JP2006246941A - Image processing apparatus and vessel tracking method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processing device capable of evaluating the cross section of a blood vessel at the same time and in real time as the vessel tracking, and reducing the time required before the evaluation. <P>SOLUTION: The image processing device comprises a three-dimensional image data storing means for storing the three-dimensional image data of a subject, a region setting means for setting a desired region of a tubular structure in the three-dimensional image data stored by the three-dimensional image data storing means from the three-dimensional image data, a core line extracting means for setting the core line of the tubular structure set by the region setting means from the three-dimensional image data, an orthogonal crossing section producing means for producing an orthogonal crossing section image orthogonally crossing the tubular structure based on the core line extracted by the core line extracting means, and a display means for displaying the core line extracted by the core line setting means and the orthogonal crossing section image produced by the orthogonal crossing section producing means at the same time as the image is produced. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an image processing apparatus, and in particular, a blood vessel, a bronchus, an intestinal tract, and the like using three-dimensional volume data of a subject obtained from an imaging apparatus such as an X-ray CT apparatus or a magnetic resonance diagnostic apparatus (MRI apparatus). The present invention relates to an image processing apparatus having a function of performing shape display and quantitative analysis of the lumen of a hollow organ. The present invention also relates to a method for tracking and displaying the travel of a luminal organ executed in such an image processing apparatus.

  In recent medical diagnosis, it is required to display the shape of a luminal organ such as a blood vessel of a subject or to subject the running state to quantitative analysis. When performing this shape display or quantitative analysis, three-dimensional volume data of a part including a luminal organ of a subject is acquired by a medical modality such as an X-ray CT apparatus MRI apparatus, and the three-dimensional volume data (stereoscopic data) is an image. It is attached to processing.

  For example, in order to find a diseased part such as a stenosis of a blood vessel, a number of cross-sectional images are taken along the direction of blood vessel extension, and screen images are sequentially searched. As if the endoscope has advanced in the digestive organs, etc. The cross-sectional images must be sequentially displayed as the images go.

  In this image processing, it is necessary to know position information of a three-dimensional core line (for example, a center line) of a hollow organ. As a method of three-dimensionally extracting the core line of a hollow organ using three-dimensional volume data, a region of a desired tubular structure in stereoscopic image data is set from the stereoscopic image data, and the set tubular structure There has been proposed one in which a core line is set from stereoscopic image data (see, for example, Patent Document 1).

More specifically, this method is specified by specifying one start point and one or a plurality of end points specified by the operator as the extraction target range of the luminal organ (for example, blood vessel) on the stereoscopic image data. In this range, a desired luminal organ, for example, a blood vessel region, is extracted three-dimensionally by clustering or the like, and a point group of the core line of the luminal organ in this region is searched using a method such as Euclidean distance conversion. Then, after connecting these point groups smoothly, the core line from the start point to the end point is obtained by repeating the process of determining the vector at the tip and calculating the next vector.
Japanese Patent Laid-Open No. 2004-230086

  However, the above-described method has the following problems. First, since the blood vessel running short axis image is displayed after repeating the above processing between the set start point and end point, the processing takes time, and the result cannot be seen unless the processing ends. The user feels that waiting time is long. That is, it is a problem that it cannot be evaluated in real time. Further, from this, even if there is an error in the detected part, it cannot be corrected until the entire process is completed.

  In this method, in order to determine a vector in the tube direction, a representative point (CT value) in the extracted region is found as a sampling target. Then, if there is a large change in the concentration of the blood vessel lumen due to adhesion of plaque between the start point and the end point, or if there is a large change in the concentration of the contrast agent, or if there is a bend or the blood vessel suddenly narrows There is also a problem that the region extraction is likely to fail.

  The present invention has been made in consideration of the above-described circumstances, and an object of the present invention is to provide an image processing apparatus that can simultaneously and in real time evaluate a blood vessel cross-section with tracking of blood vessel travel, and reduce the time required for the evaluation. To do.

  Another object of the present invention is to improve the user operability so that the blood vessel cross-section and the blood vessel cross-section can be simultaneously evaluated in parallel with the blood vessel cross-sectional detection result in real time. Another object of the present invention is to provide an image processing apparatus.

