JP2018157982A - Ultrasonic diagnosis apparatus and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To correct an inclination of the ellipse so that an inclination of an ellipse corresponding to an annular part of a fetus becomes close to the inclination of the annular part.SOLUTION: An ellipse corresponding to an annular part of a fetus (e.g., head part of the fetus) is computed based on a tomographic image acquired by the transmission and reception of an ultrasonic wave to and from the fetus in a mother's body. An angle correction part 138 computes an evaluation value in a reference area in each position and at each rotation angle in one axis direction while moving the center of the ellipse in the one axis direction (e.g., short axis direction) of the ellipse, and while rotating the reference area extending in the other axis direction (e.g., long axis direction) of the ellipse around the center after the movement, and corrects the angle of the ellipse with respect to the annular part using the evaluation value.SELECTED DRAWING: Figure 19

Description

  The present invention relates to an ultrasonic diagnostic apparatus, and more particularly to a technique related to fetal measurement.

  In some cases, the head or abdomen of a fetus is measured using an ultrasonic diagnostic apparatus. In the measurement, in order to increase the measurement accuracy, it is desirable that the position and region of the measurement target region (for example, the head and the abdomen) can be accurately detected.

  In the measurement of the fetal head, the tomographic image is converted into polar coordinates, and an ellipse that approximates the fetal head is calculated by applying dynamic programming to the image after the polar coordinate conversion, and measurement is performed using the ellipse. There are cases where it is performed (for example, Patent Document 1).

International Publication No. 2016/190256

  As described above, an ellipse corresponding to the fetal head may be calculated and the fetal head may be measured using the ellipse. In order to increase the accuracy of measurement using the ellipse, it is desirable to match the inclination of the ellipse with the inclination of the fetal head. This is also true for the case where the fetal abdomen is measured. In order to improve the measurement accuracy of the fetal abdomen, it is desirable to match the inclination of the ellipse corresponding to the fetal abdomen with the inclination of the fetal abdomen.

  An object of the present invention is to correct the inclination of the ellipse so that the inclination of the ellipse corresponding to the fetal annular part is close to the inclination of the annular part.

  The present invention provides an ellipse calculation means for calculating an ellipse corresponding to the fetal ring part based on a tomographic image obtained by transmitting and receiving ultrasonic waves to a fetus in a mother body, and the ellipse calculation means. While moving the calculated center of the ellipse in one axial direction of the ellipse and rotating the reference area extending in the other axial direction of the ellipse at the center after the movement, each position in the one axial direction and each An ultrasonic diagnostic apparatus comprising: an angle correction unit that calculates an evaluation value in the reference area at a rotation angle and corrects an angle of the ellipse with respect to the annular portion using the evaluation value. is there.

  The angle correction unit may correct the angle of the ellipse based on the evaluation value obtained at each position and each rotation angle and the variance of the evaluation value.

  The angle correction means moves the center of the ellipse in the one axial direction, rotates the reference area at the position of the movement destination, and calculates a sum of luminance values in the reference area at each rotation angle. Calculated as an evaluation value, calculates the sum of the luminance values in the reference area at each rotation angle for each position in the one axial direction, and calculates the luminance at each rotation angle for each position in the one axial direction. Calculate the variance of the sum of the values, determine the position in the one axial direction where the variance is the largest, determine the rotation angle at which the sum of the luminance values is the largest at that position, and correct the ellipse angle to that rotation angle May be.

  The angle correction means may apply a weighting process according to the amount of movement of the center of the ellipse in the one axial direction to the variance of the sum of the luminance values.

  The angle correction means is configured such that when the variance at each position in the one axial direction is equal to or less than a predetermined threshold, the ellipse is rotated at a rotation angle at which the sum of luminance values is maximized when the center of the ellipse is not moved. The angle may be corrected.

  The annular part is the head of the fetus, the one axial direction is the minor axis direction of the ellipse, the other axial direction is the major axis direction of the ellipse, and the angle correction means Based on this, the angle of the ellipse may be corrected by calculating the direction in which the midline of the fetal head extends.

  By applying template matching to the tomographic image, position calculation means for calculating a candidate for the center position of the fetal annular region, and using the center position candidate as the origin of the polar coordinate system, the tomographic image is converted into polar coordinates. The image processing apparatus further includes conversion means for converting, search means for searching for a route from the image after the polar coordinate conversion, and inverse conversion means for inversely converting the route, wherein the ellipse computing means corresponds to the route after inverse conversion. The ellipse may be calculated as an ellipse corresponding to the annular portion.

  Further, the present invention provides an ellipse computing means for computing an ellipse corresponding to the fetal annular region based on a tomographic image obtained by transmitting and receiving ultrasound to a fetus in the mother body, While moving the center of the ellipse calculated by the ellipse calculation means in one axial direction of the ellipse and rotating the reference area extending in the other axial direction of the ellipse at the center after the movement, A program that calculates an evaluation value in the reference area at each position and each rotation angle, and functions as an angle correction unit that corrects the angle of the ellipse with respect to the annular portion using the evaluation value. It is.

  According to the present invention, the inclination of the ellipse corresponding to the fetal annular part can be brought close to the inclination of the annular part.

1 is a block diagram illustrating an ultrasonic diagnostic apparatus according to a first embodiment. It is a block diagram which shows a position calculating part. It is a block diagram which shows an area | region calculating part. It is a figure which shows a B mode tomographic image. It is a figure which shows a template. It is a figure which shows an edge emphasis image. It is a figure which shows the expansion | deployment image produced | generated by polar coordinate conversion. It is a figure which shows the expansion | deployment image to which the mask was applied. It is a figure which shows the expansion | deployment image to which the mask was applied. It is a figure which shows the image for demonstrating a route search process. It is a figure which shows a search map. It is a figure which shows the image for demonstrating a route search process. It is a figure which shows the route line produced | generated by reverse conversion. It is a figure which shows an ellipse. It is a figure which shows a part of B mode tomographic image, and the axis | shaft of an ellipse. It is a figure which shows a part of B mode tomographic image, and the axis | shaft of an ellipse. It is a figure which shows the expansion | deployment image which concerns on the modification 1. FIG. It is a figure which shows the B mode tomographic image which concerns on the modification 2. It is a block diagram which shows the ultrasound diagnosing device which concerns on 2nd Embodiment. It is a figure which shows an example of the ellipse which approximates the midline of a fetal head and a fetal head. It is a figure which shows an example of the ellipse which approximates the midline of a fetal head and a fetal head. It is a figure which shows an example of the ellipse which approximates the midline of a fetal head and a fetal head. It is a figure which shows distribution of dispersion | distribution of a luminance value sum total. It is a figure which shows distribution of a luminance value sum total. It is a figure which shows the template which concerns on 3rd Embodiment. It is a figure which shows the edge emphasis image which concerns on 3rd Embodiment. It is a figure which shows the expansion | deployment image by which the mask which concerns on 3rd Embodiment was provided.

<First Embodiment>
Hereinafter, an ultrasonic diagnostic apparatus according to a first embodiment of the present invention will be described with reference to FIG. FIG. 1 shows an ultrasonic diagnostic apparatus according to the first embodiment. 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 an image of a tissue in a living body by transmitting and receiving ultrasonic waves. 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. The probe 10 has, for example, a 1D array transducer. The 1D array vibrator is formed by arranging a plurality of vibration elements in a one-dimensional manner. An ultrasonic beam is formed by the 1D array transducer and is repeatedly electronically scanned. Thereby, a scanning surface is formed in the living body for each electronic scanning. The scanning plane corresponds to a two-dimensional echo data capturing space. The probe 10 may have a 2D array transducer in which a plurality of transducer elements are two-dimensionally arranged instead of the 1D array transducer. When an ultrasonic beam is formed by the 2D array transducer and repeatedly scanned electronically, a scanning surface as a two-dimensional echo data capturing space is formed for each electronic scan, and the ultrasonic beam is scanned two-dimensionally. As a result, a three-dimensional space is formed as a three-dimensional echo data capturing space. As the scanning method, sector scanning, linear scanning, convex scanning, or the like is used. 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 12 functions as a transmission beamformer and a reception beamformer. At the time of transmission, the transmission / reception unit 12 supplies a plurality of transmission signals having a certain delay relationship to the plurality of vibration elements included in the probe 10. Thereby, an ultrasonic transmission beam is formed. At the time of reception, a 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 12. The transmission / reception unit 12 forms a reception beam by applying a phasing addition process to a plurality of reception signals. The beam data is output to the signal processing unit 14. That is, the transmission / reception unit 12 performs a delay process on the reception signal obtained from each vibration element in accordance with the delay processing condition for each vibration element, and adds a plurality of reception signals obtained from the plurality of vibration elements to receive the signal. Form a beam. The delay processing condition is defined by reception delay data (delay time). A reception delay data set (a set of delay times) corresponding to a plurality of vibration elements is supplied from the control unit 30. It should be noted that techniques such as transmission aperture synthesis may be used in ultrasonic transmission / reception. Further, the transmission / reception unit 12 may execute parallel reception processing.

  The ultrasonic beam (transmission beam and reception beam) is electronically scanned by the action of the transmission / reception unit 12, thereby forming a scanning plane. The scanning plane corresponds to a plurality of beam data, which constitute reception frame data (RF signal frame data). Each beam data is composed of a plurality of echo data arranged in the depth direction. By repeating the electronic scanning of the ultrasonic beam, a plurality of reception frame data arranged on the time axis are output from the transmission / reception unit 12. They constitute a received frame sequence.

