JP2013097586A - Device for generating moving image of internal organ and method for generating moving image of internal organ - Google Patents

Abstract

PROBLEM TO BE SOLVED: To generate a three-dimensional moving image of a prescribed portion of an internal organ which is hardly displayed in a cross-sectional image vividly, by automatically obtaining images of the prescribed portion from a plurality of cross-sectional images.SOLUTION: A device for generating a moving image of an internal organ includes cross-sectional image capturing means 2 to 5 and 7 to 14 for continuously capturing cross-sectional images of an internal organ in respective photographing positions and image analysis means 6 for image analysis of the cross-sectional images. The image analysis means 6 adds marks indicative of a prescribed portion set by a user, to all the cross-sectional images by adding the mark to a cross-sectional image captured first among the continuously captured cross-sectional images and using two-dimensional DP to add marks to subsequently captured cross-sectional images, calculates centroid positions of the marks with respect to all the cross-sectional images, obtains coordinate information of the portion in a three-dimensional space on the basis of the calculated centroid positions, obtains a three-dimensional spline curve on the basis of the coordinate information to obtain a three-dimensional shape of the portion, and generates a three-dimensional moving image of the portion.

Description

  The present invention relates to a moving image generation apparatus and a moving image generation method for a body organ, and more particularly, to a predetermined part of the organ based on a plurality of captured images obtained by continuously capturing the body organ at regular time intervals. The present invention relates to a moving image generation apparatus for a body organ and a moving image generation method for a body organ capable of providing the above state as a three-dimensional moving image.

  Conventionally, general organ imaging such as ultrasound CT (Computed Tomography), weak X-ray CT, and MRI (Magnetic Resonance Imaging) is used to determine the state of internal organs. There is known a method of photographing a state of a body organ as a two-dimensional cross-sectional image (determining a cross-sectional image by image analysis) using an organ photographing method by an apparatus. The doctor determines the state of the organ by comparing and examining a plurality of two-dimensional images having different imaging timings or imaging positions.

  Here, since the above-described organ imaging device captures an image of a cross section of the body or an image from one direction, it is not easy to specify the internal organ to be diagnosed from the entire image. There was a problem.

  Further, in the above-described organ imaging apparatus, there are internal organs that are difficult to be clearly (clearly) displayed (captured) in a cross-sectional image. For example, in weak X-ray CT, when an organ is inside a bone part, there is a problem that the organ part is blocked by the bone. In addition, the ultrasonic CT has a problem that the image resolution is not sufficient and the image becomes unclear. Furthermore, in MRI, there is no blockage of the organ part by the bone, but the organ muscles (for example, heart muscle) and tendons (for example, the tricuspid valve of the heart) appear black in the image. There was a problem that it was difficult to specify.

  When MRI is used, regarding an organ filled with blood, such as the heart atrium, the presence of the organ can be projected (taken) in a cross-sectional image relatively clearly. However, tendons and muscles that are not filled with blood are difficult to be clearly displayed (captured) on the MRI image. Therefore, the doctor has to improve the MRI image by adjusting the brightness level, etc., and then predict / specify the unclear part and diagnose the part to be diagnosed. For this reason, there is a problem that it is difficult to determine the state of a part such as a tendon or muscle, and the determination may vary depending on the experience and skill of the doctor.

  Furthermore, since the above-described organ imaging apparatus captures a two-dimensional image of an organ, it is difficult to visualize the activity state of an organ that originally performs a three-dimensional movement as a three-dimensional movement. there were.

  For example, if the movement of the tricuspid valve of the heart can be observed three-dimensionally, it is possible to easily determine whether or not the valve is operating normally. However, as described above, the tricuspid valve is a tendon and is a part that is difficult to display clearly in a CT image or an MRI image. It was not easy to identify the site.

  For this reason, studies have been made to estimate the position of the tricuspid valve from a plurality of CT images and manually estimate the shape of the tricuspid ring (see Non-Patent Document 1, for example).

Shota Fukuda, et al, "Three-Dimensional Geometry of the Tricuspid Annulus in Healthy subjects and in Patients with Functional Tricuspid Regurgitation -A Real-time, 3-Dimensional Echocardiographic Study-", American Heart Association, Circulation. 2006, 114 (Suppl I): I-492-498

  However, in the above-described method of Non-Patent Document 1, the position of the tricuspid valve ring is estimated in each of a plurality of CT images, and a three-dimensional position is calculated from each of the plurality of CT images on which the positions are estimated. Thus, it is necessary to obtain the three-dimensional position of the tricuspid valve ring by hand. For this reason, much labor is required only by obtaining a three-dimensional image of the tricuspid valve ring. Furthermore, in order to make a three-dimensional image of the tricuspid ring thus obtained a moving image, a further burden is required, and it is very difficult to obtain a three-dimensional moving image.

  The present invention has been made in view of the above problems, and a predetermined part of an internal organ that is not necessarily clearly displayed in a cross-sectional image such as an MRI image is automatically obtained in a plurality of cross-sectional images, and the state in the corresponding predetermined part It is an object of the present invention to provide a moving image generating apparatus for a body organ and a moving image generation method for a body organ capable of showing a change as a three-dimensional moving image.