  In order to solve the above-described problem, an image processing apparatus according to the present invention includes a stereoscopic image data storage unit that stores stereoscopic image data of a subject, and the stereoscopic image data storage unit as described in claim 1. Area setting means for setting a region of a desired tubular structure in the stored stereoscopic image data from the stereoscopic image data, and a core line for setting the core of the tubular structure set by the area setting means from the stereoscopic image data Created by an extraction means, an orthogonal cross-section creation means for creating an orthogonal cross-sectional image orthogonal to the tubular structure based on the core wire extracted by the core wire extraction means, and the core wire extracted by the core wire setting means and the orthogonal cross-section creation means Display means for displaying the orthogonal cross-sectional image.

  Preferably, as described in claim 2, the core wire extracting means sequentially obtains the next core wire position from the core wire positions on the orthogonal cross section created by the orthogonal cross section creating means and connects both core wire positions smoothly. The core line is extracted and the display means can be configured to display the core line gradually as the core line is extracted.

  Next, the orthogonal cross-section creating means preferably obtains the next orthogonal cross-section sequentially from the core line positions in the last two orthogonal cross-sections created as described in claim 3, and the display means It can be set as the structure which displays the orthogonal cross section located in a core wire front-end | tip as a core wire is newly extracted.

  Preferably, the display means preferably includes one or two or more Curved MPR images in which the core lines extracted by the core line extraction means are superimposed and displayed, and the orthogonal cross section creation means. It is desirable to display the created desired cross-sectional image.

  More preferably, the display means, as described in claim 5, includes the Curved MPR image, a Curved Projection created with a specified width, a MIP image, a sagittal image, a coronal image, an axial image, and a rotate. It may be configured to be switchable to any of the MPR images.

  Furthermore, the image processing apparatus preferably includes input means for instructing change of the position of the cross section, as described in claim 6, and the position of the cross section displayed on the display means is the input means. It can be configured to be sequentially changed based on the instruction to

  In order to solve the above-described problem, in the tube travel tracking method according to claim 7, a region of a desired tubular structure in the stereoscopic image data of the subject stored by the stereoscopic image data storage unit is defined as the stereoscopic image. A first step of setting from the data; a second step of setting the core line of the tubular structure set by the region setting step from the stereoscopic image data; and a tubular structure based on the core line extracted by the core line extraction step A third step of creating an orthogonal cross-sectional image orthogonal to the body, and a fourth step of displaying the orthogonal cross-section image created by the core wire extracted by the core wire setting step and the orthogonal cross-section creation step, The first step to the fourth step are repeated to gradually advance the core wire and to update the orthogonal cross section located at the core wire tip.

  On the other hand, in order to solve the above-described problem, an image processing apparatus according to an eighth aspect of the present invention is an image processing apparatus that displays an image of a blood vessel cross-section while changing the position along the traveling direction of the blood vessel. As means for obtaining the traveling direction, stereoscopic image data storage means for storing stereoscopic image data of the subject, target area determination means for determining a target area for clustering processing, and clustering for extracting blood vessels from the image data in the target area Processing means; means for determining a search direction based on blood vessel information obtained based on a result of the clustering process; and means for changing the position of the target region based on the search direction. .

  In order to solve the above-described problem, a tube travel tracking method according to claim 9 is a tube travel tracking method for obtaining a travel direction of a blood vessel. A second step of extracting blood vessels from the image data of the subject in the region using clustering processing; a third step of determining a search direction based on the blood vessel information obtained in the second step; And a fourth step of changing the position of the target area based on the search direction.

  According to the image processing apparatus of the present invention, blood vessel cross-sections can be simultaneously evaluated in real time together with tracking of blood vessel running, and the time required for evaluation can be shortened.

  In addition, the present invention can visually evaluate the blood vessel in parallel with the blood vessel cross-section as well as the blood vessel cross-section detection result performed in real time in this way, thereby improving the operability for the user. can get.

  Hereinafter, preferred embodiments of an image processing apparatus according to the present invention will be described in detail with reference to the drawings. As shown in FIG. 1, the image processing apparatus according to this embodiment is a computer apparatus integrated with a medical modality, a computer apparatus connected online to a medical modality via a communication line, or a medical modality. Is provided as an off-line computer device.

  As shown in FIG. 1, the computer device includes a calculation device 11 equipped with a CPU and a memory, a storage device 12 that stores programs and processing data, a display device 13, and an input device 14. As described above, this computer apparatus also has a function for performing data communication with the outside as necessary.