  When the ultrasonic beam is electronically scanned two-dimensionally by the action of the transmitting / receiving unit 12, a three-dimensional echo data capturing space is formed, and volume data as an echo data aggregate is formed from the three-dimensional echo data capturing space. Is acquired. By repeating the electronic scanning of the ultrasonic beam, a plurality of volume data arranged on the time axis are output from the transmission / reception unit 12. They constitute a volume data string.

  The signal processing unit 14 is a module that applies signal processing such as detection and logarithmic compression to the beam data output from the transmission / reception unit 12. The beam data after signal processing may be stored in a memory. Of course, beam data to which such signal processing is not applied may be stored in the memory, and the above processing may be applied when reading the beam data.

  The DSC (digital scan converter) 16 is a module having a conversion function (such as a coordinate conversion function and an interpolation processing function), and generates a tissue display frame sequence based on the received frame sequence output from the signal processing unit 14. Each tissue display frame sequence is data of a B-mode tomographic image. The tissue display frame sequence is displayed on the display unit 20 such as a monitor via the display processing unit 18. As a result, the B-mode tomographic image is displayed as a moving image in real time.

  The display processing unit 18 overlays necessary graphic data on the tomographic image and the like, thereby generating a display image. This image data is output to the display unit 20, and one or a plurality of images are displayed side by side in a display mode according to the display mode.

  The display unit 20 is configured by a display device such as a liquid crystal display. The display unit 20 may be configured by a plurality of display devices.

  The image processing unit 22 has a function of detecting the position and region of the fetal annular region based on the B-mode tomographic image (tissue display frame) and measuring the annular region based on the detection result. The annular portion is an annular tissue having an edge portion, for example, the head or abdomen. Note that the image processing unit 22 may perform processing on a B-mode tomographic image obtained from a two-dimensional scanning plane, or may process data corresponding to an arbitrary cross section in volume data obtained from a three-dimensional space. Processing may be performed on a two-dimensional B-mode tomographic image generated based on the target.

  The image processing unit 22 includes, for example, a position calculation unit 24, a region calculation unit 26, and a measurement unit 28.

  The position calculation unit 24 calculates candidates for the position of the fetal annular region (for example, fetal head or abdomen) by applying template matching to the B-mode tomographic image. For example, the position calculation unit 24 calculates a candidate for the center position of the fetal annular region. The candidate for the center position is used in the area calculation unit 26 described later.

  The region calculation unit 26 performs coordinate conversion of the B-mode tomographic image using the center position candidate calculated by the position calculation unit 24, and calculates the region of the fetal annular region based on the image after coordinate conversion. For example, the region calculation unit 26 calculates an ellipse that approximates the annular portion.

  The measuring unit 28 measures the annular part of the fetus using the region of the annular part (for example, an ellipse that approximates the annular part) calculated by the region calculator 26 and the B-mode tomographic image. For example, when the B-mode tomographic image representing the head of the fetus is generated by photographing the fetal head as an annular part, the measurement unit 28 measures the fetal head based on the B-mode tomographic image. When the fetal abdomen is imaged as an annular part and a B-mode tomographic image representing the fetal abdomen is generated, the measuring unit 28 measures the fetal abdomen based on the B-mode tomographic image.

  The control unit 30 has a function of performing operation control of each component shown in FIG. The control unit 30 may include a region of interest setting unit that sets a region of interest (ROI) on the B-mode tomographic image.

  An input unit 32 is connected to the control unit 30. As an example, the input unit 32 includes an operation panel including input devices such as a trackball, a keyboard, various buttons, and various knobs. The user can use the input unit 32 to specify or input the position of the scanning plane, the position of the cross section, information on the region of interest, and the like.

  In the above-described ultrasonic diagnostic apparatus, the configuration other than the probe 10 can be realized by using hardware resources such as a processor and an electronic circuit, and a device such as a memory can be used as necessary in the realization. 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. Of course, a GPU (Graphics Processing Unit) or the like may be used.

  Hereinafter, the ultrasonic diagnostic apparatus according to the first embodiment will be described in detail. In the first embodiment, the annular part of the fetus is the fetal head, and a B-mode tomographic image representing the fetal head is generated by transmitting and receiving ultrasonic waves to the fetal head.

  Hereinafter, the position calculator 24 will be described in detail with reference to FIG. FIG. 2 shows the configuration of the position calculation unit 24. The position calculation unit 24 includes a template generation unit 34, an edge enhanced image generation unit 36, and a matching processing unit 38.

The template generation unit 34 generates a template used for matching processing by the matching processing unit 38. The template has, for example, an annular shape (ring shape). For example, the template generation unit 34 generates an annular template having a size (that is, a diameter and a width) based on the fetal pregnancy days GD. Specifically, the template generation unit 34 calculates an average value (statistical value) of a child's large lateral diameter (BPD: Biparetal Diameter) and its variation (for example, standard deviation SD) based on the number of days of pregnancy GD. An annular template having a BPD average value (statistical value) as a diameter and a standard deviation SD as a width is generated. The BPD average value (statistical value) and the standard deviation SD are statistically calculated by the following formula.
BPD average value (statistical value) = − 18.3 + 3.85 × 10 ⁻¹ GD + 1.06 × 10 ⁻³ GD ² −3.70 × 10 ⁻⁶ GD ³
SD = 1.73 + 7.96 × 10 ⁻³ GD

  The above formula for obtaining the BPD average value and the standard deviation SD is a formula obtained statistically, and can vary depending on, for example, the nationality of the mother.

  Information indicating the number of days of pregnancy may be input to the ultrasonic diagnostic apparatus by a user such as a doctor, or may be input to the ultrasonic diagnostic apparatus from another apparatus.

  The edge-enhanced image generation unit 36 applies an filter to the B-mode tomographic image, thereby generating an edge-enhanced image in which the edge part of the fetal head as an annular part (the part corresponding to the skull) is emphasized. For example, a DoG filter (Difference Of Gaussian filter) is used as the filter. By using the DoG filter, a difference between images having different “blurs” is calculated, thereby extracting a portion having high or low luminance with respect to the surroundings. For example, it is possible to extract a portion with high brightness with respect to the surroundings such as a skull.

  Note that, when the DoG filter is applied, a positive value and a negative value are obtained. However, in order to use only a positive value, the negative value is set to zero (0). In addition, when the DoG filter is applied, the edge portion (the edge portion of the image) of the B-mode tomographic image itself is also emphasized, so that portion may be masked (for example, processing for setting the luminance value to 0). . As another example, the absolute value of the value obtained by applying the Dog filter may be adopted, or the obtained value may be multiplied by −1 to make the negative value zero (ie, The original negative value may be used).

  The edge-enhanced image generating unit 36 generates an edge-enhanced image by applying a DoG filter after reducing the size of the B-mode tomographic image to be processed in order to improve the processing efficiency of the matching processing unit 38 described later. May be.

  Specifically, the edge-enhanced image generation unit 36 applies a 5 tap binomial filter having a coefficient of 1: 4: 6: 4: 1 to the B-mode tomographic image two-dimensionally, and an image after the application. On the other hand, the first image down-sampled on the second floor and the second image down-sampled by one step are generated, and the difference between the first image and the second image is calculated. Thereby, an image to which the DoG filter is applied, that is, an edge enhanced image is generated. In addition, said example is an example and the kind of filter, a coefficient, and the procedure and frequency | count of various sampling processes are not limited to said example.

  The matching processing unit 38 applies a matching process using the template generated by the template generating unit 34 to the edge enhanced image generated by the edge enhanced image generating unit 36, so that the fetal head position candidate is applied. Specifically, a candidate for the center position of the fetal head (skull) is calculated. The matching processing unit 38 calculates the similarity between the template and the fetal head image (skull image) at each position while changing the position of the template on the edge enhanced image. The matching processing unit 38 specifies the position of the template that maximizes the similarity, and specifies the center position of the template at that position as a candidate for the center position of the fetal head (skull). Note that the matching processing unit 38 may calculate the center position of the fetal head by matching the template with the original B-mode tomographic image without using the edge-enhanced image. In this case, the edge enhanced image generation unit 36 may not be included in the position calculation unit 24.

  In this manner, the candidate for the center position of the fetal head (skull) is calculated. The candidate of the center position calculated in this way is used for coordinate conversion performed by the area calculation unit 26.

  Hereinafter, the region calculation unit 26 will be described in detail with reference to FIG. FIG. 3 shows the configuration of the area calculation unit 26. The area calculation unit 26 includes a conversion unit 40, a mask setting unit 42, a search unit 44, an inverse conversion unit 46, an ellipse calculation unit 48, a validity determination unit 50, a center mask unit 52, an end determination unit 54, and a center update unit 56. including.

  The conversion unit 40 performs polar coordinate conversion of the B-mode tomographic image using the candidate for the center position of the fetal head (skull) calculated by the position calculation unit 24 as the origin of the polar coordinate system. Polar coordinate conversion is a process of converting coordinates with the center position candidate as the origin, for example, with the vertical axis as the distance (radial radius r) from the center position candidate and the horizontal axis as the angle (deflection angle θ). That is, the polar coordinate system after the coordinate conversion is a polar coordinate system having a candidate for the center position of the fetal head as the origin. By this conversion processing, the B-mode tomographic image is developed with the center position candidate as the origin, and thereby a developed image (an image expressed in a polar coordinate system) is generated. In the first embodiment, by using the center position candidate (center position candidate obtained by template matching) calculated by the position calculation unit 24, polar coordinate conversion is performed using a position closer to the center position of the fetal head. This makes it possible to improve the accuracy of route search processing and ellipse calculation described later. Hereinafter, the image after the polar coordinate conversion is referred to as a “development image”.