  In order to solve the above problems, a moving image generation apparatus for a body organ according to the present invention can move the imaging position of a cross-sectional image in the internal organ in the stacking direction of the cross-sectional images, and the cross-sectional image is predetermined. Cross-sectional image photographing means capable of continuously photographing at time intervals, and the cross-sectional images photographed by the cross-sectional image photographing means, the photographing order of the cross-sectional images photographed continuously at the same photographing position, Recording means for recording in association with the photographing position of the cross-sectional image, photographing control means for controlling the photographing position and photographing timing in the cross-sectional image photographing means, and image analysis of the cross-sectional image recorded in the recording means The image analysis means to be performed and the user sets the mark position with respect to the cross-sectional image, so that the position of the predetermined site in the internal organ is matched with the mark Mark adding means capable of recording in a cross-sectional image, wherein the imaging control means causes the cross-sectional image imaging means to continuously take an image of the cross-sectional image of the internal organ at one imaging position; The continuous section images are taken for each of the photographing positions by gradually shifting the photographing positions in the stacking direction range in which the internal organs are accommodated, and the image analysis means is a cross section recorded in the recording means. An image is read out, and the cross-sectional image first taken at the photographing position among the plurality of cross-sectional images taken continuously at one photographing position is set at the mark position set by the mark adding means. Next to the cross-sectional image to which the mark is added based on the mark position of the cross-sectional image to which the mark is added using a two-dimensional continuous DP. By adding a mark to the photographed cross-sectional image, a mark is automatically added to all cross-sectional images at all photographing positions continuously photographed, and the barycentric position of the mark added to the cross-sectional image is determined. , Calculating all the cross-sectional images based on the pixel group of the mark, and based on the calculated barycentric position, the coordinates of the part of the internal organ in the three-dimensional space composed of the image plane of the cross-sectional image and the stacking direction Information is obtained for each cross-sectional image taken in the same shooting order among the cross-sectional images taken continuously at different shooting positions, and three-dimensional based on the coordinate information taken in the same shooting order A spline curve is obtained, the three-dimensional shape of the part is obtained for each photographing order, and the three-dimensional shape of the part obtained for each photographing order is determined according to the photographing order. The time-dependent change in the three-dimensional shape of the part is obtained, and a three-dimensional moving image in the part is generated.

  In addition, the method for generating a moving image of a body organ according to the present invention can move the imaging position of a cross-sectional image in the internal organ in the stacking direction of the cross-sectional images, and the cross-sectional images can be continuously captured at predetermined time intervals. The cross-sectional image photographing means capable of photographing, the photographing order of the cross-sectional images obtained by continuously photographing the cross-sectional images photographed by the cross-sectional image photographing means at the same photographing position, and the photographing position of the cross-sectional image Recording means for recording in association with the above, photographing control means for controlling the photographing position and photographing timing in the cross-sectional image photographing means, image analyzing means for performing image analysis of the cross-sectional image recorded in the recording means, By setting a mark position on the cross-sectional image, the position of a predetermined part in the internal organ is recorded in the cross-sectional image by the mark. In the internal organ moving image generating apparatus having the mark adding means capable of being recorded, the imaging control means causes the cross-sectional image imaging means to continuously capture the cross-sectional image of the internal organ at one imaging position. In addition, the imaging of the continuous cross-sectional images is performed for each imaging position by gradually shifting the imaging position in the stacking direction range in which the internal organs are accommodated, and the image analysis unit is recorded in the recording unit The mark position set by the mark adding means for the first cross-sectional image taken at the photographing position among the plurality of cross-sectional images taken continuously at one photographing position. Before the mark is added based on the mark position of the cross-sectional image to which the mark is added using a two-dimensional continuous DP. By adding a mark to the cross-sectional image photographed next to the cross-sectional image, the mark is automatically added to all cross-sectional images at all photographing positions continuously photographed, and the cross-sectional image is added to the cross-sectional image. The center-of-gravity position of the mark is calculated for all cross-sectional images based on the pixel group of the mark, and the internal organ in the three-dimensional space composed of the image plane of the cross-sectional image and the stacking direction is calculated based on the calculated center-of-gravity position. Is obtained for each of the cross-sectional images photographed in the same photographing order among the cross-sectional images continuously photographed at different photographing positions, and based on the coordinate information photographed in the same photographing order. Obtaining a three-dimensional spline curve, obtaining a three-dimensional shape of the part for each photographing order, and obtaining the three-dimensional shape of the part for each photographing order, By changing according to the imaging order, a temporal change in the three-dimensional shape of the part is obtained, and a three-dimensional moving image in the part is generated.

  In the internal organ moving image generating apparatus and the internal organ moving image generating method according to the present invention, among a plurality of cross-sectional images continuously captured at one imaging position, the first (at the very beginning) at the imaging position. ) A mark is added at the mark position set by the mark adding means to the photographed cross-sectional image, and the mark is added to the mark in the cross-sectional image photographed next to the cross-sectional image to which the mark is added. The two-dimensional continuous DP is added based on the mark position of the cross-sectional image. As described above, it is possible to automatically add marks in order using the two-dimensional continuous DP to all the cross-sectional images at all the photographing positions continuously photographed. Therefore, for example, even in a part of a body organ that is not clearly (clearly) displayed in a cross-sectional image, it is possible to continuously specify a part corresponding to that part and add a mark.

  In this way, even in a part of the internal organ that is not clearly (clearly) shown in the cross-sectional image, the position and state change of the part can be easily identified from the continuous cross-sectional image. It is possible to reduce the difference in the diagnosis result of the site depending on the experience and skill of the doctor.

  In the internal organ moving image generation apparatus and internal organ moving image generation method according to the present invention, the position of the internal organ is obtained based on the center of gravity of the mark added to the cross-sectional image, and the position of the internal organ is determined. Can be obtained as a coordinate position in a three-dimensional space consisting of the image plane of the cross-sectional image and the stacking direction, and a three-dimensional spline curve can be obtained from the coordinate position to obtain the three-dimensional shape of the internal organ. .

  For this reason, as compared with the conventional case where the doctor determines the shape of the part by estimating the three-dimensional shape based on the two-dimensional cross-sectional image regarding the part of the internal organ, It becomes possible to improve the reliability of the three-dimensional shape. In addition, it is possible to reduce the burden of shape determination, and it is possible to improve the diagnostic accuracy.

  Furthermore, in the internal organ moving image generating apparatus and the internal organ moving image generating method according to the present invention, the three-dimensional shape of the part obtained for each imaging order is changed in accordance with the imaging order, so that A three-dimensional moving image in the part can be generated by obtaining a temporal change in the dimensional shape. Therefore, it is possible for a doctor to easily grasp the movement of a part, and it is possible to diagnose internal organs in consideration of not only a three-dimensional stationary shape but also a three-dimensional movement. The accuracy can be improved. In addition, since the movable volume of the part can be easily calculated from the moving image of the three-dimensional part, judgment criteria (such as a test guideline) for statistically judging normal or abnormal based on the movable volume are provided. It can be constructed and used for diagnosis.