  Three-dimensional (stereoscopic) image data of a subject collected by an imaging apparatus (medical modality) such as an X-ray CT apparatus or an MRI apparatus is sent to the storage device 12 online or offline. The three-dimensional image data is stored in a large-capacity recording medium such as a magneto-optical disk provided in the storage device.

  In addition, a program for executing the pipe traveling tracking method according to the present embodiment is recorded in advance on a recording medium such as a hard disk of the storage device 12. Thereby, the arithmetic unit 11 reads this program at the time of starting, and sequentially executes processing for pipe travel tracking according to the procedure described in the program. In the middle of this execution, an image related to tube travel tracking is displayed on the display device 13 and operation information related to tube travel tracking processing from the operator is received via the input device 14. For this reason, the arithmetic unit 11, the display device 13, and the input device 14 can also function as an interface for executing pipe travel tracking processing for the operator.

  By executing the above-described pipe travel tracking program with such a hardware configuration, functionally, as shown in FIG. 2, a stereoscopic image data storage unit 21, a start point designating unit 22, a region extracting unit 23, The core line extracting unit 24 (including the core line searching unit 25 and the core line smoothing unit 26), the contour extracting unit 27 (including the contour smoothing unit 30), the orthogonal cross section creating unit 29, and the composite display unit 30 are realized. Hereinafter, the processing contents of the functional units 21 to 30 will be described individually.

<Stereoscopic image data storage unit 21>
As described above, the stereoscopic image data storage unit 21 realized with the storage device 12 as a main part stores three-dimensional image data obtained from an imaging apparatus such as an X-ray CT apparatus or an MRI apparatus.

<Start point designating unit 22>
The start point designating unit 22 is functionally configured with the arithmetic device 11, the display device 13, and the input device 14 as main elements. In the start point designating unit 22, the coordinates of one start point designated by the operator as the extraction target range of the luminal organ (for example, blood vessel) on the stereoscopic image data are input and stored.

<Region Extractor 23>
The region extraction unit 23 is realized by the calculation function of the calculation device 11. In the following functional units 24 to 30, the description is omitted for those realized by the arithmetic function of the arithmetic unit 11 like the region extracting unit 23.

  The region extraction unit 23 automatically extracts a desired luminal organ, for example, a blood vessel region, three-dimensionally using the three-dimensional image data sent from the stereoscopic image data storage unit 21 as input data.

  First, the region extraction unit 23 extracts three-dimensional image data of a target region to be clustered. This target area is a range based on the start point position and search direction specified by the start point specifying unit 22. For example, a spherical region having a radius of about several mm from the starting point position is set as a processing target region.

  Next, the CT value data in the target region is classified for each similar substance by clustering processing. By this clustering processing, the organs are classified into the inside and outside of the luminal organ, and are stored in the storage device 12 as binarized three-dimensional image data.

  Next, the clustering method will be described in detail. Under ideal conditions, lipids are known to exhibit CT values of -100 to 50 and vessel walls (muscles) exhibit CT values of 50 to 129. However, in actual examinations, the CT value is affected by the body size of the subject (patient), beam hardening, reconstruction function, substance size, substance state outside the region of interest, contrast agent concentration, etc. It has fluctuated. For this reason, the method of segmentation using a fixed threshold cannot cope with such a variation in CT value and cannot be classified well. In the present embodiment, the threshold value for dividing the region is automatically determined by a clustering method based on the histogram of the image.

  In the following description, the case of dividing into three regions according to the CT value will be described, but the number of regions may be other numbers.

  FIG. 3 is a configuration diagram of the region extraction unit 23. The region extraction unit 23 includes a histogram generation unit 23a, a region determination unit 23b, and a CT value conversion unit 23c.

  The histogram generation unit 23a obtains a histogram of the input low contrast image. The histogram represents the frequency distribution for each CT value.

  The region determination unit 23b classifies the histogram into three regions using the K-mean method. A threshold value for determining each region is obtained by the K-mean method so that the variance values of the CT values in each region are equal. By determining each region so that the variance values of the CT values in each region are equal, highly similar substances are likely to be included in the same region.

  FIG. 4 shows an example of area division according to the present embodiment, and CT values from −100 to 537 are classified into three areas. Pixels having CT values in the range of region A are mainly lipid portions. Pixels having CT values in the range of region B are mainly blood vessel wall portions. Pixels having CT values in the range of region C are mainly blood flow portions.

  The CT value conversion unit 23c converts the CT value of the input image into a value corresponding to each region based on the threshold value obtained by the region determination unit 23b. Thereby, it is possible to obtain three-dimensional image data classified into the area of the luminal organ and the area outside the area.

