JP2002092590A - Image display device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To display a cross-sectional image of a tubular observation object such as ramified blood vessels, the trachea and the intricately curved intestines by vertically cutting the observation object with a three-dimensional curved face extended along the observation object. SOLUTION: Intersections between a visual point path passing the approximate center of the observation object 12 included in a three-dimensional original image, and cross section images 1, 2, 3 cutting the observation object 12 at specified spaces are obtained (figure 11(a)). Cutting straight lines or cutting curves on the cross section images 1, 2, 3 passing the intersections are obtained (figure 11(b)). In the case the number of intersections is one as in the cross section image 1, a straight line is obtained using the information of the cutting straight line or cutting curve across the intersection, and a cross-sectional image is formed by adding the sequence of points on the cutting straight line or cutting curve obtained every cross section image.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image display device, and more particularly to an image display device for displaying a cross-sectional image of a tubular observation object such as a blood vessel, an intestine, or a trachea.

[0002]

2. Description of the Related Art Conventionally, there has been proposed an image display apparatus in which an observation object such as a blood vessel and an intestine is vertically cut along a curved surface (cut curved surface) along the observation object so that the inside of the observation object can be observed. (Japanese Patent Laid-Open No. 11-318884).

[0003]

However, Japanese Patent Application Laid-Open No.
The image display device described in JP-318188B sets a plurality of viewpoints along the observation target inside the observation target as a cut curved surface along the observation target, and defines the viewpoint as a plane passing through the plurality of viewpoints. I have. However, in the case where a blood vessel is branched into a plurality of blood vessels, a cut curved surface for vertically cutting each blood vessel is a three-dimensional curved surface.
Japanese Patent No. 18884 does not disclose specific means for forming a cross-sectional image cut by a three-dimensional cut curved surface.

The present invention has been made in view of such circumstances, and vertically cuts a tubular observation object such as a branched blood vessel, a trachea, or a complicated meandering intestine on a three-dimensional curved surface along the observation object. It is an object of the present invention to provide an image display device capable of displaying a cross-sectional image that has been reduced.

[0005]

According to a first aspect of the present invention, there is provided an image display apparatus, comprising: a path passing through a substantially center of an observation target included in a three-dimensional original image; Means for finding an intersection with each cutting plane to be cut at intervals, and means for finding a straight line or a curve on the cutting plane passing through the intersection, wherein when the intersection is one point, Means for obtaining a straight line using information of a straight line or a curve, and stacking the straight line or the curve obtained for each cut surface to form a three-dimensional cut surface, including the observation target cut by the cut surface It is characterized by comprising image forming means for forming a cross-sectional image of an original image, and display means for displaying the cross-sectional image formed by the image forming means.

That is, a path that passes through the approximate center of the observation target contained in the three-dimensional original image is a center projection method (Japanese Patent Laid-Open No. 7-21070) in which the viewpoint can be set inside the observation target.
No. 4, JP-A-8-16813), the path of the viewpoint used, or the observation target is extracted by an area expansion method or the like and then binarized, and the binarized value can be obtained from a thin line that has been subjected to thinning processing. . Then, an intersection point between the path and each of the cutting planes for cutting the observation target at a predetermined interval is obtained. still,
Each cutting plane is a plane having an arbitrary inclination that can traverse the path.

[0007] A straight line or a curve on the cutting plane passing through the intersection determined as described above is determined. When one or two intersections exist on one cut plane, a line passing through the intersection is a straight line, and when three or more intersections exist on one cut plane, the intersection is made. Are curves obtained by interpolation (interpolation) and extrapolation (extrapolation) calculation using an approximate polynomial.
If the number of intersections is one, a straight line is determined using information on a straight line or a curve on the cross section before and after the intersection. For example, a straight line or a straight line in the same direction as a straight line or a curved line on the front and rear cut surfaces is used. Then, the straight lines or curves obtained for each of the cut surfaces obtained as described above are stacked to form a three-dimensional cut curved surface, and a cross-sectional image of the original image including the observation target cut by the cut curved surface Is displayed.

[0008]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of an image display device according to the present invention will be described below in detail with reference to the accompanying drawings.

FIG. 1 shows a display example displayed by the image display device according to the present invention. In the figure, 1 is C
RT monitor, 2 a mouse, 3 a pseudo three-dimensional image (endoscopic image) showing the inside of a blood vessel formed by the central projection method,
Reference numeral 4 denotes a longitudinal cross-sectional image of a tubular observation target 4 such as a blood vessel, intestine, or trachea, and reference numeral 6 denotes a cross-sectional image of the observation target 12.