  The mask setting unit 42 sets a mask extending in the radial direction r in the specific region in the polar angle θ direction of the polar coordinate system with respect to the developed image (image after polar coordinate conversion). The area where the mask is set is an area where route search processing using dynamic programming described later is not performed. As another example, the luminance value in the area where the mask is set is set to 0 (zero), and the route is not searched from the information of the area where the mask is set by the search process using dynamic programming. May be. The specific region is, for example, a region including a portion where an ultrasonic beam is transmitted / received in parallel to the fetal head (skull). The luminance value of the portion where the ultrasonic beam is transmitted / received in parallel is likely to be lower than the luminance value of the other portion, and noise is likely to occur. Due to the influence of the noise, an erroneous search may occur in a route search process described later. By setting a mask in a specific area, it is possible to remove or eliminate the influence of noise and suppress or prevent the occurrence of erroneous search. For example, normally, noise is likely to occur at portions corresponding to the front and rear surfaces of the fetal head, and therefore a mask is set at those portions.

  The search unit 44 searches for a route from the developed image by applying route search processing to the developed image (image after polar coordinate conversion). This path corresponds to a candidate for an edge portion of the fetal head (a portion corresponding to the skull). For example, the search unit 44 searches for a route by applying dynamic programming to the developed image. At this time, in the developed image, the search unit 44 excludes the region where the mask is set from the search region, and searches for the route by applying the dynamic programming to the region where the mask is not set. In addition, the search unit 44 estimates a route (for example, a straight route) in an area where a mask is set in the developed image. For example, when there is a region where a mask is set between a plurality of regions where a mask is not set, the search unit 44 may replace a plurality of routes in a plurality of regions where a mask is not set with each other. The whole route is calculated from the developed image by connecting with a straight route estimated in the set region.

  The inverse transform unit 46 inversely transforms the route searched by the search unit 44 into the original image space (image space before coordinate conversion).

  The ellipse computing unit 48 computes an ellipse that approximates the fetal head using the post-inversion path obtained by the inverse transformation unit 46 and the B-mode tomographic image. Thereby, the center position of the ellipse, the length of the minor axis, the length of the major axis and the inclination are obtained as parameters of the approximate ellipse. Hereinafter, an ellipse calculated by the ellipse calculation unit 48 (an ellipse approximated to the fetal head) is referred to as an “approximate ellipse”.

  The validity determination unit 50 determines the validity of the approximate ellipse obtained by the ellipse calculation unit 48. For the determination of validity, the ratio between the length of the major axis of the ellipse and the length of the minor axis is used. The validity determination unit 50 determines that there is validity when the ratio between the length of the long axis and the length of the short axis of the approximate ellipse obtained by the ellipse calculation unit 48 is within a predetermined allowable range. If the ratio is not within the allowable range, it is determined that the ratio is not valid. In addition, for the validity determination of the ellipse, at least one of the size, position, angle, etc. of the ellipse may be used, or a plurality of combinations among them may be used.

  The center mask unit 52 sets a mask within a predetermined range from the origin for the developed image when the validity determination unit 50 determines that there is no validity. The search unit 44 searches again for the developed image in which the mask is set. In the B-mode tomographic image showing the fetal head, there is a high luminance region called a midline, and the high luminance region may be erroneously searched as a route in the route search process. That is, in the route search process, the purpose is to search for a route corresponding to the fetal skull, but there is a possibility that a route passing through the midline may be erroneously searched. Since the median line is assumed to be represented in the central portion of the B-mode tomographic image, by setting a mask in the range corresponding to the central portion and excluding the range from the route search process, the median line It becomes possible to suppress or prevent erroneous search due to the influence. Note that the mask setting unit 42 may set the mask within a predetermined range from the origin in the developed image without providing the center mask unit 52. In this case, the mask setting unit 42 may set a mask regardless of the determination result by the validity determination unit 50.

  The end determination unit 54 determines the end of processing of each unit of the region calculation unit 26. For example, when a series of processing from the conversion unit 40 to the center update unit 56 is performed a predetermined number of times (for example, 3 times), the end determination unit 54 determines that the processing is ended. When the predetermined number of processes has not been executed, a series of processes from the conversion unit 40 to the center update unit 56 are executed. Further, when the variation of the parameters such as the position, size, and angle of the approximate ellipse obtained by the ellipse calculation unit 48 becomes small (for example, when included in a predetermined allowable range), the end determination unit 54 , It may be determined that the processing is completed.

  The above processing is summarized as follows. When the processing from the conversion unit 40 to the validity determination unit 50 is executed and the validity determination unit 50 determines that there is validity, the determination by the end determination unit 54 is performed. When the validity determination unit 50 determines that there is no validity, the processing from the search unit 44 to the ellipse calculation unit 48 is executed through the processing by the center mask unit 52, and the processing of the validity determination unit 50 is skipped. Thus, determination by the end determination unit 54 is performed. When the series of processes has not been executed a predetermined number of times, the process by the center update unit 56 is performed. The series of processes described above is set as one set, and the set is executed a predetermined number of times.

  If the end determination is not made by the end determination unit 54, the center update unit 56 updates the center position of the approximate ellipse obtained by the ellipse calculation unit 48 as a new origin of the polar coordinate system in the polar coordinate conversion by the conversion unit 40. . Thereafter, a series of processing from the conversion unit 40 to the end determination unit 54 is executed. At this time, the conversion unit 40 performs polar coordinate conversion of the B-mode tomographic image using the center position (updated center position) of the approximate ellipse obtained by the ellipse calculation unit 48 as a new origin of the polar coordinate system. Thereby, polar coordinate conversion can be performed using a position closer to the center position of the fetal head, and the accuracy of the route search process and the ellipse calculation is improved.

  When the end determination is made by the end determination unit 54, the measurement unit 28 measures the fetal head using the approximate ellipse and the B-mode tomographic image obtained at the time of the end determination.

  Hereinafter, the B-mode tomographic image generated by the ultrasonic diagnostic apparatus according to the first embodiment will be described in detail with reference to FIG. FIG. 4 shows the B-mode tomographic image. In the B-mode tomographic image 58, a fetal head image 60 is represented. For example, the position and angle of the probe 10 are adjusted by the user so that the fetal head is represented in the B-mode tomographic image. In FIG. 4, the hatched area represents a high luminance (echo) area. In the B-mode tomographic image 58 suitable for measurement, the upper part 62, the lower part 64, and the midline 66 of the image in which the ultrasonic beam is easily reflected are easily drawn, and the brightness of these parts is likely to be high. In other tissues 68, the luminance may increase. On the other hand, the side surfaces 70 and 72 (surfaces corresponding to the front and rear surfaces of the fetal head in normal ultrasound diagnosis) (for example, the portion where the ultrasound beam is transmitted and received in parallel) The region including the image is difficult to draw, and the luminance of those portions tends to be low. From such a B-mode tomographic image 58, the measuring unit 28, for example, has a large lateral diameter (BPD) equivalent to the length in the direction perpendicular to the median line 66, and a longitudinal dimension (OFD). Frontal Diameter (HC), Head Circumference (HC), etc. can be measured. Of course, parameters other than these may be measured.

  Hereinafter, the template generated by the template generation unit 34 will be described in detail with reference to FIG. FIG. 5 shows an example of a template.

  The template 74 is a template (doughnut-type template) having an annular shape generated by the template generation unit 34. The shape of the template 74 is a shape assuming the fetal head (skull), and the inside 76 of the ring is a region assuming the inside of the skull. Since it can be assumed that the skull has an elliptical shape close to a circle, a template 74 having a circular overall shape is used in template matching. At the stage of template matching, the orientation (rotation angle) of the fetal head is not known. Therefore, the detection accuracy of the center position of the fetal head can be improved by using the circular template 74 having no directionality rather than using the template having the directionality such as an ellipse.

  For example, the length from the center position O of the template 74 to the center position of the annular portion of the template 74 corresponds to the radius of the template 74, and the length twice the radius corresponds to the diameter A of the template 74. The diameter A is a BPD average value (statistical value) calculated using the number of days of pregnancy. The width B of the template 74, that is, the width B of the annular portion is the standard deviation SD of the BPD average value calculated using the number of gestational days. Thus, the template 74 is a template that reflects the size estimated from the number of days of pregnancy. Since the template 74 having such characteristics reflects the actual size and shape of the skull, by performing template matching using the template 74, a position closer to the center position of the skull is a candidate for the center position. Are easily detected.

  Hereinafter, the matching process performed by the matching processing unit 38 will be described in detail with reference to FIG. FIG. 6 shows an example of the edge enhanced image.

  The edge enhanced image 78 is an image generated by the edge enhanced image generation unit 36. That is, the edge enhanced image 78 is an image generated by applying a DoG filter to the B-mode tomographic image 58 shown in FIG. In the edge-enhanced image 78, a portion corresponding to a fetal skull is emphasized.