  In addition, the internal organ that is expected to have a dramatic effect on diagnostic accuracy and judgment method by obtaining a three-dimensional moving image is the heart. The heart always moves with the heartbeat, but there is a high possibility that some abnormality has occurred in the part where the movement is unnatural. For example, valve function parts such as the pulmonary valve, tricuspid valve, mitral valve, and aortic valve of the heart are difficult to capture clearly (clearly) with ultrasonic CT, weak X-ray CT, MRI, etc. It is difficult to determine the state three-dimensionally because it is difficult to specify the region. For this reason, not only the shape of the valve function unit is three-dimensionally determined but also the moving image is generated by the moving image generating device and the moving image generating method of the internal organ according to the present invention, thereby generating the heart valve It is possible to dramatically improve the diagnostic accuracy and diagnostic method for such diseases.

  The cross-sectional image photographing means means a general organ photographing apparatus such as ultrasonic CT, weak X-ray CT, MRI, for example. A cross-sectional image is an analysis image taken by these organ imaging devices, and means an image showing a cross-sectional state of a body organ of a subject in a slice shape.

  In addition, two-dimensional continuous DP is an image analysis technique that can automatically determine the correspondence of common pixels in two corresponding images. “Ryuichi Oka, two others” on the general scheme of continuous DP -“All-pixel optimal matching for image spotting”, IEICE Technical Report, IEICE Technical Report, PRMU2010-87, IBISML2010-59 (2010-09) ” is there. Further, the two-dimensional continuous DP is described in Japanese Patent Application Laid-Open No. 2010-165104, which corresponds to a technique already known.

  Further, in the above-described internal organ moving image generation apparatus or internal organ moving image generation method, the image analysis unit adds the mark before calculating the center of gravity position of the mark for all cross-sectional images. You may perform the image process by dilation and erosion with respect to a cross-sectional image.

  Thus, before the image analysis means calculates the center of gravity position of the mark for all the cross-sectional images, the cross-sectional image to which the mark is added performs image processing by dilation and erosion. By this image processing, it becomes possible to remove isolated points (isolated pixels) of pixels constituting the mark added to the cross-sectional image and to connect and fill discontinuous pixels. For this reason, it becomes possible to collect and clarify pixels indicating marks as a group of pixels.

  Note that dilation and erosion are typical image processing called morphological processing. As described above, this process is used for removing isolated points (isolated pixels), connecting discontinuous pixels, and filling holes.

  Furthermore, in the above-described internal organ moving image generating apparatus or internal organ moving image generating method, the imaging control unit includes the internal organ when the cross-sectional image capturing unit continuously captures the cross-sectional image. A cross-sectional image from the start of one beating of the subject to the start of the next beating may be continuously taken.

  The heart, which is expected to have a dramatic effect on the results of diagnosis and determination of internal organs by obtaining a three-dimensional moving image, always moves with heartbeat. For this reason, when generating a three-dimensional moving image in the heart region, a cross-sectional image from the start of one heartbeat to the start of the next heartbeat in the subject's heart is continuously photographed to obtain a three-dimensional image. By generating a dynamic moving image, it is possible to provide a series of movements of the part accompanying the heartbeat as a moving image. For this reason, it is possible to improve the accuracy in diagnosis of a part related to the heart, and it is possible to realize early detection and diagnosis of a heart disease or the like that has been difficult to determine in the past.

  In the internal organ moving image generating apparatus and the internal organ moving image generating method according to the present invention, among a plurality of cross-sectional images continuously captured at one imaging position, the first (at the very beginning) at the imaging position. ) A mark is added at the mark position set by the mark adding means to the photographed cross-sectional image, and the mark is added to the mark in the cross-sectional image photographed next to the cross-sectional image to which the mark is added. The two-dimensional continuous DP is added based on the mark position of the cross-sectional image. As described above, it is possible to automatically add marks in order using the two-dimensional continuous DP to all the cross-sectional images at all the photographing positions continuously photographed. Therefore, for example, even in a part of a body organ that is not clearly (clearly) displayed in a cross-sectional image, it is possible to continuously specify a part corresponding to that part and add a mark.

  In this way, even in a part of the internal organ that is not clearly (clearly) shown in the cross-sectional image, the position and state change of the part can be easily identified from the continuous cross-sectional image. It is possible to reduce the difference in the diagnosis result of the site depending on the experience and skill of the doctor.

  In the internal organ moving image generation apparatus and internal organ moving image generation method according to the present invention, the position of the internal organ is obtained based on the center of gravity of the mark added to the cross-sectional image, and the position of the internal organ is determined. Can be obtained as a coordinate position in a three-dimensional space consisting of the image plane of the cross-sectional image and the stacking direction, and a three-dimensional spline curve can be obtained from the coordinate position to obtain the three-dimensional shape of the internal organ. .

  For this reason, as compared with the conventional case where the doctor determines the shape of the part by estimating the three-dimensional shape based on the two-dimensional cross-sectional image regarding the part of the internal organ, It becomes possible to improve the reliability of the three-dimensional shape. In addition, it is possible to reduce the burden of shape determination, and it is possible to improve the diagnostic accuracy.

  Furthermore, in the internal organ moving image generating apparatus and the internal organ moving image generating method according to the present invention, the three-dimensional shape of the part obtained for each imaging order is changed in accordance with the imaging order, so that A three-dimensional moving image in the part can be generated by obtaining a temporal change in the dimensional shape. Therefore, it is possible for a doctor to easily grasp the movement of a part, and it is possible to diagnose internal organs in consideration of not only a three-dimensional stationary shape but also a three-dimensional movement. The accuracy can be improved. In addition, since the moving volume of the part can be easily calculated from the moving image of the three-dimensional part, a judgment criterion (test guideline) for quantitatively and statistically determining normal or abnormal based on the moving volume. Etc.) can be constructed and used for diagnosis.

  In addition, the internal organ that is expected to have a dramatic effect on diagnostic accuracy and judgment method by obtaining a three-dimensional moving image is the heart. The heart always moves with the heartbeat, but there is a high possibility that some abnormality has occurred in the part where the movement is unnatural. For example, valve function parts such as the pulmonary valve, tricuspid valve, mitral valve, and aortic valve of the heart are difficult to capture clearly (clearly) with ultrasonic CT, weak X-ray CT, MRI, etc. It is difficult to determine the state three-dimensionally because it is difficult to specify the region. For this reason, not only the shape of the valve function unit is three-dimensionally determined but also the moving image is generated by the moving image generating device and the moving image generating method of the internal organ according to the present invention, thereby generating the heart valve It is possible to dramatically improve the diagnostic accuracy and diagnostic method for such diseases.