  This region classification can be performed by the PCNN (Pulse Coupled Neural Networks) method in addition to the K-mean method described above. However, in the pipe tracking method according to the present invention, since the range for performing region classification is short, there is no large difference in calculation accuracy between the K-mean method and the PCNN method. It is preferable to adopt.

<Core extraction unit 24>
The core line extraction unit 24 functionally includes a core line search unit 25 and a core line smoothing unit 26, and processes assigned in this order are executed.

  The core line search unit 25 performs a search process, which will be described later, on the three-dimensional volume data of the blood vessel region as the region of the luminal organ extracted by the region extraction unit 23 using the start point specified by the start point specifying unit 22 as a starting point. And the core line of the luminal organ is extracted.

<Core search unit 25>
The core line search unit 25 searches the point group of the core line of the luminal organ sequentially from the starting point. Specifically, when the direction to be searched first (initial search direction vector) is determined, the inside of the blood vessel extraction region is searched by an iterative search process.

  That is, as shown in FIG. 5, among the unit vectors radiated from the start point in the direction of all three-dimensional angles (4π), the vector having the maximum length when extended to the boundary position of the blood vessel extraction region The vector 2 in the opposite direction is defined as an initial search vector 1 and an initial search vector 2, respectively. At this time, the cross section in the extraction region orthogonal to the initial search vector 1 and the initial search vector 2 is the cross section S orthogonal to the search vector.

  Next, the search process is repeatedly performed by the core line search unit 25, and the position where the Euclidean distance in the blood vessel extraction region A in the cross section orthogonal to the search vector is maximum is obtained as the blood vessel core line position O. Specifically, as shown in FIG. 6, a 1-pixel region connected to a pixel at the vector tip position is extracted from the cross-sectional image of the region extraction data orthogonal to the search vector and including the vector tip position. The Euclidean distance conversion is performed after a hole (0-pixel area) existing in the extraction area is subjected to the hole filling process.

  This Euclidean distance conversion is a conversion in which the distance (Euclidean distance) between a pixel in the extraction region and a pixel outside the extraction region that is closest to the pixel is the value of the pixel. The value is the maximum value in the extraction area. The position of the pixel having the maximum value is obtained as the position of the core line. When the pixel at the vector tip position is 0-pixel, an area including 1-pixel at the position closest to the vector tip position is used.

  As another method for obtaining the core line position, there is a technique for obtaining the position of the center of gravity of the blood vessel extraction region A in the cross section S orthogonal to the search vector as the position of the blood vessel core line. That is, a 1-pixel (pixel having a value of 1) region connected to the pixel at the vector tip position is extracted on the cross-sectional image of the region extraction data that is orthogonal to the search vector and includes the vector tip position. Is obtained as the core line position. If the pixel at the vector tip position is 0-pixel, the barycentric position of the region including 1-pixel at the position closest to the vector tip position is obtained.

  Moreover, core line search methods other than these can also be implemented. As an example, the region extraction data extracted by the region extraction unit 23 is subjected to a cavity filling process for a cavity (0-pixel region) existing in the extraction region, and then three-dimensional Euclidean distance conversion is performed. Image data for dimensional Euclidean distance conversion may be prepared in advance. In other words, the pixel position having the maximum value may be obtained as the core line position on the cross-sectional image based on the image data that is orthogonal to the search vector and includes the vector tip position and subjected to the three-dimensional Euclidean distance transformation.

When the search for the core line ends in this way, the core line search unit 25 obtains the next vector (next orthogonal cross section). As shown in FIG. 7, this is obtained by executing the following arithmetic expression using a vector in which the next vector bnew connects the current search vector b and the last two core points.

  When the next orthogonal cross section is obtained in this way, the position of the core line in this orthogonal cross section is subsequently obtained by the above-described three-dimensional Euclidean distance conversion. FIG. 8 schematically shows how the core lines are successively determined by the three-dimensional Euclidean distance conversion. In this way, the calculation of the position of the blood vessel core line and the calculation of the next vector are repeatedly executed.

  The initial search direction of the core line, for example, when the luminal organ is a blood vessel, the operator specifies on the display device 13 via the input device 14 whether the search starts from the start point to the downstream side or the upstream side. Is determined by

<Core wire smoothing unit 26>
The core line smoothing unit 26 has a function of smoothing the three-dimensional core line extracted by the core line searching unit 25 to make a smooth curve. That is, an appropriate smoothing process is performed on the extracted three-dimensional core wire to shape it into a smooth curve.