The matching processing unit 38 applies a matching process using the template 74 to the edge-enhanced image 78 to calculate a candidate for the center position of the fetal head (skull). Specifically, the matching processing unit 38 resembles the template 74 and the fetal head image (skull image) at each position while changing the position of the template 74 on the edge enhanced image 78 as indicated by an arrow 80. Calculate the degree. As a result, a two-dimensional map of similarity (similarity map) is generated. The similarity is calculated by the following equation (1).

  Here, I (i, j) is the luminance (0 to 255) of the edge enhanced image 78 (DoG image). T (x, y) is the brightness (0 or 1) of the template 74.

  The matching processing unit 38 may apply a filter process to the similarity map. For example, two types of filters are used. The first filter is, for example, a (5 × 5), σ = 2.0 Gaussian filter, and the second filter is, for example, a Gaussian filter having a size that is half the horizontal × vertical size of the image. The first filter is a filter for removing the local high luminance value and smoothing the peripheral region so that the candidate is detected from the region having the high luminance value in the peripheral region. Assuming that the observation target is arranged at a position close to the center of the B-mode tomographic image as a characteristic of the B-mode tomographic image, the second filter makes it difficult for a candidate to be detected from the end of the B-mode tomographic image. , A filter that slightly weights the central portion of the B-mode tomographic image. Of course, these filters may not be applied.

  The matching processing unit 38 detects a position having the maximum similarity as a candidate for the center position of the fetal head (skull) in the similarity map. This center position candidate is used in polar coordinate conversion described later.

  Hereinafter, with reference to FIGS. 7 to 15, the processing by the region calculation unit 26 will be described in detail.

  FIG. 7 shows an example of a developed image generated by polar coordinate conversion. The developed image 82 is an image generated by polar coordinate conversion by the conversion unit 40. Specifically, the developed image 82 is obtained by performing polar coordinate conversion on the B-mode tomographic image 58 shown in FIG. 4 using the candidate for the center position of the fetal head (skull) obtained by the matching processing unit 38. It is the image produced | generated by. The horizontal axis indicates the angle θ (deflection angle θ) of the polar coordinate system, and the vertical axis indicates the distance r (radial radius r) of the polar coordinate system. The origin (θ = 0, r = 0) of the developed image 82 is a position corresponding to the candidate for the center position of the fetal head (skull) obtained by the matching processing unit 38. An image 84 in the developed image 82 is an image corresponding to the upper part 62, the lower part 64, and the median line 66 represented in the B-mode tomographic image 58, and the image 86 is formed on the tissue 68 represented in the B-mode tomographic image 58. Corresponding image.

  FIG. 8 shows an example of a developed image to which a mask is applied. A mask 88 is set in the developed image 82 by the mask setting unit 42. For example, in the developed image 82, a mask 88 having a width of ± 45 ° in the angle θ direction is set at a position where the angle θ is 90 ° and a position where the angle θ is −90 °. Of course, the width of the angle is merely an example, and other values may be used, or the width may be changed by the user. The mask 88 has a shape extending in the direction of the distance r from the position where the distance r = 0. The position where the angle θ is ± 90 ° is usually a position corresponding to the front surface and the rear surface of the fetal head, and since the ultrasonic beam is easily transmitted and received in parallel in that portion, it is difficult to depict that portion. If that portion is included in the route search target area, there is a possibility that an erroneous search will occur in the route search processing. In addition, if another tissue with high luminance is drawn near the portion where the skull is difficult to be drawn, there is a possibility that a false search may occur in the route search process. In order to cope with this, a mask 88 having a width in the angle θ direction is set in a portion where the skull is difficult to be drawn (position where the angle θ is ± 90 °). As a result, the portion is excluded from the route search area, so that occurrence of erroneous search can be suppressed or prevented. For example, when the image 86 shown in FIG. 7 exists in the vicinity of the angle θ of ± 90 °, a portion of the image 86 may be extracted as a route, and a portion other than the skull may be extracted as a route. . By setting a mask, such an erroneous search can be prevented.

  Hereinafter, the route search process by the search unit 44 will be described in detail. FIG. 9 shows a developed image 82. The search unit 44 searches the developed image 82 for a route that maximizes the sum of luminance values on the route. Any method may be used to search for a route that maximizes the sum of luminance values, but the search unit 44 searches for a route using dynamic programming as an example. At this time, the search unit 44 searches the route by excluding the region where the mask 88 is set from the route search region. A route 90 indicated by a solid line in FIG. 9 is a route searched by applying dynamic programming. The path 90 is along the image 84. In addition, the search unit 44 estimates the route 92 in the area where the mask 88 is set. For example, the search unit 44 calculates the entire route by connecting end portions of the plurality of routes 90 searched by the dynamic programming method with a linear route 92 (route indicated by a broken line). The path 92 is a linear path in the area where the mask 88 is set. Note that the search unit 44 searches for a route in which the r coordinate is close at the left and right ends of the developed image 82 so that the searched route becomes a closed curve in the original image space (image space before coordinate conversion). May be.

  Hereinafter, dynamic programming will be described with reference to FIGS. 10 to 12. 10 and 12 show images for explaining the route search processing. FIG. 11 shows an example of a search map.

  First, as illustrated in FIG. 10, the search unit 44 adds the luminance value of the pixel having the maximum luminance value to the target pixel 94 among the three pixels 96 arranged on the left side of the target pixel 94. . The search unit 44 performs this process from the upper end to the lower end of the image, and then performs the process from the left end to the right end of the image, thereby generating a two-dimensional search map. In the search map, the higher luminance corresponds to the region where the sum of the luminance values is larger.

  FIG. 11 shows an example of the search map obtained as described above. The search unit 44 detects the pixel 98 having the maximum luminance from the right end of the search map.

  Next, as illustrated in FIG. 12, the search unit 44 selects a pixel having the maximum luminance value among the three pixels 102 arranged on the left side of the current pixel 100 (the pixel 98 described above), This process is performed to the left end of the image. After the route reaches the left end of the image, the search unit 44 compares the r coordinate of the search start point and the r coordinate of the search end point, and if the difference is large (for example, if the difference is greater than or equal to the threshold), the search map A pixel having the second highest luminance value from the right end of is selected as a search start point, and the search is performed again. The search unit 44 repeats this process until the difference between the search start point and the search end point is included within a predetermined allowable range.

  The search unit 44 searches for a route having the maximum sum of luminance values by the above processing. The search unit 44 uses at least one parameter selected from parameter groups such as luminance gradient information, edge amount, entropy, likelihood, Hog, SaliencyMap, L1, and L2 norms as parameters other than the luminance value. The route may be searched using a combination of a plurality of selected parameters. Also, depending on the parameters used, minimization or optimization may be performed instead of maximization. Further, the direction in which the search map is generated may be the opposite direction to the above-described example.

  When the route is searched by the search unit 44 as described above, the inverse transform unit 46 inversely transforms the route into the original image space (image space before coordinate transformation). FIG. 13 shows a route line 104 generated by inverse transformation. FIG. 13 also shows the B-mode tomographic image 58 before coordinate conversion shown in FIG. 4, and the route line 104 is shown superimposed on the B-mode tomographic image 58.

  When the route line 104 is generated by the inverse transformation, the ellipse computing unit 48 computes an ellipse (approximate ellipse) that approximates the fetal head using the route line 104 and the B-mode tomographic image 58. Hereinafter, this process will be described in detail. First, the ellipse calculation unit 48 extracts a certain percentage of pixels with higher luminance (for example, 50% of pixels) among all the pixels on the route line 104. This extraction process is for increasing the accuracy of the ellipse approximation by extracting and using pixels existing on the skull among the pixels existing on the path. Next, the ellipse calculation unit 48 randomly selects five or more candidate points from the extracted pixel group, and applies an ellipse to these candidate point groups. As an ellipse fitting method, for example, a Hough transform, a least square method, an optimization method using a polynomial, or the like can be used. FIG. 14 shows an ellipse 106 applied to the candidate point group. Next, the ellipse calculation unit 48 calculates the distance between the ellipse calculated by the fitting and the extracted pixel group (the pixel group extracted from all the pixels on the route line 104). The number of pixels within a predetermined threshold (for example, the number of pixels having a distance within 5 pixels) is calculated as the number of pixels on the ellipse. The ellipse calculation unit 48 performs these processes a predetermined number of times (for example, 100 times), and adopts the ellipse having the maximum number of pixels on the ellipse as an ellipse that approximates the fetal head (approximate ellipse). Thereby, the center position of the approximate ellipse, the length of the short axis, the length of the long axis, and the angle are specified.

  The validity determination unit 50 determines the validity of the approximate ellipse adopted by the ellipse calculation unit 48 as described above. When the ratio between the length of the major axis and the length of the minor axis of the approximate ellipse is included within a predetermined allowable range, the validity determination unit 50 determines that there is validity, and the ratio is within the allowable range. If it is not included, it is determined that there is no validity. When it is determined that there is no validity, the search unit 44 searches again.

  When it is not determined by the end determination unit 54 that the process has been completed, that is, when the predetermined number of times of processing has not been executed, the center update unit 56 calculates the approximate ellipse employed by the ellipse calculation unit 48. The center position is updated as a new origin of the polar coordinate system in the polar coordinate conversion by the conversion unit 40. When the predetermined number of processes are executed, for example, the approximate ellipse obtained in the last process may be adopted as an ellipse for measurement by the measurement unit 28, or the position, size, and angle of the ellipse. An ellipse having the smallest parameter variation such as the above may be adopted as a measurement ellipse.

