JPH07121737A - Mip processing method, mip processor, method and device for surface image display - Google Patents

Abstract

PURPOSE:To provide the method and device for performing MIP processing I at high speed. CONSTITUTION:A projective reference plane rotation arithmetic part 15 of an MIP processor 100 estimates a projective reference plane vertical to a reference projective direction, takes three reference points not existing on a straight line at least, respectively rotates forward the reference points just at an angle formed by the applied projective direction and the reference projective direction and calculates a new projective reference plane specified by the respective reference points after the forward rotation. An interpolation MIP processing part 16 calculates voxel data along a normal to the new projective reference plane at respective lattice points on the new projective reference plane by performing interpolation processing, inspects those voxel data, calculates one piece of projection data among a maximum value, minimum value, average value and differential value and further calculates a projective image from the projection data concerning the respective lattice points. Thus, processing time can be shortened and real-time MIP processing is enabled as well.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention relates to MIP (generally
It is an abbreviation for Maximum Intensity Projection, but it includes not only the projection image of the maximum value but also the processing of obtaining the projection image of the minimum value, the average value, the difference value, etc. as described later) Processing method, MIP processing device, and surface image display More specifically, the present invention relates to a method and a surface image display device, and speeds up MIP processing for inspecting voxel data along a given projection direction to obtain at least one projection image of a maximum value, a minimum value, an average value and a difference value. MIP processing method and MIP processing apparatus capable of performing, and surface image display method and surface image capable of rapidly generating and displaying a surface image composed of voxel data of the surface of volume data viewed from a given projection direction Regarding display device.

[0002]

2. Description of the Related Art FIG. 9 is a block diagram of an example of a conventional MIP processing apparatus. Slice data acquired by scanning the patient K with a medical image diagnostic apparatus A such as CT, MRI, or ultrasonic diagnostic apparatus is passed to the MIP processing apparatus 500.
The MIP processing device 500 includes a computer 51 and a console 2
And a CRT 3 are provided. The computer 51 includes a slice data reading unit 11, a mask processing unit 12, a volume data conversion unit 13, and a projection direction designating unit 1.
4, volume data reverse rotation calculation unit 55, interpolation processing unit 5
6, it has the functions of the MIP processing unit 57 and the imaging unit 18.

FIG. 10 is a flowchart of the operation of the MIP processing apparatus 500. In step V1, the slice data reading unit 11 reads slice data from the medical image diagnostic apparatus A. In step V2, the mask processing unit 12 applies a special value (a value that does not appear as data, for example, -102, to data existing in a region outside the region of interest.
4) and mask data outside the region of interest. At step V3, the volume data conversion unit 13 converts the mask-processed slice data into volume data VK as shown in FIG. This volume data V
K is a reference projection direction KK (αx, αy, αz) described later.
have. Note that (αx, αy, αz) is the direction cosine of the reference projection direction KK with respect to the x-axis, y-axis, and z-axis.

In step V4, the user operates the console 2
12, the desired projection direction KR (βx, βy, βz) is designated through the projection direction designation unit 14 as shown in FIG. Note that (βx, βy, βz) is x in the projection direction KR.
It is a direction cosine with respect to the axis, the y axis, and the z axis.

At step J5, the volume data reverse rotation calculation unit 55 reversely rotates the volume data VK by an angle formed by the designated projection direction KR and the reference projection direction KK, as shown in FIG. Obtain new volume data VR. This is done by actually rotating all voxel data of the volume data VK by affine transformation by θz around the z axis, θy around the y axis, and θx around the x axis. For the new volume data VR, the designated projection direction K
R coincides with the reference projection direction KK.

At step J61, the interpolation processing section 56
From each voxel data of the new volume data VR, the voxel data of each lattice point of the xyz coordinate system is calculated by interpolation processing (for example, nearest neighbor interpolation processing, linear interpolation processing, etc.). In step J62, the MIP processing unit 57 inspects the voxel data of each grid point calculated above along the reference projection direction KK (= the designated projection direction KR), and determines the maximum value, the minimum value, and the average value of the voxel data. , At least one projection image of the difference value is obtained. Step V7
Then, the imaging unit 18 converts the projected image into a CRT image.
Display in 3. In step V8, the user selects the projection direction K
Check if you want to change R. User projected direction KR
If is changed, the process returns to step J4. If the user does not change the projection direction KR, the process ends.