<Outline extraction unit 27>
The contour extracting unit 27 also includes a contour smoothing unit 30. The contour extraction unit 27 is configured to obtain information on luminance values on a straight line extending in the radial direction from the position of the core line on the cross section S orthogonal to the core line extracted by the core line extraction unit 24 or the region extracted by the region extraction unit 23. Based on the boundary, the position of the blood vessel contour is detected.

  As shown in FIG. 9, the contour extraction is performed by setting straight lines l1 to ln at equiangular intervals around the core line position O on the cross section orthogonal to the core line obtained by the core line extraction unit 24. From the core line position O, the blood vessel contour position in the negative direction and the positive direction is detected.

  Specifically, it is checked whether or not the following three conditions are satisfied from the core line position c along the linear radiation direction. The three conditions are (1) whether or not there is a position where the luminance value (CT value) falls outside a certain threshold range, and (2) whether or not there is a position where the absolute value of the gradient of the luminance value exceeds a certain threshold value. And (3) whether the boundary position of the extraction area has been reached. Of these, the position corresponding to the first established condition is detected as the blood vessel contour position. As a result, it is possible to almost eliminate the state of erroneously detecting a position that is clearly off as a blood vessel contour.

<Outline smoothing unit 28>
The contour extracted by the contour extracting unit 27 is smoothed by the contour smoothing unit 30 and shaped into a three-dimensional blood vessel contour having a smooth curve. That is, as with the smoothing of the core wire, a smoothing process is performed on a series of obtained contour points to form a smooth curve.

<Orthogonal cross section generator 29>
The orthogonal cross section creation unit 31 has a function of determining a vector at the tip of pipe travel and creating a cross section orthogonal to the determined vector. The cross section orthogonal to the search vector was also calculated by the core line extraction unit 24 and the contour extraction unit 27. However, a cross section displayed as an image as a cross section orthogonal to the core line is created only after the vector at the tip of the pipe travel is determined. .

<Composite display unit 30>
The composite display unit 30 is realized by the display 13 and the calculation function of the calculation device 11. The composite display unit 30 includes the three-dimensional image data stored in the stereoscopic image data storage unit 21 or the data of the core line extracted by the core line extraction unit 24 and the contour data extracted by the contour extraction unit 27. The image data calculated using the three-dimensional image data is synthesized and displayed.

  The pipe traveling tracking method according to the present embodiment is executed by the functional units 21 to 30 according to the flowchart shown in FIG.

  At the time of setting the core line, the MPR image screen shown in FIG. When the operator moves the slice cursor 41 displayed on the top, bottom, left, and right of the MPR image written by the operator via the input unit 14, the corresponding MPR screen is switched, and when one of the arrow buttons on the MPR is continuously pressed, the slice is displayed. Switch. In this way, the operator finds the tube whose travel is to be tracked.

  When the operator clicks a point on the display device 13 to start tracking via the input device 14, the arithmetic unit 11 accepts designation of a desired start point (step S1).

  Then, a red line segment 43 (indicated by a solid line in the figure) is displayed downstream of the designated start point P, and a green line segment 44 (indicated by the broken line) is displayed on the upstream side, and a core initial vector ( The line / probe selection screen 42 pops up and requests to select one. Here, when the operator clicks the first path (downstream or upstream) via the input device 14, the search direction is designated (step S2).

  Then, when the operator presses the “forward” button via the input device 14 on the input screen shown in FIG. 11, the arithmetic unit 11 determines the target area for the clustering process, and the three-dimensional area in the target area. Clustering processing based on the K-means method or the like is performed on the image data, and a short blood vessel region within a certain range from the start point is extracted three-dimensionally (step S3).

  Next, the arithmetic unit 11 performs a search process on the three-dimensional extraction region of the blood vessel to extract a core line having a three-dimensional behavior (step S4), and performs a smoothing process on the core line (step S5).

  Furthermore, the computing device 11 detects the contour of the blood vessel based on the processing along the cross section at each point orthogonal to the extracted core line, and smoothes the detected contour (step S6). Here, a cross section orthogonal to the core line is determined, and a cross section orthogonal to the tip vector is created (step S7).

  The smooth and accurately obtained position information of the core line and the contour is sent from the arithmetic unit 11 to the display unit 13 and output on the screen of the display unit 13 (step S8). Thereby, the core line from the starting point 46 to the first cross section is displayed.