  Hereinafter, the process performed by the measurement unit 28 will be described in detail with reference to FIGS. 15 and 16. 15 and 16 show a part of the B-mode tomographic image 58 and the axis of the ellipse. As shown in FIG. 15, the short axis 108 and the long axis 110 of an ellipse (approximate ellipse) approximated to the fetal head (measurement ellipse) are obtained by the above processing. The measuring unit 28 detects intersections 112, 114, 116, and 118 between the short axis 108 of the approximate ellipse and the outside and inside of the fetal skull image (upper 62 and lower 64). This is because fetal head measurement methods include measuring the length between the upper outer side and the lower inner side and measuring the length between the upper outer side and the lower outer side. For example, the intersection is detected by using a gradient. As the calculation of the gradient, for example, a difference value, a Sobel filter, a Laplacian filter, a Breutit filter, or the like can be used.

  Hereinafter, a case where a Sobel filter is used will be described as an example with reference to FIG. The measurement unit 28 sets a search range on a straight line 120 obtained by extending the short axis 108 of the approximate ellipse (measurement ellipse). The search range includes an upper outer search range 122, an upper inner search range 124, a lower inner search range 126, and a lower outer search range 128. Next, the measurement unit 28 calculates the gradient strength at each point indicated by a black circle in FIG. 16, and detects the point at which the gradient strength is maximum within each search range as the intersection point. Thus, the intersection 112 is detected in the search range 122, the intersection 114 is detected in the search range 124, the intersection 116 is detected in the search range 126, and the intersection 118 is detected in the search range 128. The

  Moreover, the measurement part 28 may calculate the outer periphery ellipse 130 of the fetal head (skull) shown by FIG. 15 by performing said process about multiple directions. In this case, the measurement unit 28 may calculate the circumference of the outer periphery ellipse 130 as the head circumference HC and calculate the length of the long axis 110 as the head front-rear diameter OFD.

  When the fetal head is measured by the measuring unit 28, for example, the measurement result is displayed on the display unit 20.

  As described above, in the first embodiment, the candidate for the center position of the fetal head can be obtained by performing the matching process using the template that approximates the shape of the fetal head (skull). By performing polar coordinate conversion using the center position candidate, polar coordinate conversion that reflects the more accurate center position becomes possible, so the accuracy of the route search processing using the developed image generated by polar coordinate conversion, that is, The accuracy of elliptical approximation is improved. Since an ellipse approximated by the fetal head (skull) is obtained, the measurement accuracy of the fetal head using the ellipse is improved. As described above, in the first embodiment, in order to improve the accuracy of the route search processing for the image after the polar coordinate conversion, as a pre-processing of the polar coordinate conversion based on the route search processing, the center position of the fetal head is obtained by template matching. The candidate is computed.

(Modification 1)
Hereinafter, Modification 1 will be described with reference to FIG. FIG. 17 shows an example of a developed image. The developed image 82 shown in FIG. 17 is the same image as the developed image 82 shown in FIG.

  In the first modification, the mask setting unit 42 applies weighting processing to the developed image 82 instead of setting a mask on the developed image 82. For example, in the developed image 82, weighted regions 132 having a width of ± 45 ° in the angle θ direction are set at a position where the angle θ is 90 ° and a position where the angle θ is −90 °. The weighting region 132 has a shape extending in the distance r direction from the position where the distance r = 0. In the weighting area 132, the weighting coefficient at the position of ± 90 ° is set to be the smallest, and the larger the weighting coefficient is assigned to the position away from the position of ± 90 ° in the angle θ direction. The mask setting unit 42 multiplies the luminance value of each pixel of the developed image 82 by a weight coefficient, and the search unit 44 performs a route search process using the luminance value multiplied by the weight coefficient. As a result, the luminance value of the portion where the skull is difficult to be detected becomes difficult to be used in the route search process, so that the influence of noise in that portion can be suppressed or prevented. Note that the mask setting unit 42 may perform weighting processing on the entire developed image 82. Also in this case, the weighting coefficient at the position of ± 90 ° is set to be the smallest, and a larger weighting coefficient is assigned to a position away from the position of ± 90 ° in the angle θ direction.

(Modification 2)
Hereinafter, Modification 2 will be described with reference to FIG. FIG. 18 shows an example of a B-mode tomographic image. The B-mode tomographic image 58 shown in FIG. 18 is the same image as the B-mode tomographic image 58 shown in FIG.

  In the second modification, the mask setting unit 42 sets a mask corresponding to the type of the probe 10 in the developed image. Since the transmission / reception direction of the ultrasonic beam varies depending on the type of the probe 10, the mask setting unit 42 sets a mask corresponding to the transmission / reception direction in the developed image. The mask setting unit 42 includes, for example, a direction that is orthogonal to the transmission / reception direction of the ultrasonic beam indicated by the arrow 134 and has a predetermined width with the ellipse center position candidate calculated by the position calculation unit 24 as a reference position. A mask is set in the area 136 having the mask. Actually, a mask is set in a region corresponding to the region 136 in the developed image generated by polar coordinate conversion. In such a region 136, since the ultrasonic beam is easily transmitted and received in parallel with the fetal skull, it is difficult to depict the skull. By setting a mask in this area 136, it is possible to suppress or prevent an erroneous search for a route due to noise in the area 136. Since the position and size of the region 136 change depending on the type of the probe 10, the mask setting unit 42 changes the position and size of the region 136 according to the type of the probe 10.

  For example, mask information indicating the position and size of the region 136 corresponding to the type of each probe 10 is recorded in advance in the ultrasonic diagnostic apparatus. When the probe 10 is connected to the ultrasonic diagnostic apparatus main body, the mask setting unit 42 acquires information indicating the type of probe from the probe 10, acquires mask information associated with the type, and develops an image according to the mask information. Set the mask to.

Second Embodiment
Hereinafter, with reference to FIG. 19, an ultrasonic diagnostic apparatus according to the second embodiment of the present invention will be described. FIG. 19 shows an ultrasonic diagnostic apparatus according to the second embodiment. FIG. 19 is a block diagram showing the overall configuration.

  The ultrasonic diagnostic apparatus according to the second embodiment includes an image processing unit 22A instead of the image processing unit 22 of the ultrasonic diagnostic apparatus according to the first embodiment. In addition to the configuration included in the image processing unit 22, the image processing unit 22 </ b> A further includes an angle correction unit 138. Since the configuration other than the angle correction unit 138 is the same as the configuration of the ultrasonic diagnostic apparatus according to the first embodiment, the angle correction unit 138 will be described below, and the description of the other configuration will be omitted.

  The angle correction unit 138 corrects the angle of the ellipse (approximate ellipse) that approximates the fetal head (skull) calculated by the region calculation unit 26. As shown in FIG. 4, in the B-mode tomographic image showing the fetal head, a high-luminance region called a median line 66 extending in the front-rear direction of the fetal head is depicted. In order to increase the measurement accuracy of the fetal head, it is desirable that the major axis of the ellipse (approximate ellipse) approximating the fetal head and the midline are in the same direction. Therefore, the angle correction unit 138 corrects the angle of the approximate ellipse so that the major axis direction of the ellipse that approximates the fetal head (approximate ellipse) matches the direction of the midline. Hereinafter, the angle correction process by the angle correction unit 138 will be described in detail.

  FIG. 20 shows an example of the midline of the fetal head and an ellipse that approximates the fetal head (approximate ellipse). The high-intensity region 140 corresponds to an image representing the midline of the fetal head (the midline 66 in FIG. 4). The approximate ellipse 142 is an ellipse that approximates the fetal head calculated by the processing of the first embodiment described above. The short axis straight line 144 is a straight line obtained by extending the short axis of the approximate ellipse 142, and the long axis straight line 146 is a straight line obtained by extending the long axis of the approximate ellipse 142. The center position 148 is the center position of the approximate ellipse 142.

  First, the angle correction unit 138 sets the movement range 150 of the center position 148 on the short axis straight line 144. The movement range 150 is a range in which the center position 148 can move in the minor axis direction. The length of the movement range 150 in the minor axis direction is, for example, a predetermined length 2L, and the movement range 150 is set to a range of ± L in the minor axis direction with respect to the center position 148 before movement. Yes. The length L is a% (for example, 2.5%) of the length of the short axis of the approximate ellipse 142, for example. In order to improve processing efficiency, the length of the movement range 150 is defined. Of course, the movement range 150 may be set to the entire short axis range of the approximate ellipse 142.

  Next, the angle correction unit 138 sets the reference area 152 on the long axis straight line 146. The reference area 152 has a rectangular shape extending in the long axis direction and having a length M and a width W in the short axis direction. The length M is, for example, b% (for example, 80%) of the length of the long axis of the approximate ellipse 142. Of course, the length M may be the length of the major axis of the approximate ellipse 142 itself. The width W is, for example, the length of c pixels (for example, the length of 5 pixels). Of course, the width W may be the length of one pixel.