As can be understood from the above description, the reference projection direction KK is the volume data reverse rotation calculation unit 55.
And the projection direction KR when the calculation of the interpolation processing unit 56 is not actually required, and has a fixed positional relationship with the volume data VK. It is also the direction of MIP processing in the MIP processing unit 57.

A surface image can be obtained by performing a process of searching voxel data on the surface of the volume data viewed from the projection direction instead of the MIP process.

[0009]

In the conventional MIP processing apparatus 500, all the voxel data of the volume data VK have the designated projection direction KR as the reference projection direction K.
The volume data reverse rotation calculation is performed to reversely rotate by the angle formed with respect to K. However, the volume data V
The number of voxels of K is very large (for example, 256 × 256 ×
Therefore, even if the number of voxels is reduced by the mask processing, a huge processing time is required and real-time MIP processing is difficult. This is the same when obtaining a surface image. Therefore, a first object of the present invention is to provide a MIP processing method and a MIP processing apparatus that shorten the processing time and enable real-time MIP processing. A second object of the present invention is to provide a surface image display method and a surface image display device capable of reducing the processing time and generating and displaying a surface image in real time.

[0010]

According to a first aspect of the present invention, the present invention inspects voxel data along a given projection direction and projects at least one of a maximum value, a minimum value, an average value and a difference value. In the MIP processing method for obtaining an image, assuming a projection reference plane perpendicular to the reference projection direction of volume data, at least three reference points that are not on a straight line are set on the projection reference plane, and the given projection direction is Each of the reference points is rotated by an angle formed with the reference projection direction, and the voxel data is inspected along the normal line to the new projection reference plane defined by each reference point after the rotation, and the maximum value, the minimum value, and the average value are measured. , MIP processing method characterized by obtaining at least one projection image of the difference value.

According to a second aspect, the present invention provides a projection direction designating means for designating a projection direction and at least three points which are on a projection reference plane perpendicular to the reference projection direction of the volume data and not on a straight line. Of the selected reference point, the reference point is rotated by an angle formed by the specified projection direction and the reference projection direction, and a new projection reference plane defined by each reference point after rotation is obtained. Computing means,
MIP processing means for inspecting voxel data along a normal to the new projection reference plane to obtain at least one projected image of maximum value, minimum value, average value, and difference value. Provide a device.

According to a third aspect, the present invention provides a surface image display method for displaying a surface image composed of voxel data of the surface of volume data viewed from a given projection direction, in which the volume data is perpendicular to the reference projection direction. Assuming a different projection reference plane, at least three reference points that are not on a straight line are set on the projection reference plane, and each of the reference points is rotated by an angle formed by the given projection direction and the reference projection direction. Search voxel data in the backprojection direction along the normal to the new projection reference plane defined by each of the reference points, find the voxel data on the most surface, and generate and display the surface image from these voxel data. A surface image display method is provided.

According to a fourth aspect, the present invention provides a projection direction designating means for designating a projection direction, and at least three points which are on a projection reference plane perpendicular to the reference projection direction of the volume data and not on a straight line. Of the selected reference point, the reference point is rotated by an angle formed by the specified projection direction and the reference projection direction, and a new projection reference plane defined by each reference point after rotation is obtained. Computing means,
Surface image generation means for searching the voxel data in the back projection direction along the normal to the new projection reference plane to obtain the voxel data at the most front surface to generate a surface image, and a surface image display for displaying the surface image. There is provided a surface image display device comprising:

[0014]

In the MIP processing method and the MIP processing apparatus according to the present invention, a projection reference plane perpendicular to the reference projection direction is assumed, and at least three reference points that are not on a straight line are set on the projection reference plane, and a given projection is given. Each of the reference points is positively rotated by an angle formed by the direction with respect to the reference projection direction. Then, the voxel data is inspected along the normal line to the new projection reference plane defined by each reference point after the normal rotation, and at least one projected image of the maximum value, the minimum value, the average value and the difference value is obtained. In this way, since only at least three reference points need to be rotated, the processing time can be greatly shortened and real-time MIP processing can be performed.

In the surface image display method and the surface image display device according to the present invention, a projection reference plane perpendicular to the reference projection direction is assumed, and at least three reference points that are not on a straight line are provided on the projection reference plane. Each of the reference points is positively rotated by an angle formed by the projection direction with respect to the reference projection direction. Then, the voxel data is searched in the back projection direction (that is, toward the surface) along the normal line to the new projection reference plane defined by each reference point after the forward rotation, and the voxel data at the most front surface is obtained. In this way, since only at least three reference points need to be rotated, the processing time can be greatly shortened and a surface image can be generated and displayed in real time.