  If the “forward” button 45 is kept pressed, the process is continued (step S9: No), and then there is no blood vessel branch in the vicinity to obtain the next vector (next orthogonal cross section) (step S10: No) The next candidate point is selected (step S11), and the arithmetic unit 11 performs the above-described series of processing for the next candidate point (steps S2 to S7). By repeating in this way, an image in which the cross section is moved in real time along the search direction is displayed.

  FIG. 13 is a diagram showing how the core line is extracted. In the axial image at the lower left of the screen and the coronal image at the upper left, the core line 47 displayed in red or the like from the start point 46 gradually advances every time a cross section is created, and the tip, that is, the current core point 48 is identified in blue. . At the same time, a blood vessel vertical section short axis image (hereinafter referred to as “SA image”) at the current core point 48 is displayed in place of the sagittal image in the upper right of the screen.

  As a result, the operator can observe the running tracking in real time, and can view the blood vessel vertical section at the tip of the core wire simultaneously with the progressing core wire. Evaluation is possible.

  If there is a branch or the like in the course of tracking the pipe travel (step S10: Yes), the arithmetic unit 11 follows the thicker path if there is no instruction from the operator (step S12: No). The process is continued and the next candidate point is selected (step S11).

  At this time, if the operator presses the “return” button 49 on the core line setting screen shown in FIG. 11 in order to perform travel tracking on the narrower route, the arithmetic unit 11 reverses the route along which the core wire has been extended. (Step S12: Yes). Then, when the core line returns to the point before the branch point, stop pressing the “return” button 49 and manually set the next candidate point in the vicinity of the branch point in the narrower path, thereby calculating the arithmetic unit 11. Accepts designation of the next candidate point (step S13: Yes).

  As described above, the pipe travel tracking is automatically performed unless otherwise specified, but this can be corrected when a route different from the operator's desire is followed.

  Finally, when the core wire 46 is extended to a desired length and the operator stops pressing the “forward” button 45, the arithmetic unit 11 finishes the operation (step S9: Yes).

  By performing the above processing, the arithmetic unit 11 can obtain a cross section perpendicular to the extraction of the pipe travel and the tip vector. The obtained data is stored in the storage device 12, and the operator reads out this data from the storage device 12 and displays it in various display forms to check whether the core wire passes through the center of the pipe. Can do. Below, some confirmation forms are demonstrated.

  FIG. 14 shows an example in which a curved MPR image (upper left), an SA image (upper right), and an axial image (lower left) are displayed. A curved MPR image is an image obtained by extending a grayscale image obtained by cutting a three-dimensional image along a curved surface represented as a set of straight lines perpendicular to the plane of the MPR image through each point on the core line of the MPR image to a two-dimensional plane. is there. In the MPR image, if the blood vessel is bent three-dimensionally, the entire range of the blood vessel is not always displayed in the image. However, since the blood vessel core line is a line passing through the inside of the blood vessel, the core line is not shown in the curved MPR image. Blood vessels are drawn over the entire defined range. Therefore, by observing the curved MPR image, information such as the thickness can be obtained over the entire range of the blood vessel.

  When the operator clicks the position of the cross section for which the passing state is to be confirmed on the core line 47 (for example, displayed in red) on the curved MPR image, the SA image 50a at the clicked position is displayed. The position is displayed in blue on the curved MPR image. A cross cursor on the SA image 50a indicates a passing point of the core line. In addition to the SA image 50a at the clicked position, the SA image 50b of the immediately preceding section and the SA image 50c of the immediately following section are also displayed.

  Note that the SA images are not limited to these three images. For example, as shown in FIG. 15, nine consecutive SA images can be displayed at a time. In this case, the displayed top and bottom slice positions are displayed on the curved MPR image.

  Thereby, the operator can confirm whether the core wire has passed the pipe center in arbitrary cross sections. When the passing point needs to be corrected, the circle ROI is moved on the SA image so as to surround the outline of the blood vessel, and correction is performed by setting the cross cursor to be the passing point of the core line.

  FIG. 16 shows an example in which this pipe traveling confirmation is performed on a screen combining a coronal image (upper left), a reference image (upper right), and an axial image (lower left).

  On the reference image, the lumen of the blood vessel is displayed as a wire frame. When the reference image is dragged using a mouse or the like, the coronal image and the axial image are switched in the dragged direction, and the blood vessel core line is projected.

