three-dimensional region of interest.

  In the three-dimensional region of interest, if uterine wall data exists on the near side of the fetus data (on the projection viewpoint side that is the rendering origin), the uterine wall is imaged and the fetal image is hidden in the uterine wall image. The problem arises. Therefore, it is desirable to set the three-dimensional region of interest so that fetal data is included in the three-dimensional region of interest and uterine wall data is not included in the three-dimensional region of interest. Note that the concept of a uterine wall image that is not required for fetal imaging includes a placenta image.

  Generally, the three-dimensional region of interest has a rendering start surface as an upper surface. The rendering start surface has a function of separating an imaging target tissue and a non-target tissue, and may be referred to as a cut surface or a clipping surface, for example. It is desirable that the rendering start surface is set so that the rendering start surface is located between fetal data and uterine wall data as much as possible. Specifically, when an amniotic fluid image exists between the fetal image and the uterine wall image, the upper side (cut line) indicating the cross section of the three-dimensional region of interest on the representative cross section of the volume data is the amniotic fluid image. It is desirable to define the shape of the cut line so that it stays within. That is, it is desirable to determine the shape of the cut line so that the fetal image and the uterine wall image are not cut as much as possible. When it is desired to image the face of the fetus, the face of the fetus corresponds to the target tissue (target tissue for imaging), and the tissue other than the face of the fetus corresponds to the non-target tissue for imaging. For example, when there are fetal hands and feet in front of the fetal face, they are part of the fetus, but the images of the hands and feet correspond to the images to be separated from the fetal face image.

  In the device described in Patent Document 1, a cut surface as a convex surface or a concave surface is set manually. Patent Document 2 discloses an apparatus that sets a region of interest based on pixel values. Patent Document 3 discloses an apparatus that detects a boundary between a fetal image and another tissue image based on volume data and sets a cut surface at the boundary.

JP 2011-83439 A JP 2011-224362 A International Publication No. 2013/027526

  As a method for forming a cut line, for example, a spline interpolation method or the like is known. In order to apply the method, the user needs to directly or indirectly set a plurality of control points (for example, two end points and at least one intermediate point) for defining the shape of the cut line to be formed. There is. However, depending on the shape of the amniotic fluid image, the cut line easily protrudes from the amniotic fluid image. Further, as the number of intermediate points is increased, the complexity of setting control points for the user increases.

  An object of the present invention is to enable an appropriate setting of a cut line on an ultrasonic image. Alternatively, it is to reduce a user's operation burden required for setting a cut line.

  An ultrasonic image processing apparatus according to the present invention is an ultrasonic image processing apparatus that processes an ultrasonic image generated by transmitting / receiving ultrasonic waves to a transmission / reception region including a tissue of interest. Temporary control point setting means for setting a temporary control point group with a spatial spread between the one side and the other side, and at least one temporary control point selected from the temporary control point group while different selection targets Candidate route generating means for generating a candidate route group connecting the one side and the other side by generating candidate routes based on control points, and each candidate route as a pixel value string on the candidate route And a best route selecting means for selecting a best route from the candidate route group by evaluating based on the cut line for extracting a target tissue image based on the best route or the cut. Control points are established to define the line, characterized in that.

  In the above configuration, the temporary control point setting means sets a plurality of temporary control points between one side and the other side on the ultrasonic image. As a result, a plurality of temporary control points may be set, and they do not necessarily have to be set in advance at the same time. The candidate route generation means generates a plurality of candidate routes by selectively using them. Each candidate path is evaluated based on a pixel value sequence on the candidate path, that is, a plurality of pixel values. The individual pixel values that make up an ultrasound image generally vary depending on the tissue that it represents. Therefore, if the pixel value sequence is evaluated, it is possible to estimate which tissue the candidate route passes through. That is, by evaluating each candidate route based on each pixel value sequence, for example, a specific route that passes the longest in a specific tissue image (for example, an amniotic fluid image present around the fetal image) can be specified. It becomes possible. For example, as the evaluation value, a sum of pixel values on the candidate path, a variance value, or the like is used. Of course, other values may be used as evaluation values. As an example, the sum of pixel values is used as an evaluation value, and a candidate route having the smallest evaluation value is selected as the best route from among a plurality of candidate routes. As a result, a candidate route passing through an image having a relatively small sum of pixel values is selected as the best route. The cut line is a line that is set for the purpose of separating a target tissue image (for example, fetal image) from an imaging disturbing tissue image (for example, uterine wall image) in an ultrasonic image. Therefore, when another tissue image (for example, amniotic fluid image) exists between the target tissue image and the disturbing tissue image, the target tissue image is set by setting a cut line in the other tissue image. And the disturbing tissue image can be separated. According to the above configuration, it is possible to easily set the cut line in the amniotic fluid image. This makes it possible to properly separate the fetal image and the uterine wall image. Moreover, since the cut line or control point is set, the troublesome setting of the cut line or control point for the user is eliminated. In the above configuration, another tissue image (for example, amniotic fluid image) may be extracted from the ultrasound image, and a temporary control point group may be set in the tissue image.

  In the above configuration, the one side and the other side are relative concepts corresponding to the one end group side and the other end group side of the candidate route group. The one side and the other side are, for example, a left region and a right region that are separated from each other in the ultrasonic image. The spatial spread is, for example, at least a spread in a direction orthogonal to the direction connecting the one side and the other side.

  The ultrasonic image processing apparatus may be configured by an ultrasonic diagnostic apparatus, may be configured by a computer that processes data acquired by the ultrasonic diagnostic apparatus, or may be configured by another apparatus. Good.

  Preferably, the best route selection means includes means for calculating an evaluation value based on the pixel value sequence for each candidate route, and a plurality of evaluation values calculated for a plurality of candidate routes constituting the candidate route group. And means for selecting the best route by specifying a best evaluation value from among them.

  Preferably, each evaluation value is a sum of a plurality of pixel values constituting each pixel value column, and the best evaluation value is a minimum value among the plurality of evaluation values.

  Preferably, the target tissue is a fetus, and the cut line is a line for spatially separating fetal data and uterine wall data. When the fetal face corresponds to the tissue of interest, the data of the tissue other than the fetal face becomes a separation target from the face data. For example, hand and foot data corresponds to data to be separated from face data.

  Preferably, each of the candidate routes is generated based on the one end point designated on the one side, the other end point designated on the other side, and the selected temporary control point. Is generated. The one end point and the other end point may be points designated by the user, or may be points fixed at predetermined positions.

  Preferably, the candidate route generation means selects one temporary endpoint selected from the one temporary endpoint sequence set on the one side and the other temporary endpoint sequence set on the other side. Each candidate route is generated based on the other temporary end point that has been set and the selected temporary control point. In this configuration, one temporary end point and the other temporary end point are also selected and a candidate route is generated.

  Preferably, the temporary control point setting means sets at least a plurality of points spread in a direction perpendicular to a direction connecting the one side and the other side on the ultrasonic image as the temporary control point group. To do. The temporary control point group is, for example, a point group distributed one-dimensionally or two-dimensionally. The intervals between the control points may be equal intervals or non-equal intervals. Those intervals may be arbitrarily changed by the user.

  Desirably, the temporary control point setting means sets a plurality of points on the grid set in the ultrasonic image as the temporary control point group. The intervals between the points on the lattice may be equal intervals or non-equal intervals. Those intervals may be arbitrarily changed by the user.

  The program according to the present invention provides a computer that processes an ultrasonic image generated by transmitting and receiving ultrasonic waves to a transmission / reception region including a tissue of interest, between one side and the other side of the ultrasonic image. A candidate route based on at least one temporary control point selected from the temporary control point group while differently selecting the temporary control point setting means for setting the temporary control point group with a spatial extent to By generating candidate path generation means for generating a candidate path group connecting between the one side and the other side, and evaluating each candidate path based on a pixel value sequence on the candidate path, Functioning as a best route selecting means for selecting the best route from the candidate route group, and based on the best route, a cut line for extracting a target tissue image or the cut line is defined. Control point is determined, characterized in that.

  The ultrasonic image processing method according to the present invention includes a step of setting a candidate route group for at least one frame data in volume data acquired from a three-dimensional transmission / reception region in a living body, Identifying a best route from the candidate route group by evaluating based on a pixel value sequence on the candidate route, and setting a three-dimensional region of interest in the volume data based on the best route; And a step of forming a three-dimensional ultrasound image based on partial volume data in the three-dimensional region of interest.

  According to the present invention, it is possible to appropriately set a cut line on an ultrasonic image. Or it becomes possible to reduce the operation burden of the user required for the setting of a cut line.