[0016]

The present invention will be described in more detail with reference to the embodiments shown in the drawings. The present invention is not limited to this. FIG. 1 is a block diagram of an embodiment of the MIP processing apparatus of the present invention. The slice data acquired by scanning the patient K with the medical image diagnostic apparatus A such as CT, MRI, and ultrasonic diagnostic apparatus is the MIP processing apparatus 100.
Passed to. The MIP processing device 100 comprises a computer 1, a console 2 and a CRT 3.
The computer 1 includes a slice data reading unit 11, a mask processing unit 12, a volume data conversion unit 13, a projection direction designating unit 14, a projection reference plane rotation calculation unit 15, and an interpolation / MIP.
It has the respective functions of the processing unit 16 and the imaging unit 18.

FIG. 2 is a flowchart of the operation of the MIP processing apparatus 100. In step V1, the slice data reading unit 11 reads slice data from the medical image diagnostic apparatus A. In step V2, the mask processing unit 12 applies a special value (a value that does not appear as data, for example, -102, to data existing in a region outside the region of interest.
4) and mask data outside the region of interest. In addition,
The region of interest shall be a convex body. At step V3, the volume data conversion unit 13 converts the mask-processed slice data into volume data VK as shown in FIG. This volume data VK
Has a reference projection direction KK (αx, αy, αz).

In step V4, the user operates the console 2
The desired projection direction KR (βx, βy, βz) is designated via the projection direction designating unit 14 as shown in FIG.

In step V5, the projection reference plane rotation calculation unit 15 translates the volume data VK so that the center of gravity of the region of interest coincides with the origin (0,0,0), and is shown in FIG. Obtain volume data VB. This parallel movement is for facilitating later rotation calculation and the like. next,
As shown in FIG. 4B, assuming a projection reference plane MK that is perpendicular to the reference projection direction KK and passes through the origin (0,0,0), four reference points P1, P2, P3 that are not on a straight line , P4
On the projection reference plane MK. Next, as shown in (c) of FIG. 4, the reference points P1 and P are formed by an angle formed by the designated projection direction KR and the reference projection direction KK.
2, P3, P4 are rotated forward and new reference points R1, R
2, R3, R4 are obtained. This is actually the reference point P
The voxel data of 1, P2, P3, and P4 are positively rotated by affine transformation by θz around the z axis, θy around the y axis, and θx around the x axis. The new projection reference plane MR defined by the new reference points R1, R2, R3, R4 is perpendicular to the designated projection direction KR. However, this new projection reference plane MR is generally not correctly stored in the volume data VK. Therefore, new reference points R1 and R are added so that the internal relation between the new projection reference plane MR and the volume data VK is not excessive or insufficient.
2, R3 and R4 are moved within a plane including the new projection reference plane MR. Then, the reference points R1, R2, R3, R after the movement
The plane defined by 4 is the final new projection reference plane MR.
And

In step V6, the interpolation / MIP processing unit 1
6 performs interpolation and MIP processing. 5 and 6
The flowchart of interpolation and MIP processing is shown in FIG. In step B1, as shown in FIG. 7, the new projection reference plane MR
A unit vector Φ and a unit vector Ψ intersecting above are obtained. In step B2, as shown in FIG. 7, the new projection reference plane M is calculated from the unit vector Φ and the unit vector Ψ.
Determine the unit normal vector Π and the unit normal vector −Π for R.

At step B3, attention is paid to one of the lattice points on the new projection reference plane MR. In step B4, the voxel data of the grid point is obtained by interpolation processing (for example, nearest neighbor interpolation or linear interpolation). At step B5, it is checked whether the voxel data is larger than a special value (for example, -1024). If the voxel data is larger than the special value, the grid point is in the region of interest.
Go to step B6. If the voxel data is not larger than the special value, the grid point is outside the region of interest.
Go to step B15.

At step B6, a unit distance is advanced in the direction of the unit normal vector Π to obtain a grid point. Step B
In step 7, voxel data of the grid point is obtained by interpolation processing. In step B8, it is checked whether the voxel data is larger than a special value (for example, -1024). If the voxel data is larger than the special value, the lattice point is in the region of interest, and the process proceeds to step B9.
If the voxel data is not larger than the special value, the grid point is outside the region of interest, and the process proceeds to step B10. In step B9, if there is no projection data obtained so far, the voxel data is set as the projection data. If the projection data obtained up to that time exists, MIP processing is performed with the projection data and new voxel data, and the projection data is updated. Then, the step B6
Return to.