It is the block diagram which showed schematic structure of the MRI apparatus which concerns on embodiment. It is the block diagram which showed schematic structure of the computer of the MRI apparatus which concerns on embodiment. It is the figure which showed the structure of the heart typically. It is the flowchart which showed the process of CPU which concerns on embodiment. (A) is a figure which showed an example of the slice image of the heart image | photographed with the MRI apparatus, (b) is the positional relationship at the time of shifting and image | photographing the imaging position of the MRI image in a heart from L = 1-6. It is the schematic diagram of the heart which showed. (A) shows the image before the mark processing in the MRI image at time t = 1, and (b) shows the image after the mark processing. (A) shows an MRI image at time t = 1, (b) shows an MRI image at time t = 2, and shows an image whose mark position is obtained by two-dimensional continuous DP, (c) The image after performing dilation and erosion with respect to the MRI image of (b) is shown. (A) is the figure which showed the shape of the tricuspid valve three-dimensionally, (b) is the figure which showed the movable volume of the tricuspid valve.

  Hereinafter, an MRI (Magnetic Resonance Imaging) apparatus (magnetic resonance imaging apparatus) as an example of the moving image generating apparatus for internal organs according to the present invention will be presented and described in detail with reference to the drawings.

  FIG. 1 is a block diagram showing a schematic configuration of the MRI apparatus.

  The MRI apparatus 1 includes a static magnetic field generation magnet 2, a high frequency transmission coil 3, a high frequency reception coil 4, gradient magnetic field coils 5 and 5, a computer 6, a sequencer 7, a high frequency generation device 8, and a modulator. 9, an amplifier 10, a gradient magnetic field generator 11, an amplifier 12, a phase detector 13, an AD converter 14, a display unit 15, and an operation input unit (mark adding means) 16.

  The static magnetic field generating magnet 2 has a role of generating a uniform static magnetic field in the space where the subject X is placed. The gradient magnetic field coils 5 and 5 have a role of generating a gradient magnetic field in three orthogonal directions in the space where the subject X is placed. The gradient magnetic field coils 5 and 5 are driven by the gradient magnetic field generator 11. Power is being supplied.

  The high-frequency generator 8 is a device that outputs a high-frequency signal. The high-frequency signal output from the high-frequency generator 8 is modulated into an appropriate signal by the modulator 9, amplified by the amplifier 10, and applied to the high-frequency transmission coil 3. Supplied. By supplying a high-frequency signal to the high-frequency transmission coil 3, it is possible to irradiate a high-frequency pulse capable of causing nuclear magnetic resonance on the nuclei (protons) in the static magnetic field constituting the living tissue of the subject X. Become.

  When the subject X is irradiated with the high-frequency pulse by the high-frequency transmission coil 3, nuclear magnetic resonance occurs in protons, and a magnetic resonance signal such as an echo signal is output. This magnetic resonance signal is received by the high frequency receiving coil 4 of the MRI apparatus 1, amplified by the amplifier 12, and then detected as an analog signal by the phase detector 13. The detected analog signal is converted into measurement data of a digital signal by the AD converter 14 and output to the computer 6.

  The sequencer 7 sends a command necessary for performing magnetic resonance imaging in accordance with a control instruction of the computer 6, a magnetic field generation system of the MRI apparatus 1 (such as the gradient magnetic field coils 5 and 5 and the gradient magnetic field generator 11) and high-frequency transmission / reception. It is sent to the system (high-frequency generator 8, modulator 9, amplifiers 10, 12, high-frequency transmission coil 3, high-frequency reception coil 4, phase detector 13, AD converter 14, etc.) to control these operations Have.

  By operating the magnetic field generation system and the high-frequency transmission / reception system of the MRI apparatus 1 according to the control instruction of the computer 6, it becomes possible to continuously take cross-sectional images of the internal organs of the subject X. Further, according to the control instruction of the computer 6, it is possible to photograph the cross-sectional image by moving the photographing position of the cross-sectional image little by little in the height direction of the subject X (stacking direction of the photographed cross-sectional images). Yes.

  The static magnetic field generating magnet 2, the high frequency transmitting coil 3, the high frequency receiving coil 4, the gradient magnetic field coils 5 and 5, the sequencer 7, the high frequency generating device 8, and the modulation that perform the processing for taking a cross-sectional image in this way. The detector 9, the amplifier 10, the gradient magnetic field generator 11, the amplifier 12, the phase detector 13 and the AD converter 14 correspond to the sectional image photographing means according to the present invention.

  The display unit 15 has a role of displaying a part image and a part moving image of a body organ image-processed by the computer 6, and is connected to the computer 6. In addition, the operation input unit 16 sets and adjusts the imaging position of the MRI image of the subject X and inputs an information necessary for performing predetermined image processing on the captured MRI image (marks to be described later). For example, position setting).

  As shown in FIG. 2, the computer 6 includes a CPU (imaging control unit, image analysis unit) 20, a ROM 21, a RAM 22, and a storage unit (recording unit) 23. The ROM 21 has a role of recording a program (for example, a program having a flowchart shown in FIG. 4) in which processing contents in the CPU 20 are recorded, initial setting information in the MRI apparatus 1, and the like. The RAM 22 has a role as a work area used in the processing of the CPU 20.

  The CPU 20 performs processing for continuously capturing MRI images of a predetermined part of the organ at predetermined time intervals in accordance with a program recorded in the ROM 21. Further, the CPU 20 specifies a predetermined part from a plurality of MRI images photographed thereafter with reference to a predetermined part marked on the first MRI image by operating a manipulation input unit 16 by a doctor or the like ( A process for setting (marking) a mark by setting a corresponding mark position is performed. Further, the CPU 20 generates a three-dimensional moving image at the predetermined portion by performing the photographing while slightly shifting the photographing position of the MRI image at the predetermined portion.

  Further, the CPU 20 outputs a control instruction to the sequencer 7 according to the processing content. In response to a control instruction from the CPU 20, the sequencer 7 includes a magnetic field generation system (such as the gradient magnetic field coils 5 and 5 and the gradient magnetic field generation apparatus 11) and a high-frequency transmission / reception system (a high-frequency generation apparatus 8, a modulator 9, an amplifier 10, 12, high-frequency transmission coil 3, high-frequency reception coil 4, phase detector 13, AD converter 14 and the like.