1 is a block diagram showing an overall configuration of an ultrasonic diagnostic apparatus according to an embodiment of the present invention. It is a figure which shows an example of a cross-sectional image. It is a figure which shows an example of a cross-sectional image. FIG. 6 is a diagram illustrating an example of a candidate route group according to the first embodiment. 6 is a diagram illustrating an example of a cut line according to Embodiment 1. FIG. It is a figure for demonstrating an example of an attention pixel and a surrounding pixel. It is a figure which shows an example of the histogram of a pixel value. It is a figure which shows an example of the candidate route group which concerns on the modification 1. FIG. It is a figure which shows an example of the candidate route group which concerns on the modification 2. FIG. It is a conceptual diagram which shows an example of a three-dimensional region of interest. It is a conceptual diagram which shows another example of a three-dimensional region of interest. It is a conceptual diagram which shows another example of a three-dimensional region of interest. It is a figure which shows an example of the candidate route group which concerns on Example 2. FIG. It is a figure which shows an example of the candidate route group which concerns on Example 2. FIG. It is a figure which shows an example of a control point. It is a figure which shows an example of the cut line which concerns on Example 2. FIG. FIG. 10 is a diagram for explaining a modification of the second embodiment.

  FIG. 1 shows an embodiment of an ultrasonic diagnostic apparatus as an ultrasonic image processing apparatus according to the present invention. FIG. 1 is a block diagram showing the overall configuration. This ultrasonic diagnostic apparatus is used in the medical field and has a function of forming a three-dimensional image of tissue in a living body by transmitting and receiving ultrasonic waves. In the present embodiment, as an example, the tissue to be imaged is a fetus. Of course, other tissues may be imaged.

  The probe 10 is a transducer that transmits and receives ultrasonic waves. In the present embodiment, the probe 10 has a 2D array transducer. The 2D array vibrator is formed by two-dimensionally arranging a plurality of vibration elements. An ultrasonic beam is formed by the 2D array transducer, and the ultrasonic beam is scanned two-dimensionally. As a result, a three-dimensional space 12 is formed as a three-dimensional echo data capturing space. Alternatively, the probe 10 may incorporate a 1D array transducer and a scanning mechanism that mechanically scans the 1D array transducer. The scanning surface may be formed by electronic scanning of the ultrasonic beam by the 1D array transducer, and the scanning surface may be mechanically scanned. The three-dimensional space 12 is also formed by such a method. As electronic scanning methods, electronic sector scanning, electronic linear scanning, and the like are known. When performing ultrasonic diagnosis of the fetus, the probe 10 is brought into contact with the surface of the abdomen of the mother, and ultrasonic waves are transmitted and received in this state.

  The transmission / reception unit 14 functions as a transmission beamformer and a reception beamformer. At the time of transmission, the transmission / reception unit 14 supplies a plurality of transmission signals having a fixed delay relationship to the plurality of vibration elements of the probe 10. Thereby, an ultrasonic transmission beam is formed. At the time of reception, the reflected wave from the living body is received by the probe 10, whereby a plurality of reception signals are output from the probe 10 to the transmission / reception unit 14. The transmission / reception unit 14 performs phasing addition processing on a plurality of reception signals, and thereby outputs beam data as reception signals after phasing addition. It should be noted that techniques such as transmission aperture synthesis may be used in ultrasonic transmission / reception.

  Signal processing such as detection, logarithmic compression, and coordinate conversion is applied to the beam data by the signal processing unit 16. The beam data after the signal processing is stored in the 3D memory 18. Of course, beam data that is not subjected to such processing may be stored in the 3D memory 18. The above processing may be performed when reading the beam data.

  The 3D memory 18 has a storage space corresponding to a three-dimensional space as a transmission / reception space. The 3D memory 18 stores volume data as an echo data aggregate acquired from the three-dimensional space 12. The volume data is actually constituted by coordinate conversion and interpolation processing for a plurality of beam data. Coordinate conversion for the beam data may be performed when reading the beam data.

  The three-dimensional image forming unit 20 reads the volume data from the 3D memory 18 and executes a rendering process on the volume data in the three-dimensional region of interest (3D-ROI) according to the rendering conditions given from the control unit 30. Thereby, a three-dimensional image is formed. The image data is output to the display processing unit 26. Various methods are known as rendering processing, and various methods can be employed. For example, an image processing method such as a volume rendering method or a surface rendering method is applied.

  The cross-sectional image forming unit 22 has a function of forming a two-dimensional cross-sectional image (B-mode tomographic image). For example, the cross-sectional image forming unit 22 has a function of forming cross-sectional image data in a cross section arbitrarily set by a user. Specifically, when coordinate information or the like of an arbitrary cross section is given from the control unit 30 to the cross section image forming unit 22, the cross section image forming unit 22 reads data corresponding to the arbitrary cross section from the 3D memory 18. The cross-sectional image forming unit 22 forms a two-dimensional cross-sectional image according to the read data. This image data is output to the display processing unit 26. The cross-sectional image forming unit 22 may form an arbitrary number of cross-sectional images designated by the user.

  In the present embodiment, the cross-sectional image forming unit 22 may form a B-mode tomographic image based on beam data acquired by transmitting and receiving ultrasonic waves to and from a two-dimensional scanning plane.

  The graphic image forming unit 24 forms graphic data displayed as an overlay on the cross-sectional image or the three-dimensional image in accordance with the graphic creation parameters supplied from the control unit 30. For example, the graphic image forming unit 24 forms data such as graphic data representing a cross section of a three-dimensional region of interest and graphic data representing a cut line. The graphic data thus formed is output to the display processing unit 26.

  The display processing unit 26 performs overlay processing of necessary graphic data on the cross-sectional image or the three-dimensional image, thereby constituting a display image. The display image data is output to the display unit 28, and one or a plurality of images are displayed in a display form according to the display mode. For example, a cross-sectional image or a three-dimensional image is displayed as a moving image in real time. The display unit 28 is configured by a display device such as a liquid crystal display.

  The control unit 30 performs operation control of each configuration shown in FIG. An input unit 32 is connected to the control unit 30. The input unit 32 includes an operation panel, and the operation panel includes a keyboard, a trackball, and the like. The user can use the input unit 32 to input information such as numerical values necessary for setting the three-dimensional region of interest and coordinates of an arbitrary cross section.

  The image processing unit 34 includes a temporary control point setting unit 36, a candidate route generation unit 38, an evaluation unit 40, and a region of interest setting unit 42, and has a function of setting a cut line, a cut surface, and a three-dimensional region of interest. Hereinafter, the function of each unit of the image processing unit 34 will be described.

  The temporary control point setting unit 36 sets a temporary control point group with a spatial spread between one side and the other side in the cross-sectional image. The one side and the other side correspond to, for example, a left region and a right region that are separated from each other in the ultrasonic image. For example, the temporary control point setting unit 36 acquires cross-sectional image data from the cross-sectional image forming unit 22 and sets a temporary control point group for the cross-sectional image. Alternatively, the temporary control point setting unit 36 may read data corresponding to the cross section from the 3D memory 18 and set a temporary control point group for the data. For example, the temporary control point setting unit 36 sets a grid on the cross-sectional image, and sets a plurality of points on the grid as a temporary control point group. Each temporary control point may be set at equal intervals or may be set at non-equal intervals. Alternatively, the temporary control point setting unit 36 may set temporary control points in a randomly distributed manner.

  The candidate route generation unit 38 generates a candidate route that connects one side and the other side in the cross-sectional image based on the temporary control points selected from the temporary control point group. For example, the candidate route generation unit 38 generates a candidate route that passes through one or more selected temporary control points and connects between one side and the other side. The candidate route generation unit 38 generates a plurality of candidate routes by changing the selected temporary control points. For example, the candidate route generation unit 38 generates a route connecting the one end point designated on one side and the other end point designated on the other side as a candidate route. Alternatively, the candidate path generation unit 38 selects one temporary endpoint selected from one temporary endpoint sequence set on one side and the other selected from the other temporary endpoint sequence set on the other side. A route connecting between the temporary end points of the two may be generated as a candidate route.