  Next, the angle correction unit 138 rotates the center position 148 after moving the reference area 152 extending in the long axis direction at the center position 148 after moving while moving the center position 148 within the movement range 150. The evaluation value in the reference area 152 is calculated at each position and each rotation angle within the movement range 150 while rotating within the range of the rotation angle ± α with the long-axis straight line 146 after movement as a reference. The evaluation value is, for example, the sum of luminance values in the reference area 152. The rotation angle α is, for example, 20 °. Of course, the value of the rotation angle α is merely an example, and another value may be used, and the angle correction unit 138 has a reference area over a range of rotation angle ± 90 ° (that is, a range of all rotation angles). 152 may be rotated. The angle correction unit 138 moves within the moving range 150 by a predetermined number of pixels (for example, one pixel at a time) while moving the center position 148 and within a range of the rotation angle ± α. Further, the evaluation value in the reference area 152 is calculated while rotating the reference area 152 by each degree (for example, by 1 °). Of course, the movement amount (number of pixels) is merely an example, and other movement amounts may be used. Moreover, the angle is only an example, and other angles may be used.

  The angle correction unit 138 corrects the angle of the approximate ellipse 142 with respect to the fetal head (skull) using the luminance value sum obtained at each position and each rotation angle and the variance of the luminance value sum. Specifically, the angle correction unit 138 calculates the luminance value sum in the reference area 152 at each rotation angle for each position in the movement range 150, and the luminance at each rotation angle for each position in the movement range 150. Calculate the variance of the sum of values. Next, the angle correction unit 138 determines a position having the largest variance, determines a rotation angle having the largest sum of luminance values at the position, and corrects the angle of the approximate ellipse 142 to the rotation angle.

  Hereinafter, the angle correction processing by the angle correction unit 138 will be described in more detail with specific examples.

  FIG. 20 shows an approximate ellipse 142 in a state before the center position 148 moves. The center position 148 at this time is arranged at an initial position (position A1) within the movement range 150. The angle correction unit 138 moves the reference area 152 within the range of the rotation angle ± α with respect to the long-axis straight line 146 (for example, within the range of the rotation angles φ1 to φn) with the center position 148 disposed at the position A1. , The luminance value sum in the reference area 152 is calculated at each rotation angle. Thereby, the luminance value sum total T1 to Tn for each of the rotation angles φ1 to φn is obtained at the position A1. The angle correction unit 138 calculates the variance S1 of the luminance value sums T1 to Tn obtained at the position A1.

  Next, the angle correction unit 138 moves the center position 148 along the short axis straight line 144 within the movement range 150. FIG. 21 shows the approximate ellipse 142 after the center position 148 has moved. At this time, the center position 148 is assumed to be disposed at a position A2 within the movement range 150. The angle correction unit 138 moves the reference area 152 within the range of the rotation angle ± α with respect to the long axis straight line 146 (within the range of the rotation angles φ1 to φn) with the center position 148 disposed at the position A2. While rotating, the luminance value total in the reference area 152 is calculated at each rotation angle. Thereby, the luminance value sum total T1 to Tn for each of the rotation angles φ1 to φn is obtained at the position A2. The angle correction unit 138 calculates the variance S2 of the luminance value sums T1 to Tn obtained at the position A2.

  For example, when the positions A1 to An are set in the movement range 150, the angle correction unit 138 moves the center position 148 to each of the positions A1 to An, and performs the above processing at each of the positions A1 to An. Thus, the luminance value sum totals T1 to Tn and the variance are calculated at each of the positions A1 to An. Thereby, dispersion | distribution S1-Sn in position A1-An is obtained.

  The angle correction unit 138 specifies the largest variance among the variances S1 to Sn obtained at the positions A1 to An. For example, it is assumed that the variance Sc is the largest variance among the variances S1 to Sn. Next, the angle correction unit 138 specifies a position (referred to as position Ac) within the movement range 150 where the variance Sc is obtained. Next, the angle correction unit 138 specifies the largest luminance value total (referred to as luminance value total Tc) among the luminance value totals T1 to Tn obtained at the position Ac. The angle correction unit 138 specifies a rotation angle (referred to as a rotation angle φc) from which the luminance value total Tc is obtained. As a result, the position Ac at which the variance Sc is obtained, and the rotation angle φc at which the luminance value total Tc is obtained at that Ac are specified.

  The angle correction unit 138 matches the angle of the major axis of the approximate ellipse 142 with the rotation angle φc. In other words, the angle correction unit 138 rotates the approximate ellipse 142 so that the inclination of the major axis of the approximate ellipse 142 matches the inclination of the straight line having the rotation angle φc. Thereby, the angle of the approximate ellipse 142 is corrected. FIG. 22 shows an approximate ellipse 142 with the angle corrected. The short axis straight line 153 is a straight line obtained by extending the short axis of the approximate ellipse 142 in the angle corrected state, and the long axis straight line 154 is a straight line obtained by extending the long axis of the approximate ellipse 142 in the angle corrected state. The inclination of the long-axis straight line 154 matches the inclination of the straight line having the rotation angle φc. As shown in FIG. 22, the major axis (major axis straight line 154) of the approximated ellipse 142 whose angle has been corrected is arranged in parallel or substantially in parallel (the angular difference is slight) to the high luminance region 140 representing the midline.

  The large dispersion of the luminance value sums T1 to Tn indicates that the variation of the luminance value sums T1 to Tn is large. The luminance value sum increases as the reference area 152 matches the midline, and the luminance value sum decreases as the reference area 152 does not match the midline. The fact that the variance of the luminance value summation at a certain position Ax is small (the variation is small) indicates that the variation of the luminance value summations T1 to Tn is small within the range of the rotation angles φ1 to φn at that position Ax. Yes. If a part or all of the midline is included in the range of the rotation angles φ1 to φn at the position Ax, the luminance value sum is large at the rotation angle at which the reference area 152 intersects the midline (intersection). The larger the area to be, the larger the luminance value sum), and it is presumed that the luminance value sum is reduced at a rotation angle at which the reference area 152 does not intersect the median line. On the other hand, when there is no median line within the range of the rotation angles φ1 to φn, it is estimated that the luminance value sum at each rotation angle is small, and therefore when the variation of the luminance value sums T1 to Tn, that is, the variance is small. Guessed. Thus, when a part or all of the median line is included in the range of the rotation angles φ1 to φn at a certain position Ax, the variance of the luminance value sums T1 to Tn obtained within the range is estimated to be large. Is done. From this, it can be estimated that there is a high possibility that a midline exists in the range of the rotation angles φ1 to φn at the position Ax where the dispersion of the luminance value sums T1 to Tn becomes large.

  Further, since the luminance value total increases as the reference area 152 matches the midline, the luminance value total becomes maximum within the range of the rotation angles φ1 to φn at the position Ax where the variance of the luminance value totals T1 to Tn increases. It can be estimated that the rotation angle is closer to the angle in the direction in which the median line extends.

  Summarizing the above, it is estimated that the rotation angle φc at which the luminance value sum is maximum at the position Ac where the variance of the luminance value sums T1 to Tn is maximum is the direction in which the median line extends, and the angle correction unit 138 makes the angle of the approximate ellipse coincide with the rotation angle. Thereby, the angle of the approximate ellipse is corrected, and the inclination of the approximate ellipse is close to the inclination of the fetal head (skull).

  FIG. 23 shows the variance of the luminance value sums T1 to Tn at each position. The horizontal axis indicates the position in the short axis direction (position within the range of + L to -L), and the vertical axis indicates the variance of the luminance value sums T1 to Tn at each position in the short axis direction. For example, when the dispersion curve 156 is obtained, a position Ac at which the dispersion is maximum in the dispersion curve 156 is specified.

  FIG. 24 shows the luminance value sums T1 to Tn at the position Ac. The horizontal axis indicates the rotation angle (range of + α to −α), and the vertical axis indicates the total luminance value. For example, when the luminance value total curve 160 is obtained at the position Ac, the rotation angle φc at which the luminance value total is maximum in the luminance value total curve 160 is specified. The rotation angle φc is estimated to be an angle in the direction in which the median line extends, and the angle correction unit 138 corrects the angle of the approximate ellipse to the rotation angle φc.

  As described above, according to the second embodiment, the direction in which the midline of the fetal head extends is detected, and the direction of the long axis of the approximate ellipse coincides with the direction in which the midline extends. The angle of the approximate ellipse is corrected. Therefore, since the inclination of the approximate ellipse is close to the inclination of the fetal head (skull), the measurement accuracy of the fetal head is further improved.

  The angle correction unit 138 may weight the dispersion curve 156 according to the position in the minor axis direction. For example, the angle correction unit 138 decreases the weighting coefficient toward a position farther from the original center position 148 (position A1) before the movement, multiplies the weighting coefficient by the dispersion curve 156, and distributes the weighted dispersion in the dispersion curve. The position Ac that maximizes may be specified. That is, the angle correction unit 138 applies weighting processing according to the movement amount of the center position 148 to the dispersion curve. Since the center position 148 before movement is evaluated to be closer to the center position of the fetal head, the position of the midline may be detected with higher accuracy as the position is closer to the center position 148 before movement. There is. Therefore, by increasing the weighting coefficient closer to the center position 148 before movement, the direction of the median line can be detected with higher accuracy.

  Also, as in the dispersion curve 158 shown in FIG. 23, when the maximum value of the dispersion is less than or equal to the predetermined threshold value Sth, the angle correction unit 138 obtains the position Ac at which the maximum value was obtained. The rotation angle at which the maximum luminance value sum was obtained among the luminance value sums T1 to Tn obtained at the initial position A1 (the center position 148 not moved) without using the obtained luminance value sums T1 to Tn. In addition, the angle of the approximate ellipse may be corrected.