A method for forming the cross-sectional image 4 will be described. FIG. 2 is a diagram showing an embodiment of the method for forming a cross-sectional image. In FIG. 2A, reference numeral 10 denotes a plurality of CT images (CT1, CT2, CT) obtained by an X-ray CT apparatus or the like.
3,...) Are three-dimensional original images stacked,
Indicates an observation target 12 included in the original image 10. Reference numerals 5a to 5g denote viewpoints sequentially set in the observation target 12 by the central projection method.

The viewpoints 5a to 5g are sequentially set by operating a cursor (not shown) on the screen with a mouse or the like while viewing the endoscopic image 3 in FIG. 1, for example. It is set by being automatically updated sequentially toward a distant position. For details of the center projection method and the method of updating the viewpoint, see Japanese Patent Application Laid-Open No. 7-210704.
And JP-A-8-16813.

The viewpoints 5a to 5 set as described above
The curved surface (cut surface) including g includes a plurality of straight lines 5aL to 5gL passing through the viewpoints 5a to 5g, and a straight line 5aL.
2 to 5 gL are the CT images (C
T1, CT2, CT3,...) In the stacking direction y
Are shown for the case parallel to.

Now, a vertical cross section image 4 of the observation object 12 is obtained.
When configuring (FIG. 1), the following is performed. Viewpoint 5
A straight line 5aL parallel to the y-axis passing through a is obtained, and the CT value at each point on the straight line 5aL is calculated using a CT image (CT1, CT2, CT
3, ...). Note that the CT value between each CT image is obtained by interpolation. The CT value on the straight line 5aL thus obtained is stored in the memory 14 (FIG. 2B). Similarly, the CT values on the straight lines 5bL to 5gL passing through the other viewpoints 5b to 5g are stored in the memory 14. However, when the interval between the straight lines (viewpoints) is long, for example, as shown between the straight lines 5cL to 5dL, Straight lines at intervals at which a desired image quality is obtained are obtained by interpolation, and CT values on these straight lines are stored in the memory 14.

Further, as shown in FIG.
, A plurality of straight lines 5aL1, 5aL2, 5aL3.
It is set to rotate by a predetermined angle from the y axis, and the other viewpoint 5
A plurality of straight lines passing through b to 5 g are similarly set. The image data of the first frame is constructed by storing the CT values on each straight line having the same angle as the straight line 5aL1 in the memory, and similarly, the CT values on each straight line having the same angle as the straight lines 5aL2, 5aL3,. Is stored in the memory, it is possible to construct image data cut at any angle with the path connecting the viewpoints as an axis.

FIG. 3 shows a processing flow of the developed image construction method, which will be described step by step.

FIG. 4 is a diagram showing a method of changing the viewpoint to the center position of the observation area (corresponding to step S11 in FIG. 3). First, based on a plurality of viewpoints obtained by the central projection method, a cross section 7 orthogonal to the direction vector of the viewpoint is obtained.
Is set (FIG. 4A), and a density profile of the cross section 7 is created (FIG. 4B). Subsequently, a quadratic difference method is used to discriminate the density values inside and outside the observation target.
The quadratic difference operator is used to extract a region boundary. The area boundary is a point at which the polarity of the secondary difference value shown in FIG. As an operator, Laplacian is applied. Since this is weak against random noise, FIG.
The Laplacian operator of (a) is applied. The operator width can be changed not only the value shown but also the width value.
Depending on the target area, an empirically set value may be used. By this processing, the area boundary is obtained, and the center position of the two area boundaries becomes the center position on the cross section. Further, similarly, by obtaining the center position on the cross section when the cutting angle is changed, the viewpoint can be three-dimensionally changed to the center position of the observation target.

FIG. 4C shows a second method for changing the viewpoint to the center position of the observation area. In this method, when creating an endoscopic image from the central projection method, radial scanning is performed from the viewpoint determined earlier, and based on position information (position information of black circles) at the time of deviating from the observation target area, This is a method in which the center within the observation target is obtained by circular approximation or elliptical approximation, and is set as a new viewpoint. The circle or ellipse approximation is estimated by the least squares approximation. For example, sample point data
If {(X ₁ , Y ₁ ) (X ₂ , Y ₂ )... (X _n , Y _n )}, the objective evaluation function when this is approximated by a circle or an ellipse is

[0018]

(Equation 1) Given by The following circle function and elliptic function are determined by the obtained parameters, and the center coordinates can be estimated.

[0019]

(Equation 2) FIG. 5 is a diagram showing an embodiment of a developed image forming method for scanning the center of the observation target 12 to the outside of the observation target 12 and expanding the shaded cylindrical projection surface to display a pseudo three-dimensional image. It is.

As shown in FIG. 5A, the viewpoints 5a and 5b
And a vector n1 connecting the viewpoints 5b and 5c.
The inner surface of the cylinder orthogonal to each vector ni of 2,... Is defined as a projection plane, and the three-dimensional original image 10 is projected from each viewpoint onto the projection plane with shading. For example, the projection line from the viewpoint 5a is orthogonal to the vector n1 and the projection position a at a certain angle.
To z radially. As a shading method, any of a surface method, a volume rendering method, a depth method, and the like may be used. That is, viewpoints 5a to 5
A curve passing through g is used as a virtual line light source, and the light is projected onto a cylindrical projection surface centered on the line light source with shading.

The image information thus projected on the cylindrical projection surface is developed and stored linearly on the memory 14 as shown in FIG. 5B. The development start position (angle) a is the development position (cut line) indicated by the mouse from outside the observation target shown in FIG. Thus, a pseudo three-dimensional image in which the observation target 12 is cut open can be displayed based on the image information stored in the memory 14. Note that this method is effective only when the viewpoint set by the observation target center position setting unit is set inside the observation target.

On the other hand, when the viewpoint is set outside the observation object, as shown in FIG. 6B, the observation object is located on the line between the position set by the development position setting means and the observation object center position. Reset the viewpoint inside. That is, the viewpoint initially determined by the central projection method or the like is rotated around the viewpoint set by the observation target center position setting unit to the position set by the development position setting unit.
These processes are performed for all viewpoints, and the viewpoint group is reset (step S13 in FIG. 3).

If the viewpoint set by the observation object center position setting means and the development position setting means is determined as the development center point, the scanning angles a to outside the observation object as shown in FIG.
In the case of radial projection in z, the obtained image differs depending on whether the scanning angle is constant (step S14 in FIG. 4).
When scanning and projecting at the same scanning angle (step S1)
5) The spatial resolution of the developed image differs depending on the distance from the development center point. That is, the information expressed for each pixel of the constructed developed image expresses position information having different distances between pixels. Therefore, an isotropic scale is displayed in a mesh form as shown in FIG. 7 to supplement a developed image having anisotropic spatial resolution (step S1).
6). Since the developed image is an image developed and expanded on a plane, it satisfies the demand for intuitively diagnosing the form.

FIG. 8 shows a scanning method for reconstructing a developed image on an isotropic scale. First, the diameter is determined by circular approximation or elliptical approximation. Assuming that the angle is scanned once with respect to the length of the approximate diameter, the angle to be scanned again is set based on the average distance when scanning at an equal angle for a certain angle. After scanning by moving by that angle, the process of scanning again at the same angle from that position and setting the next angle is repeated, and the entire outer periphery of the observation target is scanned (step S17). Since this developed image has an isotropic scale, it is an effective display for measurement. From this, anisotropy,
It is desirable to provide diagnostic images in a mutually displaying manner so as to take advantage of isotropic expanded images.

FIG. 9 is a processing flow showing an embodiment of an image forming method for cutting an original image by a curved surface passing through a plurality of viewpoint paths at the same time and forming a sectional image of the original image including a sectional image of an observation target. FIG. 10 shows a cross-sectional image and the like formed by the processing. First, as shown in FIG. 9, the route (route) of the viewpoint is obtained by the central projection method (step S2).
0), a viewpoint is set at the center of the observation target (step S2)
1). Next, a reference main route is set from the viewpoint routes (step S22). Note that FIG.
In this example, a route set for an image obtained by extracting a region of an observation target by a region extension method or the like, obtaining core line information by a thinning process or the like, and shading by a surface method, a volume rendering method, or a depth method may be used.

Thereafter, a curved surface passing through a plurality of routes is reconstructed. FIGS. 11A and 11B show an image construction method.
First, planes (transverse images 1, 2, and 3) for cutting the observation target at an arbitrary inclination are set in advance, and the position information of the route on each of the transverse images 1, 2, and 3 (that is, the intersection of the cut plane and the route) Is obtained (step S23).

As shown in the figure, when a route is added, the number of intersections with the route differs depending on the position of the cross-sectional image, and a cutting straight line or a cutting curve passing through this intersection must be obtained. If the number of intersections is one or two, it is a straight line,
If there are three or more points, a sequence of points to be cut on the cross-sectional image is calculated using a curve obtained by interpolation and extrapolation by an approximate polynomial (step S24). When the number of intersections is one, the direction of the straight line is not determined. In this case, however, the direction of the straight line is substantially the same as the direction of the cutting line or the cutting curve on the cross-sectional image before and after it. To determine.

A cross-sectional image is constructed by stacking the cutting straight lines or the sequence of points on the cutting curve on each cross-sectional image obtained as described above (step S25).

FIG. 11C is a diagram for obtaining a direction vector for stacking a sequence of points and a cut image center line. Assuming that the point O of the initial center line on the cross-sectional image βn is αn, and the point C of the points A and B on the next cross-sectional image βn + 1 is αn + 1, the vector OC is a direction vector for stacking a point sequence. is there. Since the vector OC is the position where the foot lowered on the curve from the point O is vertical, geometrically,

[0030]

(Equation 3) Is obtained. This process is performed for all the points, and the direction vector of the center line and the point of the center line that are the shortest distance are obtained.

FIG. 12 is a diagram showing an embodiment of an image forming method for forming a curved surface for cutting a complicated meandering observation object. As shown in the figure, in the observation object 12 meandering in a complicated manner, the same cross-sectional image 6 may cross one viewpoint path 5 a plurality of times. Also in this case, a sequence of points to be cut on the cross-sectional image 6 by a cutting straight line or a cutting curve is calculated based on the number of intersection points on the cross-sectional image 6 arbitrarily set as in FIG. The number of intersections is 1
In the case of a point, a sequence of points to be cut is calculated using information on a cutting line or a cutting curve constituted by viewpoint groups on the cross-sectional images before and after the point, and a cut plane image is formed by stacking the point sequence. Is done.

FIG. 13 is a block diagram showing a hardware configuration example of the image display device according to the present invention. As shown in the figure, the image display device mainly includes a magnetic disk 50.
A main memory 52, a central processing unit (CPU) 54,
A display memory 56, a CRT monitor 1, a keyboard 58 for inputting various operation commands, position commands, and menu selection commands, a mouse 2, a mouse controller 60, and a common bus 62 for connecting these components. It is configured.

The magnetic disk 50 stores a three-dimensional original image 10 in which a plurality of CT images (CT1, CT2, CT3,...) Are stacked, an image configuration program, and the like. A program is stored, and an area for arithmetic processing is provided.

The CPU 54 reads out the three-dimensional original image 10 and various programs, uses the main memory 52 to construct a cross-sectional image, a pseudo three-dimensional image, and the like according to the present invention. To the display memory 56,
Display on the CRT monitor 1.

[0035]

As described above, according to the image display apparatus of the present invention, an observation target having a branch such as a blood vessel or a trachea or an observation target such as a complicated meandering intestine can be tertiarily aligned along the observation target. It can be displayed as a longitudinally sectioned image of the original curved surface, and it is possible to provide more effective diagnostic information at a preoperative stage leading to an operation plan.

[Brief description of the drawings]

FIG. 1 is a view showing a display example including a cross-sectional image displayed based on an image display device according to the present invention.

FIG. 2 is a diagram showing an embodiment of a cross-sectional image forming method.

FIG. 3 is a processing flow of a method of constructing a developed image of an observation target.

FIG. 4 is a diagram showing an observation target center calculation method.

FIG. 5 is a diagram showing an embodiment of a developed image construction method.

FIG. 6 is a diagram showing a method for setting a cut line and a development center point.

FIG. 7 is a view showing a developed image having an anisotropic scale.

FIG. 8 is a diagram showing a method of constructing a developed image having an isotropic scale.

FIG. 9 is a processing flow of a method for reconstructing a curved surface passing through a plurality of viewpoint paths at the same time.

FIG. 10 is a view used to explain a method of reconstructing a cross-sectional image by the above processing flow.

FIG. 11 is a diagram showing a method of reconstructing an image by using a curved surface passing through a plurality of viewpoint paths.

FIG. 12 is a diagram illustrating an image forming method of forming a curved surface that cuts a complicated meandering observation target.

FIG. 13 is a block diagram showing a hardware configuration example of an image display device according to the present invention.

[Explanation of symbols]

1 CRT monitor, 2 mouse, 3 endoscopic images, 4
... cross-sectional images, 5aL1 to 5aL3 ... surfaces forming cross-sectional images,
6 cross-sectional image, 7 cross-sectional image, 10-dimensional original image, 12
... observation target, 14 ... memory, 16 ... cut line, 18
... Development center coordinates, 50. Magnetic disk, 52. Main memory, 54. Central processing unit (CPU), 56. Display memory

Claims (1)

[Claims] 1. A means for determining an intersection between a path passing through a substantially center of a tubular observation object included in a three-dimensional original image and each cutting plane for cutting the observation object at a predetermined interval, and A means for obtaining a straight line or a curve on a cutting plane, wherein when the intersection is one point, a means for obtaining a straight line by using information of the straight line or the curve on the cutting plane before and after the intersection; Forming a three-dimensional cutting surface by stacking the straight lines or curves to be formed, and an image forming unit configured to form a cross-sectional image of an original image including the observation target cut by the cutting surface; and the image forming unit. An image display device, comprising: display means for displaying a cross-sectional image.