  The evaluation unit 40 evaluates the candidate route based on the pixel value sequence on the candidate route. The evaluation unit 40 selects the best route from a plurality of candidate routes by evaluating each candidate route. The selected best path is used as a cut line. Alternatively, the one or more temporary control points that are the basis for generating the best path are used as one or more control points that define the cut line. For example, the evaluation unit 40 calculates an evaluation value based on a pixel value sequence for each candidate route, and selects the best route by specifying the best evaluation value from a plurality of evaluation values. Specifically, the evaluation unit 40 calculates the sum of a plurality of pixel values (for example, luminance values and echo values) on the candidate path as an evaluation value. As another example, the evaluation unit 40 may calculate a variance value of pixel values on the candidate path as an evaluation value. As yet another example, the evaluation unit 40 may calculate an evaluation value using pixels around the pixels on the candidate path. As yet another example, the evaluation unit 40 may calculate a histogram of pixel values on the candidate path as an evaluation value. The evaluation unit 40 calculates the evaluation value of each candidate route and selects the best route based on the evaluation value. For example, the evaluation unit 40 selects a candidate route that minimizes the sum of pixel values as the best route.

  The region-of-interest setting unit 42 generates a cut surface including the best path (cut line) selected by the evaluation unit 40, and sets a three-dimensional region of interest including the cut surface for volume data. For example, the region-of-interest setting unit 42 generates a plurality of spline curves that pass through the cut line, thereby generating a cut surface. The cut surface may be generated by other methods. The cut surface corresponds to the start surface of the rendering process, and has a function of separating the target tissue and the non-target tissue to be imaged. The tissue on the near side (projection viewpoint side tissue in rendering processing) with respect to the cut plane corresponds to the non-target tissue for imaging, and the back side tissue (tissue opposite to the projection viewpoint) is an image. Corresponds to the target organization. The 3D image forming unit 20 applies rendering processing to the volume data in the 3D region of interest. Thereby, a three-dimensional image in the three-dimensional region of interest is formed.

  FIG. 2 shows an example of an image displayed on the display unit 28. The cross-sectional image 50 is, for example, a cross-sectional image on the central scanning plane of the three-dimensional space 12. The cross-sectional image 50 is a cross-sectional image representing an XY cross-section, and the cross-sectional image 50 includes a box Sxy as a graphic image. This box Sxy represents an XY cross section of the three-dimensional region of interest. The upper side of the box Sxy is a cut line Lxy. The end points Pa and Pb correspond to both end points of the cut line Lxy. The cross-sectional image 50 includes, for example, a fetal image 52, a uterine wall image 54, and an amniotic fluid image 56.

  For example, as shown in FIG. 3, the user can use the input unit 32 to change the positions of the end points Pa and Pb in the X direction and the Y direction. By changing the positions of the end points Pa and Pb, the length (width) in the X direction and the length (height) in the Y direction of the box Sxy can be changed. The width and height of the three-dimensional region of interest are changed according to the width and height of the box Sxy. Note that the positions of the end points Pa and Pb in the Y direction may be individually changed, or may be changed in conjunction with each other after being adjusted to the same height.

Example 1
Next, the cut line generation processing according to the first embodiment will be specifically described with reference to FIGS.

  FIG. 4 shows an example of a cross-sectional image, a temporary control point group, and a candidate route group. A cross-sectional image 50 shown in FIG. 4 is the same image as the cross-sectional images shown in FIGS. The temporary control point setting unit 36 sets a temporary control point group (for example, intermediate points P1, P2, P3,...) Between the one side 60a and the other side 60b in the cross-sectional image 50. The one side 60a and the other side 60b correspond to the left and right sides of the box Sxy. In the example shown in FIG. 4, the temporary control point setting unit 36 sets a grid 62 between the one side 60a and the other side 60b, and each point on the grid 62 (each point where the vertical line and the horizontal line intersect). ) Is set to the intermediate point (temporary control point). The intervals of the lattices 62 may be equal intervals or non-equal intervals. Further, the temporary control point setting unit 36 may set a plurality of intermediate points at random without setting the grid 62. In FIG. 4, only three intermediate points (intermediate points P1, P2, P3) are shown for convenience of explanation, but an intermediate point is set for each point on the grid 62.

  When the intermediate points P1, P2, P3,... Are set, the candidate route generation unit 38 generates a candidate route that passes through the intermediate point and connects the end point Pa on the one side 60a and the end point Pb on the other side 60b. . The positions of the end points Pa and Pb are specified by the user, and the end points Pa and Pb correspond to the upper end portions of the left and right sides of the box Sxy. Note that the positions of the end points Pa and Pb may be fixed. The candidate route generation unit 38 generates a plurality of candidate routes with different intermediate points. In the example illustrated in FIG. 4, the candidate route generation unit 38 generates a candidate route L1 that passes through the intermediate point P1 and connects the end point Pa and the end point Pb. As an example, the candidate route generation unit 38 generates a candidate route L1 as a spline curve by a spline interpolation calculation based on the intermediate point P1 and the end points Pa and Pb. For example, a candidate route L1 having the midpoint P1 as a vertex is generated. Similarly, a candidate route L2 connecting the end point Pa and the end point Pb is generated through the intermediate point P2, and a candidate route L3 connecting the end point Pa and the end point Pb is generated through the intermediate point P3. Candidate paths L2 and L3 are also spline curves as an example. Similarly, candidate paths are generated for other intermediate points. Note that the candidate route is not limited to a spline curve. The candidate path may be another curve such as a Bezier curve or a curve obtained by a least square method. The same applies to the following description. 4 shows only three candidate routes (candidate routes L1, L2, and L3) for convenience of explanation, a candidate route is generated for each intermediate point set on the grid 62. The

  When the candidate routes L1, L2, L3,... Are generated, the evaluation unit 40 calculates the sum of a plurality of pixel values (for example, luminance values and echo values) on the candidate route as an evaluation value of the candidate route. To do. The evaluation unit 40 may perform pixel value interpolation processing using values of neighboring pixels of the candidate path. In the example illustrated in FIG. 4, the evaluation unit 40 calculates the sum of pixel values as an evaluation value for each of the candidate paths L1, L2, L3,. Note that as the candidate path becomes longer, the number of pixels on the candidate path increases, and the sum of pixel values may increase accordingly. Therefore, the evaluation unit 40 may normalize the evaluation value (the sum of pixel values) by the number of pixels on the candidate path. Thereby, an evaluation value independent of the number of pixels, that is, an evaluation value independent of the length of the candidate path is obtained. The evaluation unit 40 may normalize the evaluation value of the candidate route included in the specific area, and may not normalize the evaluation value of the candidate route included in the other area. For example, the evaluation unit 40 may normalize the evaluation value of the candidate route included in the region above the end points Pa and Pb, and may not normalize the evaluation value of the candidate route included in the lower region. Alternatively, the evaluation unit 40 standardizes the evaluation value of the candidate route whose length is equal to or greater than the threshold, that is, the evaluation value of the candidate route whose length is equal to or greater than the threshold, and the length is less than the threshold (the number of pixels is less than the threshold). It is not necessary to normalize the evaluation value of the candidate route.

  And the evaluation part 40 selects the candidate path | route with the smallest evaluation value among candidate path | route L1, L2, L3, ... as a best path | route. For example, as shown in FIG. 5, the candidate route L1 is selected as the best route. The candidate route L1 that is the best route is used as a cut line in the rendering process. Further, the intermediate point P1 (temporary control point) from which the candidate route L1 is generated is used as a control point that defines a cut line.

  In general, the pixel values (for example, the luminance value and the echo value) of the amniotic fluid image 56 are lower than the pixel values of the fetal image 52 and the uterine wall image 54. Therefore, the evaluation value (sum of pixel values) of the candidate route passing through the amniotic fluid image 56 is relatively low, and the candidate route having the smallest evaluation value is estimated to pass through the amniotic fluid image 56. . Therefore, by calculating the evaluation value of each candidate route and selecting the candidate route having the smallest evaluation value, it is possible to select the cut line that is estimated to pass through the amniotic fluid image 56.

  As described above, according to the first embodiment, it is possible to appropriately set the cut line in the amniotic fluid image 56. That is, it is possible to set a cut line in the amniotic fluid image 56 so that the fetal image 52 and the uterine wall image 54 are not covered with the cut line as much as possible. As a result, it is possible to separate the fetal data as the imaging target tissue and the uterine wall data as the non-target tissue as much as possible. Further, the user does not need to specify a control point, and can set a cut line with a simple operation. As a result, it is possible to reduce the user's operation burden required for designating control points.

  The candidate route generation unit 38 may generate candidate routes that pass through a plurality of intermediate points. Also in this case, the evaluation value of each candidate route is calculated, and the best route is selected.

  The temporary control point setting unit 36 may extract the amniotic fluid image 56 based on the pixel value of the cross-sectional image 50 and set a plurality of temporary control points in the amniotic fluid image 56. It is assumed that the evaluation value of the candidate route passing through the temporary control point set in the amniotic fluid image 56 is relatively low. On the other hand, it is assumed that the evaluation value of the candidate route passing through the temporary control points set in the fetal image 52 and the uterine wall image 54 is relatively high. A candidate route that is assumed to have a relatively high evaluation value is unlikely to be selected as the best route. Therefore, the best route can be selected more efficiently by setting the temporary control points only in the amniotic fluid image 56 while avoiding the fetal image 52 and the uterine wall image 54. That is, compared to the case where a plurality of temporary control points are set for the entire cross-sectional image 50, it is possible to reduce the number of temporary control points that are set. As a result, it is possible to reduce the time required for generating candidate paths and calculating evaluation values.

  In the above example, the sum of the pixel values is used as the evaluation value, but other values may be used as the evaluation value. For example, a variance value of pixel values on the candidate path may be used as the evaluation value. Generally, in the amniotic fluid image 56, changes in pixel values (for example, luminance values and echo values) are relatively small. Therefore, it is assumed that the candidate path that passes through the amniotic fluid image 56 has a smaller change in pixel value on the candidate path (difference between the average value of pixel values on the candidate path and each pixel value). Therefore, it is presumed that the candidate route having the minimum variance value passes through the amniotic fluid image 56. In the example illustrated in FIG. 4, the evaluation unit 40 calculates the variance value of the pixel values on the candidate route for each of the candidate routes L1, L2, L3,. Candidate route L1) is selected as the best route. As a result, the candidate route L1 estimated to pass through the amniotic fluid image 56 is selected as the cut line.

  Another value may be used as the evaluation value. For example, each pixel on the candidate path may be a target pixel, and the evaluation value may be calculated using the values of the surrounding pixels of the target pixel. This process will be described with reference to FIG. The evaluation unit 40 sets the calculation area E on the candidate route. The calculation area E includes, for example, a target pixel e1 and peripheral pixels e2 to e25. The evaluation unit 40 calculates the sum of the values of the pixels e1 to e25 included in the calculation area E. At this time, the evaluation unit 40 changes the weighting factor for each pixel and calculates the sum of the weighted pixel values. For example, the weighting coefficient of the target pixel e1 is set to the maximum value, and the weighting coefficient is decreased as the pixel is farther from the target pixel e1. In the example shown in FIG. 6, the weighting coefficients of the peripheral pixels e2 to e9 are made smaller than the weighting coefficient of the target pixel e1, and the weighting coefficients of the peripheral pixels e10 to e25 are further reduced. The evaluation unit 40 multiplies the pixel values of the pixels e1 to e25 by a weighting coefficient, and calculates the sum of a plurality of weighted pixel values. The evaluation unit 40 sets each pixel on the candidate path as a target pixel, sequentially moves the calculation area E from the end Pa to the end Pb along the candidate path, and sets the pixel value in the calculation area E for each target pixel. Calculate the total (total in the calculation area). Then, the evaluation unit 40 calculates the total sum in the calculation area for each pixel of interest as an evaluation value. The evaluation unit 40 selects a candidate route having the smallest evaluation value as the best route. It is assumed that the candidate route having the minimum sum relating to the pixel value passes through the amniotic fluid image 56. Therefore, by selecting the candidate route having the smallest evaluation value as the best route, the candidate route presumed to pass through the amniotic fluid image 56 is selected as the best route.

  Furthermore, another value may be used as the evaluation value. For example, each pixel on the candidate path may be a target pixel, and the evaluation value may be calculated using the values of the surrounding pixels on the candidate path. Specifically, the evaluation unit 40 calculates a variance value of pixel values for pixels included in a calculation area set in advance on the candidate path with reference to the target pixel. The evaluation unit 40 sets each pixel on the candidate path as the target pixel, sequentially moves the calculation area from the end Pa to the end Pb along the candidate path, and the variance value of the pixel value in the calculation area for each target pixel Calculate (dispersion value in calculation area). Then, the evaluation unit 40 calculates the sum of the dispersion values in the calculation area for each target pixel as an evaluation value. Then, the evaluation unit 40 selects a candidate route having the smallest evaluation value as the best route. As a result, the candidate route presumed to pass through the amniotic fluid image 56 is selected as the best route.

  As yet another example, a histogram of pixel values may be used as the evaluation value. Specifically, the evaluation unit 40 calculates a ratio of the number of pixels having a pixel value equal to or smaller than a preset threshold Th in the candidate path (ratio to the total number of pixels on the candidate path), and based on the ratio To select the best route. For example, as illustrated in FIG. 7, the evaluation unit 40 counts the number of pixels (histogram) for each pixel value for each candidate path, and calculates the ratio of the number of pixels having a pixel value equal to or less than the threshold Th. Then, the evaluation unit 40 selects a candidate route having the maximum ratio as the best route. Since the fetal image 52 and the uterine wall image 54 have relatively high pixel values, it is assumed that the above ratio is relatively low in the candidate paths that pass through these images. On the other hand, since the pixel value of the amniotic fluid image 56 is relatively low, the ratio is assumed to be relatively high in the candidate route passing through the amniotic fluid image 56. Therefore, a route that is estimated to have passed through the amniotic fluid image 56 is selected by selecting a candidate route having the maximum ratio as the best route. Note that the evaluation unit 40 may calculate the ratio of the number of pixels having pixel values equal to or less than the threshold Th after binarizing the pixel values.

(Modification 1)
Next, a first modification of the first embodiment will be described with reference to FIG. In the example shown in FIG. 4 described above, the positions of the end points Pa and Pb are designated by the user or fixed. In the first modification, the end points are also handled as temporary control points.

  FIG. 8 shows an example of a cross-sectional image, a temporary control point group, and a candidate route group. The cross-sectional image 50 shown in FIG. 8 is the same image as the cross-sectional images shown in FIGS. In the cross-sectional image 50, the temporary control point setting unit 36 is provided on a row of a plurality of temporary one end points (for example, temporary one end points Pa1, Pa2, Pa3,...) Provided on one side 64a and on the other side 64b. A plurality of temporary other end points (for example, temporary other end points Pb1, Pb2,...) Are set. The temporary control point setting unit 36 sets a plurality of intermediate points (for example, intermediate points P4, P5, P6,...) Provided between the rows. For example, the temporary control point setting unit 36 sets a grid 66 for the cross-sectional image 50 and sets a temporary one end point, a temporary other end point, and an intermediate point at points on the grid 66. The one side 64 a and the other side 64 b correspond to the left and right sides of the lattice 66. The intervals of the lattice 66 may be equal intervals or non-equal intervals. Further, the temporary control point setting unit 36 may set a plurality of points randomly distributed without setting the grid 66. Specifically, the provisional one end points Pa1, Pa2, Pa3,... And the provisional other end points Pb1, Pb2,... May not be arranged in rows, but may be randomly distributed. . Intermediate points P4, P5, P6,... May also be randomly distributed. For convenience of explanation, FIG. 8 shows three temporary one end points (temporary one end points Pa1, Pa2, Pa3), two temporary other end points (temporary other end points Pb1, Pb2), and three intermediate points (intermediate points). P4, P5, and P6) are shown, but provisional one end points are set for each point on one side 64a, and provisional other end points are set for each point on the other side 64b. Is set to the intermediate point.

  When the temporary one end point, the temporary other end point, and the intermediate point are set, the candidate route generation unit 38 generates a candidate route that passes through the intermediate point and connects the temporary one end point and the temporary other end point. The candidate route generation unit 38 generates a plurality of candidate routes by changing the temporary one end point, the temporary other end point, and the intermediate point. In the example illustrated in FIG. 8, the candidate route generation unit 38 generates a candidate route L4 that passes through the intermediate point P4 and connects the temporary one end point Pa2 and the temporary other end point Pb1. As an example, the candidate route generation unit 38 generates a candidate route L4 as a spline curve by a spline interpolation calculation based on the intermediate point P4, the temporary one end point Pa2, and the temporary other end point Pb1. For example, a candidate route L4 having the midpoint P4 as a vertex is generated. Similarly, a candidate route L5 that passes through the intermediate point P5 and connects the temporary one end point Pa1 and the temporary other end point Pb1 is generated, passes through the intermediate point P6, and a candidate route L6 that connects the temporary one end point Pa3 and the temporary other end point Pb2 Generated. Candidate paths L5 and L6 are also spline curves as an example. Similarly, candidate paths are generated for other intermediate points, provisional one end point, and provisional other end point.