  Next, processing of the CPU 20 will be described. In the present embodiment, a case will be described in which the motion of the tricuspid valve of the heart is generated as a moving image as an example of a predetermined part of a body organ. However, detecting the movement of the tricuspid valve as a predetermined part of the organ is merely an example, and the predetermined part of the organ is not necessarily limited to the tricuspid valve of the heart.

  The rate at which valvular disease is involved in cardiac surgery is very high, and there is data that valvularity is attributed to over 70% of surgery. As shown in FIG. 3, there are four types of valves including pulmonary valve, tricuspid valve, mitral valve, and aortic valve in the heart. A tricuspid valve may be the cause. The tricuspid valve is a kind of tendon, and it has been difficult to clearly capture the state of an image taken by CT or MRI. For this reason, it is difficult to visually capture the state of the tricuspid valve disease from an MRI image or the like, and it is even more difficult to observe the movement of the tricuspid valve, which is a criterion for determining whether or not the disease has occurred. ing.

  The MRI apparatus 1 according to the present embodiment performs processing for generating a three-dimensional moving image of a tricuspid valve that is difficult to capture clearly in an MRI image.

  FIG. 4 is a flowchart showing the processing contents in the CPU 20 of the MRI apparatus 1. The CPU 20 captures a plurality of MRI images at a specific position in the heart (for example, an image as shown in FIG. 5A), and performs processing for automatically specifying the position of the tricuspid valve in each MRI image ( Steps S.11 to S.21 in FIG. As shown in FIG. 5B, the CPU 20 measures the tricuspid valve position detection process several times by gradually shifting the position (the shooting position L in FIG. 5B) (showing the shooting position L). Measurement is performed by shifting from L = 1 to L = 6 as shown in FIG. 5 (b) (steps S22 to S24 in FIG. 4). In the MRI apparatus 1 according to the present embodiment, MRI images are taken at a total of six positions by shifting the detection position by 12 mm in the vertical direction of the heart.

  Then, the CPU 20 detects the position of the tricuspid valve from all the measured MRI images, and generates a moving image that three-dimensionally shows the state of the tricuspid valve based on the detected position of the tricuspid valve. (Steps S.25 to 28 in FIG. 4). Details of these processes will be described below.

  First, the CPU 20 reads the initial information recorded in the storage unit 23 and performs initial setting (step S.11). Specifically, the value of t indicating the imaging timing (imaging time) of the MRI image is set to 1 (t = 1), and the value of L indicating the position where the MRI image is captured is set to 1 (imaging position L = 1) Perform processing. This initial set value is temporarily recorded in the RAM 22 and is appropriately updated according to the processing in the CPU 20.

  Next, the CPU 20 outputs a control instruction to the sequencer 7 to operate the magnetic field generation system and the high-frequency transmission / reception system of the MRI apparatus 1, so that the two-dimensional image of the heart of the subject X (the heart is in the imaging range). Are captured by the MRI apparatus 1 (step S.12). When a two-dimensional image is captured by the MRI apparatus 1, a plurality of MRI images at the same position (same imaging position) of the heart are captured at regular time intervals.

  The captured image is temporarily recorded in the storage unit 23. The MRI image taken is the first MRI image taken as the MRI image taken at time t = 1, and the MRI image taken thereafter is the MRI image at time t = 2 and at time t = 3. MRI images,... Are distinguished as MRI images at time t = T. In the present embodiment, t = 1 is set when one beat (beat) is started, and t = T is set immediately before the next beat is started. In this way, by capturing T MRI images from t = 1 to t = T at one beat interval, the state of the heart from the start to the end of one beat can be photographed.

  Next, the CPU 20 performs image mask processing called Laplacian for each of the two-dimensional MRI images recorded in the storage unit 23 (step S.13). By performing image mask processing on the MRI image, the edge of the MRI image can be emphasized, and the image can be sharpened. Note that image mask processing by Laplacian is a method that is already generally known as an image processing technique, and thus detailed description of mask processing will be omitted.

  Then, the CPU 20 makes it possible for a doctor or the like to mark the position of the tricuspid valve via the operation input unit 16 on the MRI image at time t = 1 (mark permission process by the doctor or the like). (Step S.14). In the MRI image, the tricuspid valve, which is a tendon, tends not to be clearly displayed. However, a doctor or the like can estimate the position of the tricuspid valve from the positional relationship such as a part (for example, a part with a lot of blood) that is clearly displayed in the MRI image. Therefore, the CPU 20 enables a doctor or the like to mark the position of the tricuspid valve on the first MRI image (MRI image captured at t = 1) (mark processing is performed by the doctor). Etc.), the position of the tricuspid valve (mark position) in the MRI image can be specified with high accuracy.

  FIG. 6A shows an image before the mark process is performed on the MRI image at t = 1, and FIG. 6B shows an image after the mark process is performed. Since the tricuspid valve is an annular (ring-shaped) tendon, if it is an annular end, there will be only one location marked in the MRI image, and it should be at the center of the annular shape (at least other than the end). If there are two places. In FIG. 6B, two marks M are shown.

  When the doctor or the like has finished marking the position of the tricuspid valve on the MRI image by operating the operation input unit 16, an operation indicating that the mark has been completed (for example, a mark end operation button provided in advance in the operation input unit 16). Etc.). When the operation indicating that the mark is finished is not performed (No in step S.15), the CPU 20 determines that the mark processing of the position of the tricuspid valve is being performed, and the processing Step S. The process proceeds to 14. On the other hand, when the operation indicating that the mark is finished is performed (Yes in step S.15), the CPU 20 determines that the mark processing of the position of the tricuspid valve is finished.

  When the processing for marking the position of the tricuspid valve is completed (Yes in step S.15), the CPU 20 marks the position of the tricuspid valve based on an algorithm called two-dimensional continuous DP (weak spotting version). By using the MRI image at time t as the reference image and the MRI image at time t + 1 as the input image, the correspondence between all the pixels in the reference image and the input image is obtained (step S.16). Here, 2D continuous DP is a well-known technique, “Ryuichi Oka, two others“ About the general scheme of continuous DP—all pixel optimal matching for image spotting ”, IEICE Tech. Detailed descriptions are given in the Information and Communication Society, IEICE Technical Report, PRMU2010-87, IBISML2010-59 (2010-09) and Japanese Patent Application Laid-Open No. 2010-165104.