  Thereby, it becomes possible to observe from arbitrary directions. When it is necessary to correct the passing point, the displayed core is directly edited by dragging it to a desired position using the mouse or the like.

  When the blood vessel core line is thus determined, the blood vessel property is subsequently evaluated. FIG. 17 shows an example of a display screen when this vascular property evaluation is performed. The upper left side of the screen is a curved MPR image in the horizontal long axis direction, and the right side is a curved MPR image in the vertical long axis direction. In the lower part of the screen, the SA image, the evaluation result, and the reference image are displayed from the left.

  In each curved MPR image, the blood vessel core line is displayed as a solid line such as red, and a marker is attached to the core line at regular intervals with the start point as a base point. The position of the displayed SA image is displayed as a point such as blue.

  To switch the slice position of the SA image, clicking the position where the SA image is to be displayed on the core line of the curved MPR image, the SA image at that position is displayed. In addition, the slice of the SA image is switched even if the slide bar under the SA image is moved left and right. Alternatively, on the curved MPR image, the lower left SA image can be switched by dragging the mouse along the running of the core line and corresponding to the movement. On the other hand, the slice of the curved MPR image is switched by moving the slide bar beside the curved MPR image up and down.

  Further, as shown in FIG. 18, an SA image and a profile in a straight line displayed in red or the like on the SA image may be displayed on the lower stage of the screen.

  The curved MPR image with the vertical long axis on the right side can be switched to another display mode with a menu button at the upper right of the image. They are the 3rd curved MPR image in the long axis direction, the Curved Projection that projects the curve in space in the line of sight, the MIP image created with the specified width, the sagittal image, the coronal image, the axial image, and the arbitrary cross section MPR image or the like.

  Of these, an example in which an arbitrary cross-sectional MPR image is displayed is shown in FIG. On this screen, the MPRR plane is set by dragging and rotating the normal displayed in green or the like on the lower right reference image (SA). The rotation center of the SA image is the center line, that is, the center of the ROI. If you want to change the center, select the corner ROI, draw a circle RPI on the lower left SA image, and move the ROI center to the target position. To do.

  As described above, if the tube running tracking method according to the present embodiment is used, the core line of a blood vessel that exists three-dimensionally in the subject is automatically extracted, and at the same time, the contour of the blood vessel is determined based on the core line. Automatically extracted.

  In the method of the present embodiment, the position of the display cross section is changed along the traveling direction of the blood vessel, and the clustering process is performed by sequentially changing the position of the target region on which the clustering process is performed. For example, when clustering processing is performed on a wide range of image data, extraction of blood vessel regions may not be performed satisfactorily due to variations in image density values, but with the method of this embodiment, clustering is performed for each relatively small target region. Since the processing is performed, the blood vessel portion can be extracted satisfactorily. Further, this can increase the extraction accuracy of the core wire.

  Thereby, the core line of a luminal organ such as a blood vessel and three-dimensional position information of an orthogonal cross section can be extracted with high accuracy in a short processing time, and information useful for diagnosis can be reliably provided.

  The embodiments described above are for illustrative purposes and do not limit the scope of the invention. Accordingly, those skilled in the art can employ embodiments in which each or all of these elements are replaced by equivalents thereof, and these embodiments are also included in the scope of the present invention.

1 is a block diagram of a computer apparatus according to an example of an image processing apparatus according to the present invention. FIG. 3 is a functional block diagram of an image processing apparatus when performing a three-dimensional automatic blood vessel extraction method according to the present invention. The block diagram of the outline extraction part which concerns on this embodiment. The figure explaining a clustering process. The figure explaining the process for determining the initial search vector in the pipe running tracking method which concerns on this invention. The figure explaining the method of calculating | requiring a core wire by Euclidean distance conversion on the cross section orthogonal to a search vector. The figure explaining the calculation method of the next vector (next orthogonal cross section) required for the iterative process in the pipe running tracking method which concerns on this invention. The schematic diagram which shows a mode that a core line is determined one by one by three-dimensional Euclidean distance conversion. The figure explaining the detection of the outline position on the cross section orthogonal to a core wire. The flowchart which shows the process sequence of the pipe running tracking method which concerns on this invention. The figure which shows an example of a core wire setting screen. The figure which shows an example of a search direction setting screen. The figure which shows the mode of core wire extraction. The figure which shows an example of the screen which confirms pipe | tube running. The figure which shows the modification of the screen which confirms pipe | tube running. The figure which shows the other example of the screen which confirms pipe | tube running. The figure which shows an example of the screen which performs evaluation of vascular property. The figure which shows the modification of the screen which performs evaluation of vascular property. The figure which shows the other example of the screen which performs evaluation of vascular property.