  When the candidate route is generated, the evaluation unit 40 calculates the evaluation values of the candidate routes L4, L5, L6,. Similar to the above-described example, the evaluation value is a sum of pixel values, a variance value, a value using values of surrounding pixels, a value using a histogram, or the like. Then, the evaluation unit 40 selects the best route from the candidate routes L4, L5, L6,. For example, when the sum of the pixel values is used as the evaluation value, the evaluation unit 40 selects a candidate route having the smallest evaluation value as the best route. For example, the candidate route L4 (route indicated by a solid line) is selected as the best route. Candidate route L4 is used as a cut line. Alternatively, the intermediate point P4, the provisional one end point Pa2, and the provisional other end point Pb1 that are the sources for generating the candidate route L4 are used as control points that define the cut line. According to the first modification, the end point is also a control point selection target, so that a more appropriate cut line can be set.

  Also in the first modification, the candidate route generation unit 38 may generate a candidate route that passes through a plurality of intermediate points. The temporary control point setting unit 36 may extract the amniotic fluid image 56 based on the pixel value of the cross-sectional image 50 and set a plurality of intermediate points in the amniotic fluid image 56.

(Modification 2)
Next, with reference to FIG. 9, the modification 2 of Example 1 is demonstrated. In the first modification, both end points (temporary one end point and provisional other end point) of the candidate route exist, but the candidate route may be generated based on a plurality of temporary control points without the end points. . For example, the candidate route generation unit 38 generates a candidate route that passes through any three temporary control points.

  FIG. 9 shows an example of a cross-sectional image, a temporary control point group, and a candidate route group. The cross-sectional image 50 shown in FIG. 9 is the same image as the cross-sectional images shown in FIGS. The temporary control point setting unit 36 sets a plurality of temporary control points (for example, temporary control points P7, P8, P9,...) In the cross-sectional image 50. For example, the temporary control point setting unit 36 sets a grid 68 for the cross-sectional image 50 and sets a temporary control point at each point on the grid 68 (each point where the vertical line and the horizontal line intersect). The intervals of the lattices 68 may be equal intervals or non-equal intervals. Further, the temporary control point setting unit 36 may set a plurality of temporary control points in a random manner without setting the grid 68. The temporary control point setting unit 36 may extract the amniotic fluid image 56 from the cross-sectional image 50 and set a plurality of temporary control points in the amniotic fluid image 56.

  When a temporary control point is set, the candidate route generation unit 38 generates a candidate route that passes through any three temporary control points. For example, the candidate route generation unit 38 generates a candidate route L7 that passes through the temporary control points P7, P8, and P9. As an example, the candidate route generation unit 38 calculates a spline curve passing through the temporary control points P7, P8, and P9 by a spline interpolation calculation based on the temporary control points P7, P8, and P9. This spline curve is the candidate path L7. For example, a candidate route L7 having the intermediate temporary control point P8 as a vertex is generated. At this time, the candidate route generation unit 38 extends the candidate route L7 outside the temporary control points P7 and P9. For example, the candidate route generation unit 38 calculates a curve outside the temporary control points P7 and P9 by applying a known extrapolation method based on a spline curve passing through the temporary control points P7, P8, and P9. . Similarly, a candidate route L8 that passes through temporary control points P10, P11, and P12 is generated, and a candidate route L9 that passes through temporary control points P13, P14, and P15 is generated. Candidate paths L8 and L9 are also spline curves as an example. Similarly, candidate paths are generated for other temporary control points. The candidate route generation unit 38 may generate a candidate route that passes through four or more temporary control points.

  When the candidate route is generated, the evaluation unit 40 calculates the evaluation values of the candidate routes L7, L8, L9,. This evaluation value is, for example, the sum of pixel values as in the above-described example. Then, the evaluation unit 40 selects the best route from the candidate routes L7, L8, L9,. For example, the candidate route L7 (route indicated by a solid line) is selected as the best route. Candidate route L7 is used as a cut line. Further, the temporary control points P7, P8, and P9 that are the sources for generating the candidate route L7 are used as control points that define the cut line. The candidate route generation unit 38 may generate candidate routes that pass through four or more temporary control points. The temporary control point setting unit 36 may extract the amniotic fluid image 56 based on the pixel value of the cross-sectional image 50 and set a plurality of temporary control points in the amniotic fluid image 56.

  As described above, when the best path (cut line) is selected according to the first embodiment or the first and second modifications, the region-of-interest setting unit 42 generates a cut surface including the cut line, and a tertiary including the cut surface. Set the original region of interest.

  Here, an example of cut surface generation processing will be described with reference to FIG. FIG. 10 shows a three-dimensional region of interest V1. The three-dimensional region of interest V1 is a region that virtually exists in the data processing space. That is, the three-dimensional region of interest V1 only exists numerically as a rendering process range condition, and such a shape is not actually generated. However, in the description of the present embodiment, it is assumed that the figure can visually recognize the three-dimensional region of interest V <b> 1 in order to help the understanding.

  In FIG. 10, the eight corners of the three-dimensional region of interest V1 are represented by reference symbols a1, a2, a3, a4, a5, a6, a7, and a8. The midpoints of the sides are represented by reference signs a12, a23, a34, a14, a56, a67, a78, and a58.

  The box Sxy shown in FIG. 2 corresponds to the center XY cross section in the Z direction in the three-dimensional region of interest V1. Accordingly, the cross-sectional images 50 shown in FIGS. 2 to 5 and FIGS. 8 and 9 represent images in the XY cross-section at the center in the Z direction. The box Sxy is not limited to the center XY cross section in the Z direction, and may correspond to a cross section at an arbitrary position in the Z direction. The position may be designated by the user. For example, it is assumed that the cross section is set in the central cross section of the fetal face data. Of course, the cross section may be set at a place other than this.

  The sizes of the three-dimensional region of interest V <b> 1 in the X direction, the Y direction, and the Z direction are determined by the user using the input unit 32. For example, the respective sizes in the X direction and the Y direction of the three-dimensional region of interest V1 are determined by the respective sizes in the X direction and the Y direction of the box Sxy shown in FIG. Each size in the Z direction is determined according to a parameter input from the input unit 32.

  The generation process of the cut surface 70 will be described in detail. Here, it is assumed that the candidate route L1 shown in FIG. 5 is selected as the best route. The candidate route L1 is a route that passes through the intermediate point P1 (control point) and connects the midpoints a14 and a23 (corresponding to the end points Pa and Pb).

  When the candidate route L1 is determined as described above, the region-of-interest setting unit 42 has a plurality of spline curves (for example, spline curves d1, d2,...) Between the sides b2 and b4 parallel to the Z direction. Are generated sequentially. Specifically, the region-of-interest setting unit 42 sets a plurality of end points c1, c2, c3, c4,... At equal intervals for the sides b1, b3 parallel to the X direction, and similarly candidates Also on the route L1, a plurality of passing points are set at equal intervals in the X direction. Then, the region-of-interest setting unit 42 generates a spline curve by performing a spline interpolation calculation using two end points and one passing point at each position in the X direction. In the example illustrated in FIG. 10, the spline curve d1 is a curve that connects the end points c1 and c2 and passes through a passing point on the candidate route L1. The spline curve d2 is a curve that connects the end points c3 and c4 and passes through a passing point on the candidate route L1. Generating such a spline curve at each position in the X direction results in the formation of a curve array, thereby forming the cut surface 70. The cut surface 70 is actually composed of three-dimensional shape data.

  As described above, the region-of-interest setting unit 42 generates the cut surface 70, and sets the three-dimensional region of interest V1 including the cut surface 70 for the volume data. The 3D image forming unit 20 forms a 3D image in the 3D region of interest V1 by applying a rendering process to the volume data in the 3D region of interest V1. The cut surface 70 corresponds to a rendering start surface. In the example shown in FIG. 10, each ray in the rendering process is set along the Y direction, and the projection viewpoint is above the Y direction. Therefore, the tissue on the projection viewpoint side with respect to the cut plane 70 is not imaged, but the tissue on the side opposite to the projection viewpoint (tissue in the three-dimensional region of interest V1) is imaged. In the rendering process, each voxel data is generated by an interpolation process by referring to a plurality of echo data existing in the vicinity. When performing the interpolation process, data outside the three-dimensional region of interest V1 may be referred to.

  When the three-dimensional image is formed, the three-dimensional image is displayed on the display unit 28. For example, the cross-sectional image 50 and the three-dimensional image are displayed side by side on the display unit 28.

  Next, another example of the cut surface generation process will be described with reference to FIG. FIG. 11 shows a three-dimensional region of interest V2. The three-dimensional region of interest V <b> 2 has a cut surface 72. The cut surface 72 is a surface generated by a process different from that of the cut surface 70 shown in FIG. The configuration other than the cut surface in the three-dimensional region of interest V2 is the same as the configuration of the three-dimensional region of interest V1.