  And CPU20 calculates | requires the pixel corresponding to all the pixels of the tricuspid valve marked in the MRI image of the time t in the MRI image of the time t + 1 from the correspondence of all the pixels based on two-dimensional continuous DP (step S.17). ).

  Next, the CPU 20 performs a pixel increase process called dilation and a pixel reduction process called erosion on the pixel group of the MRI image at time t + 1 for which the correspondence of the mark is obtained by the two-dimensional continuous DP. (Step S.18). By applying this dilation and erosion to the MRI image at time t + 1, removal of isolated points (isolated pixels) of the pixel group obtained as pixels corresponding to the tricuspid valve at time t + 1, discontinuous pixels And filling the hole, the mark (pixel group at the mark position) of the MRI image at time t + 1 becomes clear.

  In the MRI apparatus 1 according to the present embodiment, as an example, dilation (dilation: pixel increase processing) is performed twice and erosion (erosion: pixel reduction processing) is performed four times. However, the number of times of dilation and erosion is not particularly limited.

  Note that pixel increase / reduction processing using dilation and erosion is a typical image processing called morphological processing, and as described above, removal of isolated points (isolated pixels) is performed. This is a process used for discontinuous pixel connection and hole filling. Since this image processing is a known processing method, detailed description thereof is omitted here.

  FIG. 7 shows the dilation and erosion of the t = 2 MRI image mark obtained based on the two-dimensional continuous DP from the tricuspid valve pixel group marked in the t = 1 MRI image. Is an MRI image showing an example of clarification by repeating a plurality of times.

First, FIG. 7 (a), as previously described, a t = 1 of the MRI image, is already an image in which the mark M ₁ to the position of the tricuspid valve has been made. On the other hand, the MRI image shown in FIG. 7B shows an MRI image at t = 2 in which a corresponding mark is obtained based on the two-dimensional continuous DP. As shown in the enlarged image of FIG. 7 (b), the mark _{M 2} of t = 2 of the MRI image obtained based on the two-dimensional continuous DP as a collection of marks _{M 1} and different, fine point of (a) It is shown.

The CPU 20 performs dilation and erosion a plurality of times on the MRI image obtained as shown in FIG. By executing the dilation and erosion, as shown in FIG. 7 (c), the mark M ₃ is a group of pixels of the lump, it is possible to realize a clarification of the mark M _3.

  Thereafter, the CPU 20 records the MRI image with the clarified mark in the storage unit 23 (step S.19). The MRI image is recorded in association with the value of the imaging position L and the value of the imaging time t.

  Next, the CPU 20 determines whether or not the MRI image at time t + 1 is the MRI image at time T (that is, whether or not t + 1 = T, and whether or not the t + 1 image is the last MRI image at each imaging position). (Step S.20). When the MRI image at time t + 1 is not the MRI image at time T (No in step S.20), the CPU 20 increases the value of t by one by the process of t = t + 1 (step S.21). The process is described in step S. The process described above is repeatedly executed.

  On the other hand, when the MRI image at time t + 1 is the MRI image at time T (Yes in step S.20), the CPU 20 determines whether or not the imaging position L of the MRI image is the sixth imaging position. That is, it is determined whether or not L = 6 (step S.22).

  When the imaging position L of the MRI image is not the sixth position (No in step S.22), the CPU 20 outputs the control instruction to the sequencer 7, thereby setting the imaging position of the MRI image in the MRI apparatus 1 to 12 mm. (Step S.23). Then, the CPU 20 increases the value of the photographing position L by 1 (step S.24), and the process is performed in step S.24. To 12.

  Here, the direction in which the imaging position moves is taken as the Z-axis direction, and the cross section (slice plane) of the heart where the MRI image is taken is taken as the XY plane. MRI images are generally taken with a resolution of 512 pixels × 512 pixels, and this plane physically corresponds to a size of 35 cm × 35 cm. Therefore, the section of 1 pixel corresponds to 0.684 mm.

  When a region in which a heart is reflected is cut out from an MRI image of 512 pixels × 512 pixels, the size is 250 pixels × 250 pixels, and the size of the MRI image used for detecting the position of the tricuspid valve is 250 pixels × It will be included in the size of 250 pixels. Therefore, the actual length of one side of the image of the clipped area is 250 pixels × 0.684 mm = 170.9 mm.

  From this, the three-dimensional space required for detecting the position of the tricuspid valve has a length in the X-axis direction and the Y-axis direction of 170.9 mm, and the length of the Z-axis report is 12 mm × 5 (shooting position L = 1 To the distance L = 6) = 60 mm. Accordingly, the physical space in which the position of the tricuspid valve is detected is similar to a space of 171 mm × 171 mm × 60 mm. The length of 60 mm in the Z-axis direction is 60 mm ÷ 0.684 mm (length of one pixel) = 88 in terms of pixels, and corresponds to 88 pixels.

  When the imaging position L of the MRI image is the sixth position (Yes in step S.22), that is, when the mark processing of the T-th MRI image at the last imaging position (L = 6) is completed. The CPU 20 calculates the barycentric position of the pixel group constituting the mark area from the marks of all photographed MRI images (step S.25). Thus, by obtaining the position of the center of gravity of the mark in each MRI image, the position of the tricuspid valve that moves with the heartbeat can be extracted in time series.

  In addition, as is apparent from the above-described processing, the tricuspid slice image based on the MRI image cannot be simultaneously photographed at six locations. However, the MRI image at each imaging position is acquired at the timing when the heart beat is started, at the time t = 1, and until the next heartbeat starts, the MRI image at time t = T. Is taken. Since this imaging timing is the same timing from the imaging position L = 1 to L = 6, it can be determined that MRI images at the same timing were captured at each imaging position corresponding to the movement of the heartbeat. Therefore, if the MRI image at time t = 1, which is the timing at which the heartbeat starts, is identified as the initial time, the MRI images captured at a plurality of imaging positions can be processed as MRI images captured at the same time. It becomes possible to obtain the three-dimensional movement of the tricuspid valve.