  Further, in the luminance value summation curve indicating the luminance value summations T1 to Tn obtained at the position Ac at which the dispersion is maximum, as in the luminance value summation curve 162 shown in FIG. 24, the maximum value of the luminance value summation. Is equal to or less than a predetermined threshold Tth, the angle correction unit 138 does not have to correct the angle of the approximate ellipse. This is because a significant rotation angle was not detected.

  Note that the angle correction unit 138 does not correct the center position 148 of the approximate ellipse 142 because the center position of the approximate ellipse calculated by the region calculation unit 26 is estimated to be close to the center position of the fetal head. . In the example shown in FIG. 22, the center position 148 of the approximate ellipse 142 is not corrected, and the center position 148 of the approximate ellipse 142 in the angle-corrected state is the center position of the approximate ellipse 142 in the state before the angle correction. It is the same position as 148. Thereby, it is possible to make the inclination of the approximate ellipse 142 close to the inclination of the fetal head in a state where the center position 148 of the approximate ellipse 142 is arranged at or near the center position of the fetal head. Of course, the angle correction unit 138 may correct the center position 148. For example, the position Ac may be adopted as the center position 148.

  In the above example, the luminance value sum is used, but pattern matching, angle estimation by principal component analysis, or the like may be used as a correction method. Further, although the variance of the luminance value sum is used, another value (for example, standard deviation) may be used as an index representing variation. Further, angle correction may be performed using the maximum value of the luminance value sum without using variance.

  In the second embodiment, the approximate ellipse calculated by the process according to the first embodiment is targeted, but the ellipse that is the target of the angle correction process by the angle correction unit 138 is the process according to the first embodiment. It may be an ellipse calculated by other processing. For example, when an ellipse that approximates the fetal head is calculated by applying a known technique, angle correction processing may be applied to the ellipse. In this case, the image processing unit 22A may not include the position calculation unit 24 and the region calculation unit 26, and the angle correction unit 138 applies the angle correction process to the ellipse calculated by other processes. For example, when an approximate ellipse is obtained by applying an ellipse approximation technique according to a known technique, the angle correction unit 138 corrects the angle of the approximate ellipse by applying an angle correction process to the approximate ellipse. May be.

<Third Embodiment>
Hereinafter, an ultrasonic diagnostic apparatus according to a third embodiment of the present invention will be described. In the third embodiment, it is assumed that the fetal annulus is the fetal abdomen, and ultrasonic waves are transmitted to and received from the fetal abdomen to generate a B-mode tomographic image representing the fetal abdomen. The ultrasonic diagnostic apparatus according to the third embodiment has the same configuration as the ultrasonic diagnostic apparatus according to the first embodiment.

  Also in the third embodiment, as in the first embodiment, the position calculation unit 24 applies a template matching to the B-mode tomographic image to obtain a fetal abdominal position candidate (for example, a center position candidate). Calculate. The region calculation unit 26 performs polar coordinate conversion on the B-mode tomographic image using the center position candidate calculated by the position calculation unit 24, and approximates the fetal abdomen based on the image (development image) after the polar coordinate conversion (approximation). Ellipse) is calculated. The measuring unit 28 measures the fetal abdomen based on the B-mode tomographic image using the approximate ellipse. The measurement unit 28 includes, for example, an anterior posterior trunk diameter (APTD), a transverse trunk diameter (TTD), an abdominal circumference (AC), a fetal trunk area (FTA). parameters such as sectional area) can be measured. Of course, parameters other than these may be measured.

  Hereinafter, the ultrasonic diagnostic apparatus according to the third embodiment will be described in detail.

  For example, the template generation unit 34 generates an annular template having a diameter and a width based on the fetal pregnancy days GD. Specifically, the template generation unit 34 calculates the average value (statistical value) of the diameter R of the fetal abdomen and its variation (for example, standard deviation SD) based on the number of days of pregnancy GD, and calculates the average value of the diameter R. An annular template (for example, template 74 shown in FIG. 5) having (statistical value) as diameter A and standard deviation SD as width B is generated. As an average value (statistical value) of the diameter R of the fetal abdomen, for example, an average value (statistical value) of a transverse trunk diameter (TTD) is used, and a variation in the average value of the diameter R (eg, standard deviation SD) As such, a variation (for example, standard deviation SD) of the TTD average value (statistical value) may be used.

  As in the first embodiment, the edge-enhanced image generation unit 36 applies, for example, a DoG filter to the B-mode tomographic image to generate an edge-enhanced image in which the edges of the fetal abdomen are enhanced.

  In the processing using the DoG filter, a positive value is used. However, the fetal abdomen has fewer bones than the fetal head, and information on the edge portion is extremely difficult to be drawn. For this reason, when some edge portions are strong or there are very many weak edge portions, the edge portions of the fetal abdomen are not easily depicted due to the influence of those portions. Therefore, the edge-enhanced image generation unit 36 calculates the average value of positive values obtained by applying the DoG filter, sets the value of a pixel having a luminance value equal to or higher than the average value to “1”, A value of a pixel having a luminance value less than the average value is set to “0”. As described above, the edge-enhanced image generation unit 36 binarizes the image after the DoG filter is applied. Thereby, weak edge portions are removed. Next, the edge-enhanced image generation unit 36 applies a median filter (1 × 3) in the horizontal direction to the binarized image to fill the hollow portion generated by the binarization process. Next, the edge-enhanced image generation unit 36 applies weighting processing to the central portion of the image to which the median filter is applied. The weighting function is, for example, a function (normal distribution) in which the weighting coefficient at the center of the image is 1.0 and the value decreases toward the end of the image. As a result, it is possible to emphasize a portion that is likely to be a circle existing near the center of the image.

  Similar to the first embodiment, the matching processing unit 38 applies the matching process using the template to the edge-enhanced image, so that the fetal abdominal position candidate, specifically, the fetal abdominal part can be detected. The center position candidate is calculated.

  Also in the third embodiment, as in the first embodiment, polar coordinate conversion by the conversion unit 40, mask processing by the mask setting unit 42, route search processing by the search unit 44, inverse conversion by the inverse conversion unit 46, and an ellipse calculation unit Ellipse calculation processing by 48 is performed. The conversion unit 40 performs polar coordinate conversion of the B-mode tomographic image using the above-described candidate for the center position of the fetal abdomen. In addition, the process by the validity determination part 50 and the process by the center part mask part 52 may be performed.

  The measuring unit 28 measures the fetal abdomen using the approximate ellipse and the B-mode tomographic image obtained by the above processing.

  In the above example, the matching process is performed using the annular template 74 (see FIG. 5), but a template having a different shape may be used. The fetal abdomen has fewer bones than the fetal head, and the edge information is extremely difficult to depict. In addition, as described in the first embodiment, the B-mode tomographic image includes a portion where an edge portion is difficult to be drawn (for example, a region in which the declination θ direction is ± 90 ° in the polar coordinate system). Information from that part may be noise.

  Therefore, in the third embodiment, a template having a shape in which a part of the ring is missing (a shape in which a part of the ring is cut) may be used. FIG. 25 shows an example of the template. The template 164 is a template generated by the template generation unit 34. The shape of the template 164 is a shape assuming the fetal abdomen. Template 164 includes an upper arch portion 166 and a lower arch portion 168 having an arcuate shape. The upper arch portion 166 and the lower arch portion 168 are arranged in opposite directions. Spaces 170 and 172 are formed between both ends of the upper arch portion 166 and both ends of the lower arch portion 168. The positions of the spaces 170 and 172 correspond to portions where the edge portion is difficult to be drawn in the B-mode tomographic image, that is, regions where the declination θ direction is ± 90 ° in the polar coordinate system. A region 174 surrounded by the upper arch portion 166 and the lower arch portion 168 is a region assuming the inside of the fetal abdomen. Since it can be assumed that the fetal abdomen has an elliptical shape close to a circle, a template 164 simulating the shape is used in template matching. When the template 164 is regarded as an annular shape including the spaces 170 and 172, the diameter A of the ring is an average value (statistical value) of the diameter R calculated using the number of gestation days (for example, a TTD average value). (Statistical value)), and the width B of the annular ring, that is, the width B of the upper arch part 166 and the lower arch part 168, is the standard deviation of the diameter R average value (statistical value) calculated using the number of days of pregnancy. SD. Thus, the template 164 is a template that reflects the size estimated from the number of days of pregnancy and the variation in the size. Further, in the template 164, spaces 170 and 172 are formed in the portion where the edge portion is difficult to be drawn in the B-mode tomographic image, so that it is difficult to detect noise in the matching process. Since the template 164 having such characteristics reflects the actual size and shape of the fetal abdomen, the template 164 has a shape in which noise is difficult to be detected. It is easy to detect a position closer to the center position as a candidate for the center position.

  Hereinafter, the matching process using the template 164 will be described in detail with reference to FIG. FIG. 26 shows an example of the edge enhanced image.

  For example, the position and angle of the probe 10 are adjusted by the user so that the fetal abdomen is represented in the B-mode tomographic image, and a B-mode tomographic image representing the fetal abdomen is generated. The edge-enhanced image generation unit 36 generates an edge-enhanced image 176 from the B-mode tomographic image. In FIG. 26, the hatched area indicates a high luminance area. The fetal abdomen has fewer bones than the fetal head, and the edge portion of the fetal abdomen is hardly depicted in the edge-enhanced image 176 (B-mode tomographic image). In the example shown in FIG. 26, an image 178 representing a part (upper and lower) of the fetal abdomen is shown in the edge-enhanced image 176, but since many other tissues are also drawn, it is different from the fetal head. Therefore, it is difficult to distinguish the fetal abdomen. Further, as described above, since the ultrasonic beam is transmitted and received in parallel on the side surface of the fetal abdomen (the region where the declination θ direction is ± 90 ° in the polar coordinate system), it is difficult to depict that portion. In order to cope with this, median filter processing and weighting processing are performed on the B-mode tomographic image in addition to DoG filter processing, so that a region having a shape close to a circle can be easily drawn. Further, the template 164 is used so that noise is not easily detected in the matching process.