In step B10, the grid point of interest on the new projection reference plane MR is returned to. In step B11, a lattice point is obtained by advancing by a unit distance in the direction of the unit normal vector −Π. In step B12, the voxel data of the grid point is obtained by interpolation processing and stored in the memory. At step B13, it is checked whether the voxel data is larger than a special value (for example, -1024). If the voxel data is larger than the special value, the grid point is in the region of interest, and the process proceeds to step B14. If the voxel data is not larger than the special value, the grid point is outside the region of interest. Therefore, the projection data obtained up to that point is stored in the memory, and then the projection data is erased.
Go to 5. In step B14, if there is no projection data obtained so far, the voxel data is set as the projection data. If the projection data obtained up to that time exists, MIP processing is performed with the projection data and new voxel data, and the projection data is updated. Then, the process returns to step B11.

In step B15, it is checked whether or not the processing in step B5 has been completed for all the grid points on the new projection reference plane MR. If all grid points have not ended, the process proceeds to step B16. If all grid points have been completed, the process proceeds to step B17. Step B16
Then, move to an unprocessed grid point on the new projection reference plane MR,
The procedure returns to step B4. In step B17, a projection image is obtained from the projection data of each grid point stored in the memory in step B14. Then, the process proceeds to step V7 in FIG.

In step V7 of FIG. 2, the imaging unit 18
Displays the projected image on the CRT 3. In step V8, it is checked whether the user wants to change the projection direction KR. If the user changes the projection direction KR, the process returns to step V4. If the user does not change the projection direction KR, the process ends.

According to the above MIP processing apparatus 100, 4
Since only one reference point P1, P2, P3, P4 needs to be rotated, the processing time can be greatly shortened and real-time M
IP processing is also possible.

It should be noted that in steps B9 and B14, the voxel data is simply stored in the memory, and when the process proceeds from step 13 to step 15, all the voxel data stored in the memory is inspected to find the maximum and minimum values. , At least one of the average value and the difference value may be obtained and stored in the memory. In this case, since all voxel data is stored in the memory, there is another advantage that it can be used. On the other hand, in the case of the above embodiment, since the MIP processing is performed every time the voxel data is obtained, there is an advantage that the memory capacity is small and no extra calculation processing is required.

The configuration of an embodiment of the surface image display device of the present invention is such that a surface image generation processing section is provided instead of the MIP processing section 16 in FIG. Further, in that operation, interpolation and projection data search processing is executed instead of step V6 of FIG. FIG. 8 shows a flow chart of the interpolation and projection data search processing. In step C1, as shown in FIG. 7, a unit vector Φ and a unit vector Ψ that intersect on the new projection reference plane MR are obtained. Step C2
Then, as shown in FIG. 7, a unit normal vector Π with respect to the new projection reference plane MR is obtained from the unit vector Φ and the unit vector Ψ. This unit normal vector Π points in the back projection direction.

At step C3, attention is paid to one of the lattice points on the new projection reference plane MR. In step C4, the voxel data of the grid point is obtained by interpolation processing (for example, nearest neighbor interpolation or linear interpolation). At Step C5, it is checked whether the voxel data is larger than a special value (for example, -1024). If the voxel data is larger than the special value, the grid point is in the region of interest.
Go to step C6. If the voxel data is not larger than the special value, the grid point is outside the region of interest.
Go to step C10.

At step C6, a unit distance is advanced in the direction of the unit normal vector Π to obtain a grid point. Step C
In step 7, voxel data of the grid point is obtained by interpolation processing. At Step C8, it is checked whether the voxel data is larger than a special value (for example, -1024). If the voxel data is larger than the special value, the lattice point is in the region of interest, so the process proceeds to step C9.
If the voxel data is not larger than the special value, the lattice point is outside the region of interest, so the surface data is stored in the memory, and the process proceeds to step C10. In step C9, the voxel data is set as surface data. And
The procedure returns to step C6.

In step C10, it is checked whether or not the processing in step C5 has been completed for all the grid points on the new projection reference plane MR. If all the grid points have not ended, the process proceeds to step C11. If all grid points have been completed, the process proceeds to step C12. Step C11
Then, move to an unprocessed grid point on the new projection reference plane MR,
The procedure returns to step C4. In step C12, a surface image is obtained from the surface data of each grid point stored in the memory in step C8. Then, the process proceeds to step V7 in FIG.