  Since the tricuspid valve has a ring shape, the MRI image (slice image) of the tricuspid valve captures one or two cross-sectional images of the ring-shaped cross section of the tricuspid valve. This ring-shaped cross section varies depending on the photographing position, and when photographing the end portion side of the tricuspid valve, the tricuspid valve has one cross section. Therefore, it is possible to three-dimensionally determine the ring positions of 10 to 12 tricuspid valves (the center of gravity positions of the marks) according to the time t from the MRI images of the six shooting positions with different shooting positions. By changing the ring positions of 10 to 12 tricuspid valves shown three-dimensionally from t = 1 to t = T according to time t, the state of the tricuspid valve obtained three-dimensionally is changed. It becomes possible to obtain in time series, and the movement of the tricuspid valve can be obtained.

  However, the ring position of the tricuspid valve obtained as described above is 10 to 12 pieces of position information at time t, and is a part of the overall position information of the tricuspid valve in three dimensions. There is only. On the other hand, in order to obtain a large amount of position information, it is preferable to perform slice imaging of the tricuspid valve with the MRI apparatus 1 at a number of imaging positions, but it is actually difficult.

  For this reason, the ring position of the tricuspid valve at each obtained time t is set as a three-dimensional sample point of the tricuspid valve, and the ring shape of the tricuspid valve in the three-dimensional space is formed by connecting the sample points. By assuming a smooth three-dimensional curve, the ring shape of the tricuspid valve with reasonableness can be obtained.

  Accordingly, the CPU 20 calculates a spline curve that forms a smooth closed curve through the ring positions (the center of gravity positions of the marks) of 10 to 12 tricuspid valves obtained from the respective MRI images at every time t (step). S.26). Since a method for calculating a spline curve based on some sample points is already known, a detailed description thereof is omitted here. By obtaining the three-dimensional spline curve in this way, it is possible to express the three-dimensional tricuspid ring shape by connecting, for example, 1000 spatial points.

  FIG. 8A is a diagram expressing the ring shape of the tricuspid valve by disposing a small sphere at each point of the spline curve passing through the sample point (white sphere). When the total time from the start of one heart beat (beat) to immediately before the next heart beat is considered as the aforementioned time T, T spline curves can be obtained. Therefore, based on T spline curves from time t = 1 to T, the three-dimensional movement of the tricuspid valve in the time from the start to the end of one beat can be obtained.

  The CPU 20 obtains three-dimensional coordinate information of the tricuspid valve based on T spline curves from time t = 1 to T (step S.27), and obtains the obtained three-time dimension of the tricuspid valve. A three-dimensional moving image of the tricuspid valve corresponding to one heartbeat is generated based on the change in the coordinate information (step S.28).

  The three-dimensional moving image of the tricuspid valve thus obtained is repeatedly displayed (reproduced and displayed) on the display unit 15, thereby changing the state of a part such as a tendon that cannot be clearly imaged by the MRI apparatus 1, this embodiment In this case, it is possible to extract the image of the state change of the tricuspid valve as a video clearly showing the change and provide it to a doctor or the like. For this reason, the doctor visually determines whether or not there is any abnormality in the ring shape and whether or not there is any abnormality in the movement based on the clear three-dimensional image and three-dimensional moving image of the tricuspid valve. It becomes possible to do.

  Further, in the moving image of the tricuspid valve, all position information of the ring of the tricuspid valve and its position change (movement) are recorded as three-dimensional electronic data. For this reason, the three-dimensional space is divided into small dice shapes, and the number of dice-shaped divided portions that come into contact with the movement of the tricuspid valve in one heartbeat is added based on the moving image of the tricuspid valve. The movable volume of the tricuspid valve can be calculated. As described above, the dice-shaped division interval can be obtained in units of millimeters (mm), so that the movable volume in units of cubic millimeters can be obtained.

  FIG. 8B is a three-dimensional view showing the movable volume of the tricuspid valve. By collecting a large number of normal tricuspid valve movable volumes and abnormal tricuspid valve movable volumes based on reliable samples, it is possible to determine the presence or absence of a disease by a statistical test.

  As described above, the moving image generating apparatus and the moving image generating method of the internal organ according to the present invention have been described using the MRI apparatus 1 as an example and imaging the tricuspid valve with the MRI apparatus 1. The in-vivo organ moving image generation apparatus and in-vivo organ moving image generation method according to the present invention are not limited to the configuration and processing method of the MRI apparatus 1 shown in the above-described embodiment, and the tricuspid valve is imaged. It is not limited only to the case.

  For example, in the MRI apparatus 1 according to the embodiment, the case of generating a ring-shaped moving image of the tricuspid valve of the heart has been described. However, by marking other parts on the MRI image at time t = 1, It is possible to generate a moving image of a part corresponding to the position. Therefore, for example, it is possible to generate a moving image in the atrium by setting a mark on a part such as the atrium of the heart, and it is possible to perform visualization of the atrial operation state and calculation of the movable volume in the same manner. is there.

  In particular, in the moving image generating apparatus for internal organs according to the present invention, a two-dimensional continuous DP can be obtained by setting a mark in an MRI image at time t = 1 even if the region is not necessarily clearly shown in the MRI image. Thus, the pixel group for which the correspondence relationship is obtained can be automatically obtained. For this reason, it is possible to automatically obtain the state of the part at the corresponding mark position by reliably performing the mark processing of the part in the MRI image at time t = 1. Therefore, it is possible to visualize the shape (mode) of the part that could not be visually confirmed in the past and the operation state thereof, and it is possible to improve the diagnostic accuracy and reduce the diagnostic burden in clinical diagnosis. Become.

  Further, the MRI apparatus 1 according to the present embodiment obtains a cross-sectional image mask process, a mark identification process based on a two-dimensional continuous DP, an image process using dilation and erosion, and a three-dimensional spline curve. The case where processing etc. are performed only by CPU20 was demonstrated. However, the present invention is not limited to a configuration in which all these processes are performed by only one CPU. A different CPU may be provided for each process, and a plurality of CPUs may sequentially perform processes, and as a result, a three-dimensional moving image of the internal organ may be generated.