  In this example, the processing (cut line generation processing) by the temporary control point setting unit 36, the candidate route generation unit 38, and the evaluation unit 40 is a cross-sectional image (XY cross-sectional image) at each position set at equal intervals in the Z direction. Is executed against. For example, the image processing unit 34 sets a plurality of cross sections Sxya, Sxyb, Sxyc,... At equal intervals in the Z direction. These cross sections represent XY cross sections. Then, the cut line generation process is executed on the cross-sectional images in each of the cross sections Sxya, Sxyb, Sxyc,. Thereby, candidate routes La, Lb, Lc,... Are selected as the best routes. For example, when the cut line generation process is executed on the cross-sectional image of the cross-section Sxya, the candidate route La is selected as the best route. The candidate route Lb is the best route in the cross-sectional image of the cross section Sxyb. The candidate path Lc is the best path in the cross-sectional image of the cross section Sxyc. By selecting such a candidate route at each position in the Z direction, a curved array is formed as a result, and the cut surface 72 is formed. Thereby, the three-dimensional region of interest V2 having the cut surface 72 is set. Each of the cross sections Sxya, Sxyb, Sxyc,... Does not have to be set at equal intervals in the Z direction. For example, each of the cross sections Sxya, Sxyb, Sxyc,... May be set at an arbitrary interval at an arbitrary position in the Z direction. As an example, where the imaging target tissue and the non-target tissue are close to each other, such as the vicinity of the fetal face, each cross-section is provided at a narrow interval, and the distance between the imaging target tissue and the non-target tissue is sufficient Then, it is assumed that each cross section is provided at a wide interval. By setting in this way, efficient and accurate processing becomes possible.

  Further, the region-of-interest setting unit 42 sets a plurality of end points c1, c2, c3, c4,... At equal intervals, similarly to the processing shown in FIG. 10, and similarly, candidate routes La, Lb, Lc, In the above, a plurality of passing points may be set at equal intervals in the X direction. Then, the region-of-interest setting unit 42 generates a spline curve by performing a spline interpolation calculation using the two end points and the passing points on each candidate route at each position in the X direction. In the example shown in FIG. 11, the spline curve d10 is a curve that connects the end points c1 and c2 and passes through the passing points on the candidate routes La, Lb, Lc,. The spline curve d11 is a curve that connects the end points c3 and c4 and passes through the passing points on the candidate routes La, Lb, Lc,. Generating such a spline curve at each position in the X direction results in the formation of a curve array, thereby forming the cut surface 72.

  Next, still another example of the cut surface generation process will be described with reference to FIG. FIG. 12 shows a three-dimensional region of interest V3. The three-dimensional region of interest V <b> 3 has a cut surface 74. The cut surface 74 is a surface generated by a process different from that of the cut surfaces 70 and 72 described above. In the three-dimensional region of interest V3, the configuration other than the cut surface is the same as the configuration of the three-dimensional regions of interest V1, V2.

  The box Sxy shown in FIG. 2 corresponds to the XY cross section at the center in the Z direction in the three-dimensional region of interest V3. Therefore, the cross-sectional image 50 shown in FIG. 2 represents an image in the center XY cross-section in the Z direction. The box Sxy is not limited to the center XY cross section in the Z direction, and may correspond to a cross section at an arbitrary position in the Z direction. Here, it is assumed that the candidate route L1 shown in FIG. 5 is selected as the best route.

  In this example, the cut line generation process is performed on cross-sectional images (images of a YZ cross-section orthogonal to the XY cross-section) at each position set at equal intervals in the X direction. Similarly to the processing shown in FIG. 10, the image processing unit 34 sets a plurality of end points c1, c2, c3, c4,... At equal intervals, and similarly in the X direction on the candidate route L1. A plurality of passing points c12, c34,... Are set at intervals. Then, with one end point as one end point and the passing point on the candidate route L1 as the other end point, a cut line generation process is executed between these end points. This selects the best path between those endpoints. This cut line generation process is executed at each position in the X direction. In the example shown in FIG. 12, the end point c1 is used as an end point on one side, and the passing point c12 on the candidate route L1 is used as an end point on the other side. Then, as shown in FIG. 4, a plurality of temporary control points are set between one side and the other side, passing through the selected temporary control points while changing the selection target, and the end point c1 and the passing point c12 A plurality of candidate routes connecting the two are generated. Then, the evaluation value of each candidate route is calculated, and the candidate route with the smallest evaluation value is selected as the best route (the best route in the Z direction). In the example shown in FIG. 12, the candidate route d1a is selected as the best route between the end point c1 and the passing point c12. By the same processing, the candidate route d1b is selected as the best route between the end point c2 and the passing point c12. The candidate route d2a is selected as the best route between the end point c3 and the passing point c34, and the candidate route d2b is selected as the best route between the end point c4 and the passing point c34. Generating such candidate paths at each position in the X direction results in the formation of a curved array, thereby forming the cut surface 74.

  When the cut surface is generated and the three-dimensional region of interest is set as described above, the three-dimensional image forming unit 20 applies the rendering process to the volume data in the three-dimensional region of interest. Thereby, a three-dimensional image in the three-dimensional region of interest is formed. As described above, since the cut line is set in the amniotic fluid image, the cut surface generated based on the cut line can be set in the amniotic fluid image. This makes it possible to appropriately separate the fetal image and the uterine wall image to form a three-dimensional image of the fetus.

(Example 2)
Next, the cut line generation processing according to the second embodiment will be specifically described with reference to FIGS. In the second embodiment, a plurality of control points that define the cut line are specified in stages.

  FIG. 13 shows an example of a cross-sectional image, a temporary control point group, and a candidate route group. The cross-sectional image 50 shown in FIG. 13 is the same image as the cross-sectional images shown in FIGS. First, the image processing unit 34 executes a first stage process. The temporary control point setting unit 36 sets a row 80a composed of a plurality of temporary control points (for example, intermediate points P20, P21, P22,...) Between the one side 60a and the other side 60b. The column 80a is a column parallel to the Y direction. That is, intermediate points P20, P21, P22,... Aligned in a line along the Y direction are set between the one side 60a and the other side 60b. The intermediate points may be arranged at equal intervals or may be arranged at non-equal intervals. As an example, the row 80a is set at an intermediate position between the one side 60a and the other side 60b. This is an example, and the row 80a may be set at an arbitrary position between the one side 60a and the other side 60b. In FIG. 13, for convenience of explanation, only three intermediate points (P20, P21, P22) are shown, but a plurality of intermediate points are set on the column 80a.

  When the intermediate points P20, P21, P22,... Are set, the candidate route generation unit 38 generates a candidate route that passes through the intermediate point and connects the end point Pa on one side 60a and the end point Pb on the other side. The positions of the end points Pa and Pb are specified by the user and correspond to the upper end portions of the left and right sides of the box Sxy as shown in FIG. Note that the positions of the end points Pa and Pb may be fixed. The candidate route generation unit 38 generates a plurality of candidate routes with different intermediate points. In the example illustrated in FIG. 13, the candidate route generation unit 38 generates a candidate route L20 that passes through the intermediate point P20 and connects the end point Pa and the end point Pb. As an example, the candidate route generation unit 38 generates a candidate route L20 as a spline curve by a spline interpolation calculation based on the intermediate point P20 and the end points Pa and Pb. For example, a candidate route L20 having the midpoint P20 as a vertex is generated. Similarly, a candidate route L21 that connects the end point Pa and the end point Pb is generated through the intermediate point P21, and a candidate route L22 that connects the end point Pa and the end point Pb is generated through the intermediate point P22. Candidate paths L21 and L22 are also spline curves as an example. Similarly, candidate paths are generated for other intermediate points.

  When the candidate routes L20, L21, L22,... Are generated, the evaluation unit 40 calculates the evaluation values of the candidate routes L20, L21, L22,. This evaluation value is, for example, the sum of pixel values as in the above-described example. Then, the evaluation unit 40 selects a provisional best route from the candidate routes L20, L21, L22,. For example, when the sum of the pixel values is used as the evaluation value, the evaluation unit 40 selects a candidate route having the smallest evaluation value as the temporary best route. For example, the candidate route L20 (route indicated by a solid line) is selected as the temporary best route.

  The evaluation unit 40 uses the intermediate point P20 from which the candidate route L20 is generated as a control point that defines the final best route (cut line). Here, the processing of the first stage is completed.

  When the intermediate point P20 used as the control point is specified as described above, the image processing unit 34 executes the second stage process. For example, as shown in FIG. 14, the temporary control point setting unit 36 includes a plurality of temporary control points (for example, intermediate points P30, P31,...) Between the intermediate point P20 (row 80a) and the one side 60a. The column 80b is set. Further, the temporary control point setting unit 36 sets a column 80c composed of a plurality of temporary control points (for example, intermediate points P40, P41,...) Between the intermediate point P20 (column 80a) and the other side 60b. . The columns 80b and 80c are columns parallel to the Y direction. That is, the intermediate points P30, P31,... Aligned in a line along the Y direction are set between the intermediate point P20 and the one side 60a. Further, intermediate points P40, P41,... Aligned in a line along the Y direction are set between the intermediate point P20 and the other side 60b. The intermediate points may be arranged at equal intervals or may be arranged at non-equal intervals. As an example, the column 80b is set at an intermediate position between the intermediate point P20 and the one side 60a, and the column 80b is set at an intermediate position between the intermediate point P20 and the other side 60b. These are merely examples, and the row 80b may be set at an arbitrary position between the intermediate point P20 and the one side 60a. Similarly, the column 80c may be set at an arbitrary position between the intermediate point P20 and the other side 60b. In FIG. 14, for convenience of explanation, only four intermediate points (P30, P31, P40, P41) are shown, but a plurality of intermediate points are set on columns 80b and 80c, respectively.

  When the columns 80b and 80c are set, the candidate route generation unit 38 generates a candidate route that passes through the intermediate point included in the column 80b and connects the end point Pa on one side 60a and the intermediate point P20 as the control point. The candidate route generation unit 38 generates a plurality of candidate routes by changing the intermediate points included in the column 80b. In the example illustrated in FIG. 14, the candidate route generation unit 38 generates a candidate route L30 that passes through the intermediate point P30 and connects the end point Pa and the intermediate point P20 (control point). As an example, the candidate route generation unit 38 generates a candidate route L30 as a spline curve by spline interpolation processing based on the intermediate point P30, the end point Pa, and the intermediate point P20. For example, a candidate route L30 having the midpoint P30 as a vertex is generated. Similarly, a candidate route L31 that passes through the intermediate point P31 and connects the end point Pa and the intermediate point P20 is generated. Candidate course L31 is also a spline curve as an example. Similarly, candidate paths are generated for the other intermediate points included in the column 80b.

  In addition, the candidate route generation unit 38 generates a candidate route that passes through the intermediate point included in the column 80c and connects the end point Pb on the other side 60b and the intermediate point P20 as the control point. The candidate route generation unit 38 generates a plurality of candidate routes by changing the intermediate points included in the column 80c. In the example illustrated in FIG. 14, the candidate route generation unit 38 generates a candidate route L40 that passes through the intermediate point P40 and connects the end point Pb and the intermediate point P20 (control point). As an example, the candidate route generation unit 38 generates a candidate route L40 as a spline curve by spline interpolation processing based on the intermediate point P40, the end point Pb, and the intermediate point P20. For example, a candidate route L40 having the midpoint P40 as a vertex is generated. Similarly, a candidate route L41 that passes through the intermediate point P41 and connects the end point Pb and the intermediate point P20 is generated. Candidate course L41 is also a spline curve as an example. Similarly, candidate paths are generated for the other intermediate points included in the column 80c.

  When a candidate route (candidate routes L30, L31,...) Connecting the end point Pa and the intermediate point P20 is generated, the evaluation unit 40 calculates each evaluation value of the candidate routes L30, L31,. . This evaluation value is, for example, the sum of pixel values as in the above-described example. Then, the evaluation unit 40 selects a provisional best route connecting the end point Pa and the intermediate point P20 from the candidate routes L30, L31,. For example, when the sum of the pixel values is used as the evaluation value, the evaluation unit 40 selects a candidate route having the smallest evaluation value as the temporary best route. For example, the candidate route L30 (route indicated by a solid line) is selected as the temporary best route.

  The evaluation unit 40 uses the intermediate point P30 from which the candidate route L30 is generated as a control point that defines the final best route (cut line).

  In addition, when a candidate route (candidate routes L40, L41,...) Connecting the end point Pb and the intermediate point P20 is generated, the evaluation unit 40 obtains the evaluation values of the candidate routes L40, L41,. Calculate. This evaluation value is, for example, the sum of pixel values as in the above-described example. Then, the evaluation unit 40 selects a provisional best route connecting the end point Pb and the intermediate point P20 from the candidate routes L40, L41,. For example, when the sum of the pixel values is used as the evaluation value, the evaluation unit 40 selects a candidate route having the smallest evaluation value as the temporary best route. For example, the candidate route L40 (route indicated by a solid line) is selected as the temporary best route.

  The evaluation unit 40 uses the intermediate point P40 from which the candidate route L40 is generated as a control point that defines the final best route (cut line).

  Here, the processing of the second stage ends. Through the processing so far, the intermediate points P20, P30, and P40 used as control points are specified.

  Similarly, the intermediate point used as the control point is specified thereafter. That is, the temporary control point setting unit 36 sets a row of a plurality of intermediate points between two adjacent control points, and a plurality of intermediate points between the one side 60a and the adjacent control points. A column composed of a plurality of intermediate points is set between the other side 60b and a control point adjacent thereto.

  For example, in the third stage of processing, as shown in FIG. 15, the temporary control point setting unit 36 sets a column 80d composed of a plurality of temporary control points between the intermediate point P30 adjacent to each other and the one side 60a. To do. In addition, the temporary control point setting unit 36 sets a row 80e including a plurality of temporary control points between the intermediate points P30 and P20 adjacent to each other. Further, the temporary control point setting unit 36 sets a column 80f including a plurality of temporary control points between the intermediate points P20 and P40 adjacent to each other. Further, the temporary control point setting unit 36 sets a row 80g including a plurality of temporary control points between the intermediate point P40 and the other side 60b adjacent to each other. In FIG. 15, the provisional control points included in each column are not shown. The columns 80d, 80e, 80f, and 80g are columns parallel to the Y direction. That is, a plurality of intermediate points arranged in a line along the Y direction are between one side 60a and the intermediate point P30, between the intermediate point P30 and the intermediate point P20, between the intermediate point P20 and the intermediate point P40, And it is set between the intermediate point P40 and the other side 60b.

  When the columns 80d, 80e, 80f, and 80g are set, the temporary best path that connects the end point Pa and the midpoint P30, the midpoint P30, and the midpoint by the processing of the candidate path generation unit 38 and the evaluation unit 40 in the same manner as described above. A temporary best path connecting the point P20, a temporary best path connecting the intermediate point P20 and the intermediate point P40, and a temporary best path connecting the intermediate point P40 and the end point Pb are selected.

  The evaluation unit 40 uses the intermediate point from which each best path is generated as a control point that defines the final best path (cut line). FIG. 16 shows the intermediate point. The intermediate points P50, P60, P70, and P80 are control points specified by the third stage process. The intermediate point P50 is a point between the end point Pa and the intermediate point P30. The intermediate point P60 is a point between the intermediate point P30 and the intermediate point P20. The intermediate point P70 is a point between the intermediate point P20 and the intermediate point P40. The intermediate point P80 is a point between the intermediate point P40 and the end point Pb.

  The temporary control point setting unit 36, the candidate route generation unit 38, and the evaluation unit 40 repeat the above processing a plurality of times. This number of times is preset. Further, the number of times may be set by the user using the input unit 32. In the above example, intermediate points P20, P30, P40, P50, P60, P70, and P80 used as control points are specified. That is, the intermediate point P20 is specified in the first stage process, the intermediate points P30, P40 are specified in the second stage process, and the intermediate points P50, P60, P70, P80 are specified in the third stage process. In this way, a plurality of intermediate points used as control points are specified in stages.

  When the control point is specified as described above, the evaluation unit 40 generates the best route (cut line) using the control point. For example, the evaluation unit 40 generates a curved or straight path that passes through all control points as the best path. As another example, the evaluation unit 40 may apply a least square method to a plurality of control points, and adopt a path formed thereby as the best path. Alternatively, the evaluation unit 40 may connect the provisional best route and adopt the route formed thereby as the best route.

  FIG. 16 shows a route L50 as the best route. This route L50 is a route generated based on a plurality of control points specified by the processing up to the third stage. As an example, the path L50 is a curve that passes through all of the intermediate points P20, P30, P40, P50, P60, P70, and P80. This curve may be a spline curve or a Bezier curve. Alternatively, a curve formed by applying the least square method may be adopted as the best path. Alternatively, a straight path that passes through all of the intermediate points P20, P30, P40, P50, P60, P70, and P80 may be adopted as the best path.