Explanation of symbols

DESCRIPTION OF SYMBOLS 11 Computation device 12 Storage device 13 Display device 14 Input device 21 Stereo image data storage unit 22 Start point designation unit 23 Region extraction unit 23a Histogram generation unit 23b Region determination unit 23c CT value conversion unit 24 Core line extraction unit 25 Core line search unit 26 Core line smoothing Section 27 Outline extraction section 28 Outline smoothing section 29 Orthogonal section creation section 30 Composite display section 41 Slice cursor 42 Core line initial vector (line segment / probe) selection screen 43, 44 Line segment 45 "Forward" button 46 Start point 47 Core line 48 Current Core 49 “Back” button 50 SA image A Blood vessel extraction region B Extraction region boundary C Blood vessel contour S Cross section orthogonal to search vector

Claims (9)

Stereoscopic image data storage means for storing stereoscopic image data of the subject;
Region setting means for setting a region of a desired tubular structure in the stereoscopic image data stored by the stereoscopic image data storage means from the stereoscopic image data;
A core line extracting means for setting the core line of the tubular structure set by the region setting means from the stereoscopic image data;
Orthogonal cross section creating means for creating an orthogonal cross section image orthogonal to the tubular structure based on the core line extracted by the core line extracting means;
Display means for displaying the core wire extracted by the core wire setting means and the orthogonal cross-sectional image created by the orthogonal cross-section creating means;
An image processing apparatus comprising: The core wire extracting means sequentially obtains the next core wire position from the core wire position on the orthogonal cross section created by the orthogonal cross section creating means, extracts the core wire by connecting and smoothing both core wire positions, and the display means extracts the core wire. The image processing apparatus according to claim 1, wherein the core line is displayed gradually as it is moved. The orthogonal cross section creating means sequentially obtains the next orthogonal cross section from the core line position in the last two orthogonal cross sections created,
The image processing apparatus according to claim 1, wherein the display unit displays an orthogonal cross section positioned at the tip of the core wire as the core wire is newly extracted at the same time as the creation. The display means displays one or more Curved MPR images in which the core lines extracted by the core line extraction means are superimposed and a desired orthogonal cross-sectional image created by the orthogonal cross-section creating means. The image processing apparatus according to claim 1. The display means is configured to be able to switch the Curved MPR image to any one of a Curved Projection, a MIP image, a sagittal image, a coronal image, an axial image, and a rotate MPR image created with a specified width. The image processing apparatus according to claim 4, wherein: Comprising input means for instructing a change in the position of the cross section;
The image processing apparatus according to claim 1, wherein the position of the cross section displayed on the display unit is sequentially changed based on an instruction to the input unit. A first step of setting a region of a desired tubular structure in the stereoscopic image data of the subject stored by the stereoscopic image data storage means from the stereoscopic image data;
A second step of setting the core line of the tubular structure set by the region setting step from the stereoscopic image data;
A third step of creating an orthogonal cross-sectional image orthogonal to the tubular structure based on the core line extracted by the core line extraction step;
A fourth step of displaying the core wire extracted in the core wire setting step and the orthogonal cross-section image created in the orthogonal cross-section creation step;
And a step of repeating the first to fourth steps to gradually advance the core wire and update the orthogonal cross section located at the tip of the core wire. An image processing apparatus that displays an image of a blood vessel cross-section while changing the position along the traveling direction of the blood vessel,
As a means for obtaining the traveling direction of the blood vessel,
Stereoscopic image data storage means for storing stereoscopic image data of the subject;
A target area determining means for determining a target area for clustering processing;
Clustering processing means for extracting blood vessels from the image data in the target region;
Means for determining a search direction based on blood vessel information obtained based on the result of the clustering process;
An image processing apparatus comprising: means for changing a position of the target area based on the search direction. In the tube travel tracking method for determining the travel direction of the blood vessel,
A first step of determining a target area for clustering processing;
A second step of extracting blood vessels from the image data of the subject in the target region using a clustering process;
A third step of determining a search direction based on the blood vessel information obtained in the second step;
And a fourth step of changing the position of the target area based on the search direction.

Images