According to the above surface image display device, since it is necessary to rotate only the four reference points P1, P2, P3 and P4, the processing time can be greatly shortened and the surface image can be generated and displayed in real time. Becomes

[0033]

The MIP processing method and MIP of the present invention
According to the processing device, when performing the MIP process of inspecting the voxel data along the given projection direction and obtaining at least one projection image of the maximum value, the minimum value, the average value, and the difference value, at least three points Since it is only necessary to rotate the reference point, the processing time can be greatly shortened and real-time MIP processing can be performed. According to the surface image display device of the present invention, at the time of displaying the surface image composed of voxel data of the surface of the volume data viewed from the given projection direction, it suffices to rotate at least three reference points. The processing time can be greatly reduced, and surface images can be generated and displayed in real time.

[Brief description of drawings]

FIG. 1 is a block diagram of an embodiment of a MIP processing device of the present invention.

FIG. 2 is a flowchart illustrating an operation of the MIP processing device of FIG.

FIG. 3 is a conceptual diagram illustrating a relationship between volume data, a reference projection direction, and a projection direction.

FIG. 4 is a conceptual diagram illustrating rotation of a projection reference plane.

FIG. 5 is a flowchart illustrating an operation of a MIP processing unit.

FIG. 6 is a subsequent flowchart illustrating the operation of the MIP processing unit.

FIG. 7 is a conceptual diagram of a projection reference plane and a unit normal vector.

FIG. 8 is a flowchart illustrating an operation of a front surface image generation processing unit.

FIG. 9 is a block diagram of an example of a conventional MIP processing device.

10 is a flowchart illustrating an operation of the MIP processing device of FIG.

FIG. 11 is a conceptual diagram of volume data.

FIG. 12 is a conceptual diagram illustrating a relationship between volume data and a reference projection direction and a projection direction.

FIG. 13 is a conceptual diagram of reverse rotation of volume data.

[Explanation of symbols]

100,500 MIP processing device 1 computer 2 console 3 CRT 14 projection direction designation unit 15 projection reference plane rotation calculation unit 16 interpolation / MIP processing unit KK reference projection direction KR projection direction MR projection reference plane VK, VB volume data Φ, Ψ projection Unit vector on the reference plane Π, −Π Unit vector normal to the projected reference plane

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification number Office reference number FI technical display location 9287-5L G06F 15/62 390 B

Claims (4)

[Claims] 1. A MIP processing method for inspecting voxel data along a given projection direction to obtain at least one projection image of a maximum value, a minimum value, an average value, and a difference value. Assuming a projection reference plane perpendicular to the direction, at least three reference points that are not on a straight line are set on the projection reference plane, and the reference points are rotated by an angle formed by the given projection direction and the reference projection direction. , The voxel data is inspected along the normal to the new projection reference plane defined by each reference point after rotation, and at least one projection image of the maximum value, the minimum value, the average value, and the difference value is obtained. And MIP processing method. 2. A projection direction designating means for designating a projection direction, and at least three reference points which are on a projection reference plane perpendicular to the reference projection direction of the volume data and which are not on a straight line are selected. Projection reference plane rotation calculating means for respectively rotating the reference points by an angle formed by the designated projection direction and the reference projection direction, and obtaining new projection reference planes defined by the rotated reference points, and the new projection reference. A MIP processing device comprising: MIP processing means for inspecting voxel data along a normal to a surface to obtain at least one projected image of a maximum value, a minimum value, an average value and a difference value. 3. A surface image display method for displaying a surface image consisting of voxel data of the surface of volume data viewed from a given projection direction, assuming a projection reference plane perpendicular to the reference projection direction of the volume data, At least three reference points that are not on a straight line are set on the projection reference plane, the reference points are rotated by an angle formed by the given projection direction and the reference projection direction, and are defined by the respective reference points after rotation. A surface image display method characterized by searching voxel data in the backprojection direction along the normal to the new projection reference plane to find the voxel data on the most surface, and generating and displaying the surface image from these voxel data. . 4. A projection direction designating means for designating a projection direction, and at least three reference points which are on a projection reference plane perpendicular to the reference projection direction of the volume data and which are not on a straight line are selected. Projection reference plane rotation calculating means for respectively rotating the reference points by an angle formed by the designated projection direction and the reference projection direction, and obtaining new projection reference planes defined by the rotated reference points, and the new projection reference. A surface image generating means for searching the voxel data in the back projection direction along the normal to the surface to obtain the voxel data on the most surface to generate a surface image, and a surface image display means for displaying the surface image. A surface image display device characterized by the above.