  When the best route is generated as described above, the region-of-interest setting unit 42 generates a cut surface based on the best route (cut line), and converts the three-dimensional region of interest including the cut surface into volume data. To set. Then, rendering processing is applied to the volume data in the three-dimensional region of interest, thereby forming a three-dimensional image in the three-dimensional region of interest.

  In the second embodiment, the temporary best path is partially selected between the one side 60a and the other side 60b, and a plurality of control points are specified in stages. By partially selecting the provisional best route, a plurality of provisional best routes along a more complicated shape can be selected. By using the intermediate point from which these provisional best paths are generated as a control point, it becomes possible to form a cut line having a complicated shape. As a result, a cut surface having a complicated shape can be formed. As a result, it is possible to set the cut line and the cut surface at a more appropriate position and more appropriately separate the fetal image and the uterine wall image.

  Note that the modified example 1 of the first embodiment may be applied to the second embodiment, and the candidate paths may be generated by changing the end points Pa and Pb.

  Next, a modification of the second embodiment will be described with reference to FIG. In the above example, the number of control point specifying processes is set in advance or set by the user. In the modification, the number of times is automatically set.

  FIG. 17 shows the processing results of the second stage and the third stage. That is, the intermediate points P20, P30, and P40 are specified as the second-stage processing results, and the candidate routes L30 and L40 are selected as temporary best routes. In addition, intermediate points P50, P60, P70, and P80 are specified as the processing results of the third stage.

  The evaluation unit 40 obtains an intersection between the candidate route obtained in the second stage and a column composed of a plurality of temporary control points set in the third stage. And the evaluation part 40 calculates | requires the difference of the position of the intersection and the intermediate point obtained at the 3rd step. If the position difference is equal to or greater than the threshold value, the control point specifying process is continued for the candidate route section, and the next fourth-stage process is executed for the candidate route section. If the position difference is less than the threshold, the control point specifying process for the candidate route section is terminated. In this case, it is assumed that the position of the route does not change greatly even if the control point specifying process is repeated in the section of the candidate route. Therefore, the control point specifying process is terminated.

  A specific example will be described. For example, the intersection between the candidate route L30 and the column 80d is obtained, and the difference in position between the intersection and the intermediate point P50 on the column 80d is obtained. If the difference is not less than the threshold value, the control point specifying process is continued between the end point Pa and the intermediate point P30. If the difference is less than the threshold value, it is assumed that the position of the candidate route L30 will not change greatly even if the control point specifying process is repeated between the end point Pa and the intermediate point P30, and thus the control point specifying process ends. .

  In the example shown in FIG. 17, for the intermediate points P50, P60, and P80, the position difference is less than the threshold value. Therefore, between the end point Pa and the intermediate point P30, between the intermediate point P30 and the intermediate point P20, and between the intermediate point P40 and the end point Pb, the setting process of the intermediate points P50, P60, and P80 is executed. The control point specifying process ends.

  On the other hand, for the intermediate point P70, the difference in position is equal to or greater than the threshold value. Therefore, between the intermediate point P20 and the intermediate point P40, even after the intermediate point P70 is set, the division of the sections and the control point specifying process are repeatedly executed until the position difference becomes less than the threshold value.

  According to the above processing, the number of repetitions of the control point specifying process is reduced in a simple structure portion, and the number of repetitions is increased in a complicated structure portion. Therefore, in the complicated structure portion, the control points are set more finely, and a more accurate cut line can be set.

  In the present embodiment, the whole image of the fetus corresponds to the target tissue image, but a part of the image of the fetus may correspond to the target tissue image. For example, when it is desired to image the fetal face, the fetal face corresponds to the target tissue (image target tissue), and tissues (hands and feet) other than the fetal face correspond to non-image target tissue. Applicable. In this case, the image of the fetus hand or foot corresponds to an image to be separated from the image of the fetal face. The cut line is used to separate the facial image from other tissue images (hand and foot images).

  The configuration other than the probe 10 shown in FIG. 1 can be realized by using hardware resources such as a processor and an electronic circuit, for example, and even if a device such as a memory is used in the realization, the configuration can be realized. Good. The configuration other than the probe 10 may be realized by a computer, for example. That is, all or part of the configuration other than the probe 10 may be realized by cooperation of hardware resources such as a CPU, a memory, and a hard disk included in the computer and software (program) that defines the operation of the CPU and the like. . The program is stored in a storage device (not shown) via a recording medium such as a CD or DVD, or via a communication path such as a network. As another example, the configuration other than the probe 10 may be realized by a DSP (Digital Signal Processor), an FPGA (Field Programmable Gate Array), or the like.

  DESCRIPTION OF SYMBOLS 10 Probe, 12 Three-dimensional space, 14 Transmission / reception part, 16 Signal processing part, 18 3D memory, 20 Three-dimensional image formation part, 22 Cross-section image formation part, 24 Graphic image formation part, 26 Display processing part, 28 Display part, 30 Control unit, 32 input unit, 34 image processing unit, 36 temporary control point setting unit, 38 candidate route generation unit, 40 evaluation unit, 42 region of interest setting unit.

Claims (10)

In an ultrasonic image processing apparatus for processing an ultrasonic image generated by transmitting / receiving ultrasonic waves to a transmission / reception region including a tissue of interest,
Temporary control point setting means for setting a temporary control point group with a spatial spread between one side and the other side in the ultrasonic image;
A candidate route group connecting the one side and the other side by generating a candidate route based on at least one temporary control point selected from the temporary control point group while changing the selection target Candidate route generation means for generating
Best path selection means for selecting the best path from the candidate path group by evaluating each candidate path based on a pixel value sequence on the candidate path;
Including
Based on the best path, a cut line for extracting a target tissue image or a control point that defines the cut line is determined.
An ultrasonic image processing apparatus. The ultrasonic image processing apparatus according to claim 1,
The best route selection means is:
Means for calculating an evaluation value based on the pixel value sequence for each candidate path;
Means for selecting the best route by specifying a best evaluation value from among a plurality of evaluation values calculated for a plurality of candidate routes constituting the candidate route group;
An ultrasonic image processing apparatus comprising: The ultrasonic image processing apparatus according to claim 2,
Each evaluation value is a sum of a plurality of pixel values constituting each pixel value sequence,
The best evaluation value is a minimum value among the plurality of evaluation values.
An ultrasonic image processing apparatus. The ultrasonic image processing apparatus according to any one of claims 1 to 3,
The tissue of interest is a fetus;
The cut line is a line for spatially separating fetal data and uterine wall data,
An ultrasonic image processing apparatus. The ultrasonic image processing apparatus according to any one of claims 1 to 4,
Each candidate route generation means generates each candidate route based on one end point designated on the one side, the other end point designated on the other side, and the selected temporary control point. ,
An ultrasonic image processing apparatus. The ultrasonic image processing apparatus according to any one of claims 1 to 4,
The candidate route generation means includes one temporary endpoint selected from one temporary endpoint sequence set on the one side and the other selected from the other temporary endpoint sequence set on the other side. Each of the candidate routes based on the temporary end point of the selected temporary control point and the selected temporary control point,
An ultrasonic image processing apparatus. The ultrasonic image processing apparatus according to any one of claims 1 to 6,
The temporary control point setting means sets at least a plurality of points spread in a direction perpendicular to a direction connecting the one side and the other side on the ultrasonic image as the temporary control point group.
An ultrasonic image processing apparatus. The ultrasonic image processing apparatus according to claim 7,
The temporary control point setting means sets a plurality of points on a grid set in the ultrasonic image as the temporary control point group.
An ultrasonic image processing apparatus. A computer for processing an ultrasonic image generated by transmitting and receiving ultrasonic waves to a transmission / reception region including a tissue of interest;
Temporary control point setting means for setting a temporary control point group with a spatial spread between one side and the other side in the ultrasonic image;
A candidate route group connecting the one side and the other side by generating a candidate route based on at least one temporary control point selected from the temporary control point group while changing the selection target Candidate route generation means for generating
Best path selection means for selecting the best path from the candidate path group by evaluating each candidate path based on a pixel value sequence on the candidate path;
Function as
Based on the best path, a cut line for extracting a target tissue image or a control point that defines the cut line is determined.
A program characterized by that. Setting a candidate route group for at least one frame data in volume data acquired from a three-dimensional transmission / reception region in a living body;
Identifying the best route from the candidate route group by evaluating each candidate route based on a pixel value sequence on the candidate route; and
Setting a three-dimensional region of interest in the volume data based on the best path;
Forming a three-dimensional ultrasound image based on partial volume data in the three-dimensional region of interest;
An ultrasonic image processing method comprising:

Images