  Similar to the first embodiment, the matching processing unit 38 applies a matching process using the template 164 to the edge-enhanced image 176 to calculate a fetus abdominal center position candidate. Specifically, the matching processing unit 38 calculates the similarity between the template 164 and the fetal abdomen image at each position while changing the position of the template 164 on the edge enhanced image 176 as indicated by an arrow 180. As a result, a two-dimensional map of similarity (similarity map) is generated. The similarity is calculated by the above equation (1). The matching processing unit 38 may perform filter processing using two types of filters, as in the first embodiment. The matching processing unit 38 detects a position where the similarity is maximum in the similarity map as a candidate for the center position of the fetal abdomen. This center position candidate is used in polar coordinate conversion.

  The region calculation unit 26 calculates an ellipse (approximate ellipse) that approximates the fetal abdomen by performing polar coordinate conversion, mask processing, route search processing, inverse conversion, and ellipse calculation processing using the center position candidates described above. To do. The measuring unit 28 measures the fetal abdomen using the approximate ellipse and the B-mode tomographic image.

  As described above, according to the third embodiment, a candidate for the center position of the fetal abdomen can be obtained by performing the matching process using the template that approximates the shape of the fetal abdomen. By performing polar coordinate conversion using the center position candidate, polar coordinate conversion reflecting a more accurate center position is possible, and thus the accuracy of ellipse approximation is improved. Thereby, since the ellipse approximated by the fetal abdomen is obtained, the measurement accuracy of the fetal abdomen using the ellipse is improved.

  As described above, the fetal abdomen has fewer bones than the fetal head, and the edge portion is difficult to depict. Therefore, the area to which the route search process is applied in the developed image may be narrowed by the mask process. FIG. 27 shows an example of a developed image to which the mask is applied. The developed image 182 is an image generated by polar coordinate conversion by the conversion unit 40. Specifically, the developed image 182 is an image generated by performing a polar coordinate conversion of the B-mode tomographic image using the fetus abdominal center position candidate obtained by the matching processing unit 38. The origin (θ = 0, r = 0) of the developed image 182 is a position corresponding to the candidate for the center position of the fetal abdomen obtained by the matching processing unit 38. An image 184 in the developed image 182 is an image corresponding to the image 178 represented in the B-mode tomographic image.

  Masks 186, 188, and 190 are set in the developed image 182 by the mask setting unit 42. Similar to the first embodiment, in the developed image 182, a mask 186 having a width of ± 45 ° in the angle θ direction is set at a position where the angle θ is 90 ° and a position where the angle θ is −90 °. Mask 186 has a shape extending in the direction of distance r from the position where distance r = 0. The position where the angle θ is ± 90 ° is a position corresponding to the side surface of the fetal abdomen, and since the ultrasonic beam is easily transmitted and received in parallel in that portion, it is difficult to depict that portion. If that portion is included in the route search target area, there is a possibility that an erroneous search will occur in the route search processing. In addition, if another tissue having high luminance is depicted in a portion where the fetal abdomen is difficult to depict, an erroneous search may occur in the route search process. In order to cope with this, a mask 186 having a width in the direction of the angle θ is set in a portion where the fetal abdomen is difficult to be drawn (position where the angle θ is ± 90 °). As a result, the portion is excluded from the route search area, so that occurrence of erroneous search can be suppressed or prevented.

  Further, the mask setting unit 42 estimates a range where there is a high possibility that the edge part (boundary) of the fetal abdomen does not exist based on the mean diameter R (statistical value) of the fetal abdomen estimated from the number of days of pregnancy. Masks 188 and 190 are set in the range. The mask 188 is a mask set in a predetermined range (distance a) from the position where the distance r = 0 (mask set in the range 0 to a), and the mask 190 is an edge portion of the fetal abdomen. Is a mask set in a range assumed to be outside (a mask set in the range of Ra). By setting the mask in this way, it is possible to remove or eliminate noise from an area where it is assumed that there is no edge portion of the fetal abdomen, thereby suppressing or preventing the occurrence of a false search.

  Note that the mask setting unit 42 may apply weighting processing to the developed image instead of setting the mask, as in the first modification of the first embodiment, or the modification according to the first embodiment. Similar to 2, a mask corresponding to the type of the probe 10 may be set in the developed image.

  The ultrasonic diagnostic apparatus according to the third embodiment may have the same configuration as the ultrasonic diagnostic apparatus according to the second embodiment. In this case, the angle correction unit 138 corrects the angle of the ellipse that approximates the fetal abdomen. For example, the angle may be corrected based on the spine of the fetus.

  The template 164 used in the third embodiment may be used in the first embodiment. As described above, in the B-mode tomographic image, the side surface of the fetal head is hard to be drawn, and the computation processing of the candidate for the center position of the fetal head may be affected by noise caused by that portion. By using the template 164 lacking a part of the ring, the influence of noise on the side surface of the fetal head can be reduced, so that the detection accuracy of the candidate for the center position of the fetal head is increased. Also, the masks 188 and 190 shown in FIG. 27 may be set in the developed image according to the first embodiment.

22 image processing units, 24 position calculation units, 26 area calculation units, 28 measurement units, 34 template generation units, 36 edge-enhanced image generation units, 38 matching processing units, 40 conversion units, 42 mask setting units, 44 search units, 46 Inverse conversion unit, 48 ellipse calculation unit, 50 validity determination unit, 52 center mask unit, 54 end determination unit, 56 center update unit, 138 angle correction unit.

Claims (8)

Based on the tomographic image obtained by transmitting and receiving ultrasonic waves to the fetus in the mother body, an ellipse computing means for computing an ellipse corresponding to the annular part of the fetus,
While moving the center of the ellipse calculated by the ellipse calculating means in one axial direction of the ellipse and rotating the reference area extending in the other axial direction of the ellipse at the center after the movement, the one axial direction Angle correction means for calculating an evaluation value in the reference area at each position and each rotation angle, and correcting the angle of the ellipse with respect to the annular portion using the evaluation value;
An ultrasonic diagnostic apparatus comprising: The ultrasonic diagnostic apparatus according to claim 1,
The angle correction means corrects the angle of the ellipse based on the evaluation value obtained at each position and each rotation angle and the variance of the evaluation value.
An ultrasonic diagnostic apparatus. The ultrasonic diagnostic apparatus according to claim 2,
The angle correction means moves the center of the ellipse in the one axial direction, rotates the reference area at the position of the movement destination, and calculates a sum of luminance values in the reference area at each rotation angle. Calculated as an evaluation value, calculates the sum of the luminance values in the reference area at each rotation angle for each position in the one axial direction, and calculates the luminance at each rotation angle for each position in the one axial direction. The variance of the sum of the values is calculated, the position in the one axial direction where the variance is the largest is determined, the rotation angle at which the sum of the luminance values is the largest is determined, and the angle of the ellipse is corrected to the rotation angle ,
An ultrasonic diagnostic apparatus. The ultrasonic diagnostic apparatus according to claim 3.
The angle correction means applies a weighting process according to the amount of movement of the center of the ellipse in the one axial direction to the variance of the sum of the luminance values.
An ultrasonic diagnostic apparatus. The ultrasonic diagnostic apparatus according to claim 3 or 4,
The angle correction means is configured such that when the variance at each position in the one axial direction is equal to or less than a predetermined threshold, the ellipse is rotated at a rotation angle at which the sum of luminance values is maximized when the center of the ellipse is not moved. Correct the angle of
An ultrasonic diagnostic apparatus. The ultrasonic diagnostic apparatus according to claim 1,
The annular portion is the fetal head;
The one axial direction is a minor axis direction of the ellipse,
The other axial direction is the major axis direction of the ellipse,
The angle correction means corrects the angle of the ellipse by calculating the direction in which the midline of the fetal head extends based on the evaluation value.
An ultrasonic diagnostic apparatus. The ultrasonic diagnostic apparatus according to any one of claims 1 to 6,
By applying template matching to the tomographic image, position calculating means for calculating a candidate for the center position of the fetal annular region,
Conversion means for converting the tomographic image into polar coordinates using the center position candidate as the origin of the polar coordinate system;
Search means for searching for a route from the image after the polar coordinate transformation;
Inverse transformation means for inversely transforming the path;
Further including
The ellipse calculation means calculates an ellipse corresponding to the path after inverse transformation as an ellipse corresponding to the annular portion,
An ultrasonic diagnostic apparatus. Computer
Based on a tomographic image obtained by transmitting and receiving ultrasonic waves to the fetus in the mother's body, an ellipse calculation means for calculating an ellipse corresponding to the fetal ring part,
While moving the center of the ellipse calculated by the ellipse calculating means in one axial direction of the ellipse and rotating the reference area extending in the other axial direction of the ellipse at the center after the movement, the one axial direction Angle correction means for calculating an evaluation value in the reference area at each position and each rotation angle, and correcting the angle of the ellipse with respect to the annular portion using the evaluation value;
A program characterized by functioning as