DESCRIPTION OF SYMBOLS 1 ... MRI apparatus 2 ... Static magnetic field generating magnet (cross-sectional image imaging means)
3 ... Coil for high frequency transmission (cross-sectional image photographing means)
4 ... Coil for high frequency reception (cross-sectional image photographing means)
5. Gradient magnetic field coil (cross-sectional image photographing means)
6 ... Computer 7 ... Sequencer (cross-sectional image photographing means)
8 ... High-frequency generator (cross-sectional image photographing means)
9: Modulator (section image photographing means)
10: Amplifier (cross-sectional image photographing means)
11: Gradient magnetic field generator (cross-sectional image photographing means)
12 ... Amplifier (cross-sectional image photographing means)
13 ... Phase detector (cross-sectional image photographing means)
14 ... AD converter (cross-sectional image photographing means)
15: Display unit 16: Operation input unit (mark adding means)
20 CPU (shooting control means, image analysis means)
21… ROM
22 ... RAM
23 ... Storage unit (recording means)
X: Subject

Claims (6)

A cross-sectional image photographing means capable of moving the photographing position of the cross-sectional image in the internal organ in the laminating direction of the cross-sectional image, and capable of continuously photographing the cross-sectional image at a predetermined time interval;
Recording means for recording the cross-sectional images photographed by the cross-sectional image photographing means in association with the photographing order of the cross-sectional images continuously photographed at the same photographing position and the photographing position of the cross-sectional image;
Photographing control means for controlling the photographing position and photographing timing in the cross-sectional image photographing means;
Image analysis means for performing image analysis of the cross-sectional image recorded in the recording means;
Mark adding means capable of recording a position of a predetermined site in the internal organ by a mark in the cross-sectional image by a user setting a mark position with respect to the cross-sectional image;
The imaging control means causes the cross-sectional image imaging means to continuously take the cross-sectional images of the internal organs at one imaging position, and the internal organs fit in the continuous imaging of the cross-sectional images. In the stacking direction range, shift the shooting position little by little and let it be done for each shooting position,
The image analysis means includes
Read the cross-sectional image recorded in the recording means,
Add the mark to the mark position set by the mark adding means for the first cross-sectional image taken at the shooting position among the plurality of cross-sectional images taken continuously at one shooting position,
Using a two-dimensional continuous DP, based on the mark position of the cross-sectional image to which the mark has been added, by adding a mark to the cross-sectional image taken after the cross-sectional image to which the mark has been added, Automatically add marks to all cross-sectional images at all shooting positions
The center of gravity of the mark added to the cross-sectional image is calculated for all cross-sectional images based on the pixel group of the mark,
Based on the calculated center-of-gravity position, the coordinate information of the part of the internal organ in the three-dimensional space composed of the image plane of the cross-sectional image and the stacking direction is taken at different imaging positions and the cross-sectional images taken continuously. Find each cross-sectional image taken in the same shooting order among the images,
Obtain a three-dimensional spline curve based on the coordinate information photographed in the same photographing order, obtain a three-dimensional shape of the part for each photographing order,
The time-dependent change of the three-dimensional shape of the part is obtained by changing the three-dimensional shape of the part obtained for each photographing order according to the photographing order, and the three-dimensional shape in the part is obtained. A moving image generating apparatus for organs in a body characterized by generating a moving image. The image analysis means performs image processing by dilation and erosion on the cross-sectional image to which the mark is added before calculating the gravity center position of the mark for all cross-sectional images. The apparatus for generating a moving image of a body organ according to claim 1. When the cross-sectional image photographing means continuously captures the cross-sectional image, the photographing control means is a cross-sectional image from the start of one heartbeat of the subject having the internal organ to the start of the next heartbeat. The apparatus for generating a moving image of a body organ according to claim 1 or 2, wherein the image is continuously photographed. A cross-sectional image photographing means capable of moving the photographing position of the cross-sectional image in the internal organ in the laminating direction of the cross-sectional image, and capable of continuously photographing the cross-sectional image at a predetermined time interval;
Recording means for recording the cross-sectional images photographed by the cross-sectional image photographing means in association with the photographing order of the cross-sectional images continuously photographed at the same photographing position and the photographing position of the cross-sectional image;
Photographing control means for controlling the photographing position and photographing timing in the cross-sectional image photographing means;
Image analysis means for performing image analysis of the cross-sectional image recorded in the recording means;
A moving image generating apparatus for internal organs having mark adding means capable of recording a position of a predetermined site in the internal organs in the cross-sectional image with a mark by a user setting a mark position with respect to the cross-sectional image In
The photographing control means is
The cross-sectional image photographing means continuously photographs the cross-sectional images of the internal organs at one photographing position, and the continuous cross-sectional images are photographed little by little in a stacking direction range in which the internal organs are accommodated. Shift the shooting position for each shooting position,
The image analysis means is
Read the cross-sectional image recorded in the recording means,
Add the mark to the mark position set by the mark adding means for the first cross-sectional image taken at the shooting position among the plurality of cross-sectional images taken continuously at one shooting position,
Using a two-dimensional continuous DP, based on the mark position of the cross-sectional image to which the mark has been added, by adding a mark to the cross-sectional image taken after the cross-sectional image to which the mark has been added, Automatically add marks to all cross-sectional images at all shooting positions
The center of gravity of the mark added to the cross-sectional image is calculated for all cross-sectional images based on the pixel group of the mark,
Based on the calculated center-of-gravity position, the coordinate information of the part of the internal organ in the three-dimensional space composed of the image plane of the cross-sectional image and the stacking direction is taken at different imaging positions and the cross-sectional images taken continuously. Find each cross-sectional image taken in the same shooting order among the images,
Obtain a three-dimensional spline curve based on the coordinate information photographed in the same photographing order, obtain a three-dimensional shape of the part for each photographing order,
The time-dependent change of the three-dimensional shape of the part is obtained by changing the three-dimensional shape of the part obtained for each photographing order according to the photographing order, and the three-dimensional shape in the part is obtained. A method for generating a moving image of a body organ characterized by generating a moving image. The image analysis means is
The image processing by dilation and erosion is performed on the cross-sectional image to which the mark is added before calculating the center-of-gravity position of the mark for all cross-sectional images. A method for generating moving images of internal organs. The photographing control means is
When the cross-sectional image capturing unit continuously captures the cross-sectional image, the cross-sectional image from the start of one heartbeat of the subject having the internal organ to the start of the next heartbeat is continuously captured. The method for generating a moving image of a body organ according to claim 4 or 5, wherein: