JP3891442B2 - 3D image processing method - Google Patents

Description

  The present invention relates to a three-dimensional image processing method, and more particularly, to calculate and display an image obtained by projecting a three-dimensional shape onto a two-dimensional plane from three-dimensional data composed of voxel data obtained from an X-ray CT apparatus or the like. The present invention relates to an image processing method.

  As a method of generating a three-dimensional image projected on a two-dimensional plane from three-dimensional data composed of voxels, a volume rendering method capable of obtaining a high-quality image is widely used in the medical field and the like. In addition, in order to display such a three-dimensional image and use it for medical treatment or data analysis, a technique for performing special image processing on a local area as well as merely displaying is known.

  For example, in “Preoperative simulation of cerebellar pontine corner tumor by helical scan CT (HES-CT)” (2nd Computer Surgery Study Group, pp.49-50 (1993)) Therefore, a method called synthesized oblique display is described in which a cross-sectional image at a specified depth position is displayed for a specified region on a three-dimensional image.

Japanese Patent Application Laid-Open No. 3-219377 discloses a method of displaying and synthesizing three-dimensional data of a specified area using the three-dimensional mask when displaying measured three-dimensional voxel data. Yes.
Furthermore, when displaying pre-segmented three-dimensional data, a method of displaying an internal state by displaying a specific organization for a specific region is known, “Surface Rendering”, IEEE CG & A , Vol. 10, pp. A display example is shown in 41-53 (March, 1990).

“Clinical Planning Support System-CliPSS”, IEEE CG & A, vol. 13, no. 6, pp. 76-84 (Nov., 1993) describes an example in which a three-dimensional image obtained by synthesizing a state when a guide tube is inserted into a head tumor is displayed.
In this method, surface data obtained by extracting a tumor and a cranium in advance are processed, and the insertion direction of the guide tube is determined by numerical input of parameters.

JP-A-3-219377 "Preoperative simulation of cerebellopontine angle tumor by helical scan CT (HES-CT)" (Proceedings of the Second Computer Surgery Study Group, pp.49-50 (1993)) “Surface Rendering”, IEEE CG & A, vol. 10, pp. 41-53 (March, 1990) “Clinical Planning Support System-CliPSS”, IEEE CG & A, vol. 13, no. 6, pp. 76-84 (Nov., 1993)

Each of the above prior arts has the following problems.
First, since the above-described synthesized oblique method can only display a cross section, it is difficult to grasp the three-dimensional shape.
Further, in the above-described method of cutting out data by masking a certain area, by rotating the display object, if the depth direction in which the area is cut out and the hole is formed does not match the line-of-sight direction, The area at the bottom of the can not be displayed. In this case, even if it is attempted to remove the data and observe the attention area in the back, the purpose cannot be achieved.

(the purpose)
Accordingly, a basic object of the present invention is to provide a three-dimensional image processing method that makes it possible to grasp the three-dimensional shape of a region of interest and the positional relationship between the entire three-dimensional data and the region of interest from all angles. Is to provide.

  Another object of the present invention is to perform segmentation without segmentation or at high speed and easily because segmentation takes a very long time in the method of segmenting and displaying three-dimensional data for internal display. By providing a three-dimensional image processing method that makes it possible to display an internal region of interest. In this case, how to specify the attention area is important.

  Still another object of the present invention is to provide a three-dimensional solution that can solve this problem in view of the fact that the above-mentioned method for displaying an approach to a tumor requires definition of display data, which takes time for preprocessing. An image processing method is provided. Another object of the present invention is to provide a three-dimensional image processing method that can solve the problem that image quality is a problem in measurement data whose surface definition is difficult, and that the method of specifying the direction of the approach is not intuitive.

The above object of the present invention can be achieved by the following configurations.
That is, first, a three-dimensional shape can be grasped by means for displaying by volume rendering. In addition, the attention point is set in a three-dimensional space, the position of the attention point after rotation is calculated, and the ROI (Region of Interest) that is the region of interest centered on the attention point on the projection plane is automatically calculated. Further, by setting a parameter for volume rendering calculation in the ROI along the line-of-sight direction separately from other areas, and realizing the means for calculating and displaying the line-of-sight method and the ROI Match the depth direction.

Further, a means for displaying the region of interest with high definition is realized by increasing the number of samplings in the rendering in the ROI as compared with other regions.
Furthermore, in order to delete a region having a desired depth, a means for designating the depth of the rendering start point in the ROI is realized.
Also, a means for specifying the size and shape of the ROI is realized.
Furthermore, a means for displaying the position of the target point superimposed on the three-dimensional image when there is no display target between the target point and the projection plane is realized.
As a method for specifying the position of the target point, after the temporary target point is specified, a means for displaying a cross-sectional image that passes through the temporary target point and is parallel to the projection plane, and the cutting position of the cross section are moved in parallel. And means for designating the true attention point on the cross-sectional image.
Furthermore, a means for setting an approach direction parallel to the line-of-sight direction passing through the attention point and a means for displaying the three-dimensional image, the attention point and the approach direction in an overlapping manner are realized.

(Function)
With the above-described means, the ROI is always set around the attention point, and the ROI depth coincides with the line-of-sight direction. Therefore, even if the rotation center and the attention point are different, the attention region can be observed from all angles.
Further, by realizing a means for displaying only the ROI portion with parameters different from those of other regions, the following operations can be provided. First, in many cases, simple segmentation is possible using threshold values, and an object of interest can be displayed simply by changing the threshold value, which is a parameter, without performing extraction.

  If the number of samplings is further increased, a detailed image can be obtained. However, since the amount of calculation basically increases in the order of the third power, there is a problem in the conventional technique in calculating the whole with high definition. By means, it becomes possible to display only the ROI portion with high definition, and since the calculation area is small, it is possible to observe in detail the area of interest in a short calculation time.

Furthermore, by realizing a means for designating the ROI rendering start point, it is possible to display a low-opacity voxel region inside a high-opacity voxel region, and to create a hole with a depth specified for the ROI. You can perform operations like opening them.
In addition, by realizing the means for inputting the size and shape of the ROI, it is possible to secure a field of view of the ROI that matches the size and shape of the region of interest.
Furthermore, by realizing a means for displaying the target point in a three-dimensional image when there is no display target in front of the target point, the effect is as if the target point was written in the three-dimensional data. The three-dimensional position can be confirmed with the same feeling as the three-dimensional data.

In addition, it is possible to set the attention point not only on the surface of an object but also on the center thereof by means of specifying the position of the attention point using the temporary attention point and the cross-sectional image. As a result, the object of interest can always be rotated and displayed at the center of the ROI.
Furthermore, the approach direction can be intuitively designated by setting the approach direction to coincide with the line-of-sight direction.

According to the present invention, a region of interest is always set centering on a portion of interest regardless of the observation direction, and the relationship between the region of interest and the other region is changed to be three-dimensional by changing display parameters interactively. You will be able to grasp. Further, by rendering only the region of interest, high-speed processing can be performed and a high-definition image can be obtained.
Furthermore, it is possible to plan an operation by displaying a display of a hole to be displayed or an approach direction.

Embodiments of the present invention will be described below with reference to FIGS.
FIG. 1 is a flowchart showing a three-dimensional image processing method according to the first embodiment of the present invention, and FIG. 2 is a diagram for explaining a volume rendering method of the present embodiment.
In FIG. 2, 1 is three-dimensional data, 2 is a projection plane, 3 is a pixel whose value is to be projected on the projection plane 2, and 4 is on a line 7 perpendicular to the projection plane from the pixel 3. A resampled voxel set, 5 is a coordinate system A of three-dimensional data, coordinate values are represented by (x, y, z), and 6 is a coordinate system V that performs projection, and the coordinate values are It is represented by (X, Y, Z).
FIG. 3 shows an example of re-sampling voxels of the three-dimensional data 1 in the three-dimensional image processing method according to the present embodiment, and the diagram on the right side is an enlarged part of the diagram on the left side.
In FIG. 3, 11 is the center point of the voxel of the three-dimensional data 1, 12 is the center position of the voxel to be resampled, and 13 is a cell composed of eight voxel centers surrounding the voxel 12 to be resampled.

FIG. 4 is a diagram showing the correspondence between voxel values and opacity in the present embodiment, in which 21 is a threshold value for setting opacity to 0, and 22 is a proportional constant.
FIG. 5 is an example of a head three-dimensional image displayed according to this embodiment. In FIG. 5, 31 is the head surface that is the display target, 32 is the attention point, 33 is the ROI centered on the attention point 32, and 34 is the brain surface that is the display target according to the display parameters.
The operation of the three-dimensional image processing method according to the present embodiment will be described below with reference to FIGS.
First, a three-dimensional image of the entire three-dimensional data is displayed by volume rendering (step 101). This volume rendering method will be described with reference to FIG. The projection plane 2 is arranged with respect to the three-dimensional data 1 as shown in the figure, and the line-of-sight direction is perpendicular to the projection plane.

Each voxel of the voxel 4 resampled into the coordinate system V system to be projected has a value of opacity α and a shaded color C. Re-sampling will be described later. The opacity α takes a value of “0” if the voxel is completely transparent and transmits light, and takes “1” if it is completely opaque and does not transmit light.
The color is a color emitted by the voxel, and the reflection of light on the virtual surface at the voxel position with respect to the color of the voxel itself is a voxel data value (hereinafter referred to as “voxel value”). Considering the inclination vector as the direction of the virtual surface, it is shaded by the direction of the light source, the direction of the projection surface, and the like. At this time, numbers are assigned to the voxels of the voxel set 4 in the order from the projection plane, and the opacity and color of each voxel are α (1), C (1),..., Α (n), respectively. , C (n).

Here, for simplification, it is assumed that the light is parallel light, and the light ray is incident on the three-dimensional data from the projection surface side in a direction perpendicular to the projection surface.
In such a case, the value P projected onto the pixel 3 is calculated by Expression (1) by transmitting light from the voxel closest to the projection surface to the voxel far from the projection plane and adding the reflection amount at each voxel. To do. In the present embodiment, it is assumed that only diffuse reflection is handled in consideration of the simplification of the description and the nature of data to be handled, and the light is parallel light and is perpendicular to the projection plane.

なお、式（１）中のＡ（ｉ）はｉ番目のボクセルに入射する光の透過率を表わしている。

In addition, A (i) in Formula (1) represents the transmittance | permeability of the light which injects into an i-th voxel.

Here, re-sampling will be described with reference to FIG.
As shown in the left diagram of FIG. 3, when trying to resample a certain voxel, among the cells surrounded by the eight voxel centers of the coordinate system A as shown in the right diagram, those voxel centers and resampling are performed. It is assumed that the ratio of the distance to the voxel center 12 to be performed is s: (1-s), t: (1-t), u: (1-u) as shown in the figure. At this time, the resampled voxel values such as opacity and color are obtained by the trilinear method of the following equation (2). Here, Q is used to represent values such as opacity and color.

As a method of performing resampling, another method such as a nearest neighbor method using the value of the original voxel located closest to the voxel to be resampled may be used.
When performing the calculation by the method as described above, parameters for volume rendering calculation are defined.
First, the opacity α is defined by the voxel value, but various displays are possible by changing this association. Therefore, in this embodiment, the opacity is determined in proportion to the voxel value. As shown in FIG. 4, the voxel value f and the opacity α are associated with each other, and the threshold value fth of the voxel value for which the opacity is 0 and the proportionality constant k are used as parameters. This is to make it possible to easily set the opacity, and when displaying three-dimensional data such as a human body measured by X-ray CT, the higher the voxel value, the higher the opacity. This is because when set, natural display can be performed.

According to the above embodiment, the display target can be changed by changing the threshold value fth and the proportionality constant k which are rendering parameters.
In addition, by changing the positional relationship between the three-dimensional data and the projection plane, it is possible to display a three-dimensional image from various directions. Therefore, a matrix for performing coordinate transformation between the coordinate system A in the three-dimensional data space and the coordinate system V in the projection space, particularly in this embodiment, only rotation is considered, and the rotation matrix is used as a parameter.
In this embodiment, for simplification, only the above elements are used as parameters, the voxel color is monochrome, and the direction of the light source is fixed. Of course, these may be added to the parameters and made variable. Needless to say.
Now, when performing volume rendering in step 101 by the above-described method, the voxel on the line when obtaining the projection value of each pixel, which exceeds the threshold fth, from the projection plane of the voxel closest to the projection plane. The distance is stored as a Z buffer (step 102).

Next, the point of interest on the three-dimensional image calculated and displayed as described above is designated with the mouse and displayed (step 103).
With reference to the Z buffer of the designated point of interest, the coordinate value of the point of interest in the coordinate system V after rotation is obtained (step 104).
In step 105, a circle having a predetermined initial radius centered on the point of interest is set as the ROI on the three-dimensional image, and the ROI range is displayed on the three-dimensional image.
The coordinates of the point of interest in the coordinate system V are converted, and the coordinates of the point of interest in the original coordinate A are obtained and stored (step 106).
Thus, the setting of the point of interest and the setting of the ROI are completed, and the display parameter change / input waiting state is entered. The three-dimensional image displayed at this point is as shown in FIG.

Next, a case where the display direction is changed by performing rotation will be described. Here, it is assumed that the parameter given to the rendering processing unit when the rotation is instructed is a rotation matrix for converting the coordinate values of the coordinate system A and the coordinate system V.
A parameter is input (step 107), and it is checked in step 108 if it is a rotation parameter. Here, since the rotation parameter has been input, the routine proceeds to step 109.
In step 109, a coordinate value in the coordinate system V is obtained from the coordinates of the point of interest in the stored coordinate system A using the input rotation matrix.
The coordinates on the projection plane of the target point in the obtained coordinate system V are set as the center. A circular ROI is set and displayed (step 110).
Thereafter, in step 111, volume rendering is performed on the entire projection surface. At this time, it is determined whether or not the pixel position being calculated is within the set ROI area (step 112), and the volume rendering is performed with the ROI display parameters in the ROI and the other areas with the normal display parameters. Then, display is performed (steps 113 and 114). The image displayed at this time is as shown in FIG. 5B, and the ROI 33 centered on the attention point 32 rotates together with the attention point.

Next, when display parameters other than rotation are changed, it is determined in step 114 whether the display parameters for ROI are changed. In the case of the parameters for ROI, volume rendering is performed with the parameters for ROI only for the pixels in the ROI, and only the ROI portion is displayed again (step 116). The image displayed at this time is as shown in FIG. 5C, and a different part can be displayed only for the ROI.
When the normal parameter is changed, volume rendering is performed with the normal parameter only for pixels other than the ROI, and only the portion other than the ROI is displayed again (step 117). In the image displayed at this time, the display part can be changed for a portion other than the ROI as shown in FIG.
When the display is completed, the process waits for the next input. If the rendering parameter is input, the process returns to step 107.
According to the embodiment, the ROI can be set around the attention point, and the attention point can be displayed with the special display parameter even when the rotation is performed. Furthermore, since the ROI depth direction coincides with the line-of-sight direction, the region of interest can be displayed without being hidden.

As an application of the present embodiment, for example, the body surface is displayed as a whole, and the internal organs can be displayed for the region of interest, so that the internal state and the external positional relationship can be easily grasped. Can do.
In the above embodiment, the volume rendering method is used as the projection image calculation method, but other methods may be used. For example, assuming that there is a surface at a voxel position exceeding a threshold value, the inclination of the surface is obtained from the position of the surface voxel, and reflection on the surface is calculated to generate a projection image. A method called a voxel method may be used.
Furthermore, in the above embodiment, the mouse is used to set the attention point, but other coordinate indicating devices such as a keyboard and a pen input device may be used.
Further, when determining the position of the Z buffer, the position of the voxel exceeding the threshold closest to the projection plane is taken, but the position of the voxel with the largest amount of reflected light of the voxel on the line of sight may be taken. This is how to define a virtual surface, and the actual surface is ambiguous in such data, so neither is a problem.

Further, after setting a point of interest once, the process may return to step 103 to reset the point of interest, or the setting of the point of interest may be canceled and normal three-dimensional display may be performed.
In the above embodiment, high-definition rendering can be performed by increasing the number of ROI samplings as a display parameter.
FIG. 6 shows this embodiment. For simplification, it is assumed that the three-dimensional data 1 is parallel to the projection plane 2. 41 is an object on a line existing in the three-dimensional data, 42 is an ROI on the projection plane, 43 is an area on the three-dimensional data corresponding to the ROI, and 44 is an ROI when displayed at the same sampling interval as the area other than the ROI. , 45 is an ROI image displayed at half the sampling interval of the area other than the ROI, and 46 is an area where the object 41 is projected.
A finer and more accurate image can be obtained when the sampling interval is finer as shown in FIG. Since the number of samplings increases, the calculation time increases. However, since only the limited ROI of interest is calculated, the calculation amount is far less than the detailed calculation of the whole.

According to the above embodiment, for example, when a package is inspected by a three-dimensional image using high energy X-ray CT, the entire display takes a certain sampling interval in consideration of the processing time. If an abnormal object such as a pistol is found, the area is marked, and only the surrounding ROI is sampled finely for accurate observation, and rotated and displayed. This makes it possible to grasp the three-dimensional shape.
Here, an embodiment in which the depth of ROI is designated as a rendering parameter for data as shown in FIG. 7 will be described. In FIG. 7, to simplify the description, consider the rendering of a certain section in three-dimensional data. In the figure, 51 is a cross section of the three-dimensional data of interest, 52 is an object existing in the three-dimensional data, and has a round shape 53, a square shape 54, and a triangular shape 55, respectively. The voxel values in the region are 10, 50, 100, and 30 indicated by encircled numbers.

Reference numeral 56 denotes an area where the cross section 51 is projected onto the projection plane, and 57 denotes an area corresponding to the ROI among 53 areas. Reference numerals 58, 59, and 60 indicate the positions of the start points of the rendering calculation in the ROI area.
First, when the calculation is performed from the normal rendering calculation start position 58, if the threshold value is smaller than 10, the surface of the square object is displayed, the round object is displayed in 10 to 50, and the rod-shaped object is displayed in 50 to 100. . Thus, the inside of the object in the ROI can be displayed by changing the threshold value.
Next, when the calculation is performed from the rendering calculation start position 59, if the threshold value is smaller than 30, the surface of the triangular object and the inside of the round object are displayed. In the case of 30 to 50, the inside of the round object is displayed. A rod-shaped object is displayed.
Further, when the calculation is performed from the rendering calculation start position 60, if the threshold is smaller than 50, the inside of the round object is displayed, and 50 to 100, a bar-shaped object is displayed.

As in the above-described embodiment, by setting the rendering start position in the back, it is possible to make a hole in the object and display the inside. Furthermore, a triangular object having a low voxel value in an object having a high voxel value, which could not be displayed from the normal rendering position, can be displayed.
In the above embodiment, the rendering start position is set by the distance from the projection plane. However, other setting methods may be used. For example, the rendering start position may be set by the depth from the surface of the object having a voxel value of 10.
Next, an embodiment for changing the shape and size of the ROI will be described with reference to FIG. In the figure, 35 is the ROI after the change.
First, it is assumed that the initial state of the ROI is, for example, a square centered on the point of interest 32 and one side is 20 in length. At this time, the size of the ROI is changed by dragging the vertex of the square ROI with the mouse. It is assumed that the ROI changes around the attention point, and the changed ROI 35 has 30 on one side. Here, for example, if the ROI shape is selected from the pop-up menu and changed to a circle, the ROI becomes a circle with a diameter of 30. The size of the ROI is changed by dragging this circumference with the mouse.

In the above embodiment, the mouse is used to intuitively change the size, but numerical values such as the length of one side, the diameter, and the radius may be input from the keyboard. The shape of the ROI may be a rectangle or a triangle, and may be a free shape traced with a mouse (however, the point of interest is included inside the ROI). Furthermore, the designation of the ROI shape may not be selected from the menu, and the shape may be switched each time the middle button of the mouse is pressed, for example.
Next, an embodiment for displaying the position of the attention point is shown in FIG. In the figure, reference numeral 61 denotes a display object.
First, the point of interest 32 is set when displayed at the angle (a). Rotating this and displaying it from another angle are (b), (c) and (d).
(B) is a case where the display object 61 exists in the coordinate system V after the rotation, in front of the attention point, that is, between the attention point and the projection plane. In such a case, the position of the attention point is shown. Is not displayed.
In this way, it is possible to obtain the effect of writing the attention point in the three-dimensional data, and display can be performed in a natural context.

However, in the method (b), the position of the point of interest in the shadow of the object cannot be known. Therefore, in (c), the position of the point of interest is displayed in any case. However, with this method, the sense of depth at the point of interest may not be known.
In (d), when an object exists between the attention point and the projection plane, the object is displayed in an average color of the attention point and the object, and displayed in a translucent manner. This has the advantages and disadvantages of both (b) and (c).
Therefore, the above methods (b), (c), and (d) are used in some cases.
The point of interest position is displayed by the method described above using the point of interest coordinates in the V system obtained in step 109 and the Z buffer of the three-dimensional image after step 111 or 116 in FIG.
In the above embodiment, since the attention point can be set only on the surface of the object, an embodiment that can also be set inside the object will be described with reference to FIGS.

FIG. 10 shows a flowchart of the present embodiment, and FIG. 11 shows an example of operations and images corresponding to the flowchart shown in FIG. 65 is an object to be displayed, 66 is a temporary attention point on the object surface, 67 is a three-dimensional image on which the object 65 is projected, 68 is a position of the attention point on the three-dimensional image, and 69 is a temporary attention point. A cross-sectional image 70 of the ROI set around is a true attention point designated on the cross-section.
At this time, it is assumed that the display target in the three-dimensional data 1 is an object having a spherical content 65 as shown in FIG. The attention point setting method will be described below using a flowchart.
First, in step 201, a temporary attention point 68 is designated on the three-dimensional image 67 on which the object 65 is projected. Then, the attention point is set at the surface position 66 of the object in the three-dimensional space (see FIG. 11A).
In the next step 202, a cross-sectional image 69 passing through the temporary attention point 66 set in step 201 and parallel to the projection plane 2 is calculated and displayed in step 203. At this stage, since the cross section is in a position where it touches the surface of the object, it is displayed as a point as shown in FIG.

Subsequently, in order to search for a cross-sectional position designated as a true attention point, the cross-sectional position is translated in step 205 to display as shown in FIG.
When the position for setting the attention point is determined while performing the parallel movement, the process proceeds to step 206 for setting the attention point in step 204, and the position of the true attention point is designated as shown in FIG. The three-dimensional position of the attention point designated on the cross-sectional image can be obtained and stored, and used for the processing after step 107 in the embodiment shown in FIG.
With the above embodiment, it is possible to specify a point of interest inside the object. Thereby, for example, the center of the tumor can always be displayed as the center of the region of interest, and the tumor can be rotated so as not to protrude from the region of interest.
Now, an embodiment for setting and displaying the approach direction to the attention point will be described with reference to the flowchart of FIG. 12 and FIG. The present embodiment can be used, for example, when three-dimensional data obtained by imaging a patient with a tumor is designated as a point of interest and the approach for performing an operation is examined in advance.

This embodiment can be used in combination with the above-described embodiment such as the method of setting the ROI centered on the attention point shown in FIG. 1, but here, for simplicity of explanation, the setting and display of the approach are performed alone. An example is shown. The approach is set in a direction perpendicular to the projection plane through the specified point of interest. In FIG. 13, 71 is an approach direction, 72 is an approach direction unit vector, 73 is a projection image of an object 65, 74 is a projection of a point of interest, and 75 is a projection of an approach.
First, the point of interest is set as in the embodiment shown in FIG. 1 (steps 101, 102, 103, 104, 106).
After setting the attention point, rotation is performed, the display object is observed from various angles, an optimum direction for approaching the attention point is searched, and an approach direction is determined (step 301).
In the post-rotation coordinate system V system, the value in the original coordinate system A system of the unit vector in the Z axis direction (perpendicular to the projection plane) is obtained and stored as an approach direction unit vector (step 302).

The approach direction can be set by the above steps. Next, the set approach is displayed.
When the approach is set, the approach direction is perpendicular to the projection plane, so that it is displayed as a point on the projection plane at a position overlapping the point of interest (FIG. 13A). If the approach direction and the line-of-sight direction do not coincide after rotation, the approach is displayed as a straight line, but at this time, if there is an object between the straight line indicating the approach and the projection plane, the approach is not displayed. (FIG. 13B). In addition, the approach is displayed in a certain direction from the attention point in the direction of the approach direction unit vector. This length is assumed to be initialized.
When an approach is set or parameters such as rotation are input (step 311), in step 312, the coordinates of the point of interest in the rotated V system are obtained.

Next, an approach direction unit vector in the V system is obtained (step 313).
Further, the end point which is not the attention point at the time of displaying the approach is obtained from the approach length, approach direction unit vector and the attention point coordinates already specified (step 314). Therefore, a straight line connecting the target point and the end point is projected onto the projection plane, and the Z buffer in the V system at that time is stored (step 315). Volume rendering is performed to obtain a three-dimensional image (step 316), and the obtained Z buffer (step 317) is compared with the approach Z buffer, and the image Z buffer is closer to the projection plane. Displays a 3D image, otherwise it displays an approach (step 318).
Thereafter, the attention point is displayed as in the embodiment of FIG. 9, for example.
According to the present embodiment, it is possible to perform natural combined display in consideration of the context, but other methods may be used as long as the approach and the three-dimensional image are combined and displayed.
Further, the length and thickness of the approach may be changed.

According to the above embodiment, the approach direction can be determined and then rotated, and the approach can be examined by observing from various angles. Here, by performing a display in which the set approach direction and the line-of-sight direction coincide with each other, it is possible to simulate a state of observing from the direction approached during the operation. FIG. 14 shows an embodiment for displaying the approach view.
The following processing is added to the flowchart in FIG.
After the approach direction is determined in step 302, the rotation matrix from the coordinate system A to the coordinate system V at the current display angle is stored (step 401). Thereby, the approach view direction is stored.
Here, when an instruction to perform approach view display is input (step 411), the rotation matrix stored in step 401 is input as a parameter (step 412).
Then, the process proceeds to step 311 in FIG. 12, and an approach view in which the approach direction coincides with the line-of-sight direction is displayed according to the embodiment in FIG.

With the above-described embodiments, a surgical approach can be planned, but then the surgical planning can be further supported by determining where on the body surface to approach.
FIG. 15 shows an embodiment for obtaining the approach surface position.
First, before setting the attention point and approach, volume rendering is performed, and parameters for displaying the body surface are set and stored (step 501). In this embodiment, a threshold value is stored.
Here, after the approach direction is set in steps 301 and 302 in the embodiment of FIG. 12, the parameter (threshold value) stored in step 501 is called and input to the rendering parameter (step 511).
Then, according to the embodiment (steps 311 to 318) shown in FIG. 12, the three-dimensional image of the body surface, the approach, and the attention point are combined and displayed. At this time, since the approach is perpendicular to the projection plane, the approach becomes a point, and the approach and the point of interest overlap, as shown in FIG.

Then, the position where the body surface and the approach direction intersect is obtained by the Z buffer obtained in step 317 (step 512). Since the obtained approach surface position is a V-system coordinate, it is converted into an A-system coordinate and stored (step 513).
Thereafter, the stored approach surface position is converted from the stored coordinate system A to the coordinate system V when displaying the attention point, and displayed in the same manner as in the embodiment of FIG.
According to the above embodiment, the approach direction and approach surface position can be examined before surgery.
Each of the above-described embodiments shows an example of the present invention, and it goes without saying that the present invention should not be limited thereto.

6 is a flowchart illustrating a method of setting a region of interest centered on a point of interest and a rendering method according to an embodiment of the present invention. It is a figure explaining the Example of volume rendering. It is a figure explaining the Example of resampling. It is a figure which shows the example of a correspondence of opacity and a voxel value. It is a figure explaining the example of a rendering result when an attention point and a region of interest are set. It is a figure which shows the example when performing a high-definition display. It is a figure explaining the Example which changes a rendering calculation start position. It is a figure which shows the example which changes the magnitude | size and shape of a region of interest. It is a figure which shows the example which displays an attention point. It is a figure which shows the example which designates an attention point inside an object. It is a figure which shows the example which designates an attention point inside an object. It is a flowchart which shows the example of the setting and rendering of an approach direction. It is a figure which shows the example of a display of an approach direction. It is a flowchart which shows the example of the return to an approach view. It is the flowchart which showed the Example which calculates | requires the display position on an approach.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 3D data 2 Projection surface 5 Original coordinate system of 3D data 6 Projection coordinate system after rotation 21 Threshold value 22 Proportional constant defining opacity 32 Point of interest 33 Region of interest 66 Temporary point of interest 70 True point of interest 71 Approach direction

Claims (3)

A three-dimensional image processing method for generating and displaying an image obtained by projecting a display target object in the three-dimensional data onto two dimensions using three-dimensional data composed of a three-dimensional array of voxel data,
At least a step of designating a point of interest on the three-dimensional image, a step of obtaining a depth coordinate value of the point of interest on the designated three-dimensional image, a step of storing the three-dimensional coordinate value of the point of interest, A step of setting an approach to the attention point, which is a line passing through the attention point in the line-of-sight direction, and each time a projection image generation calculation parameter is changed / input, the three-dimensional image by the parameter, the attention point, A three-dimensional image processing method comprising a step of combining and displaying an approach to a point of interest.   The three-dimensional image processing method according to claim 1, further comprising the step of matching the approach direction and the line-of-sight direction by one operation in addition to the steps. In addition to the above steps, a step of storing projection image generation calculation parameters when displaying the surface of the object, a step of changing the parameters to display the object surface when setting an approach to the point of interest, and an approach A step of obtaining a position of a point on the object surface, and instead of the step of combining and displaying according to claim 1, each time a parameter for calculation of projection image generation is changed / input, 2. The three-dimensional image processing method according to claim 1, further comprising the step of combining and displaying a three-dimensional image, a point of interest, an approach to the point of interest, and a point on the object surface on the approach.

Images