JP2007190413A - X-ray ct apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an X-ray CT apparatus capable of reconstructing images of high image quality. <P>SOLUTION: An image reconstruction part 31 is provided with with an interpolation processing part 33, a filtered back projection part 35 to perform image-reconstruction of data on a desired slice position by a filtered back projection method, a memory 37 used commonly by the interpolation processing part 33 and the filtered back projection part 35, and an information storage part 39 to store parameters etc. The interpolation processing part 33 carries out a mathematically correct filtering process at high speed by determining a weighing factor for data included in a range in the slice direction among data stored in the collected data storage device 29 on the basis of size of the range and the sampling position of the data. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

The present invention relates to an X-ray CT apparatus, and more particularly, X-ray C that can reduce image noise by filtering data obtained by helical scanning in the slice direction.
It relates to the T device.

(1) Single slice CT
In recent years, an X-ray CT apparatus has an X-ray focal point for generating a fan-shaped X-ray beam (fan beam) and a detector in which a large number of channels, for example, 1000 channels of detection elements are arranged in a row. Single-slice CT having

The X-ray focal point and the detector are rotated around the subject, and X-ray intensity data (referred to as projection data) that has passed through the subject is collected. Collecting projection data, for example 1000 times in one rotation,
Based on this data, an image is reconstructed by a method described later. One data collection is referred to as one view, data of one detection element in one view is combined with one beam, and all beams in one view (data of all detection elements) are collectively referred to as actual data.

(2) Two scanning methods Two scanning methods of the X-ray CT apparatus will be described.
The first scan method is a conventional scan. As shown in FIG. 6A, this is a scanning method in which a target section, for example, the periphery of the section A is rotated once. When it is desired to obtain images of a plurality of cross sections, for example, the cross section A and the cross section B, as shown in FIG. 6A, first, data is collected while rotating once around the cross section A, and then the subject is placed. Sleeper or X-ray focus 10
1 and the detector 103 are moved so that the section B and the rotation surface are aligned. After that, data is collected while rotating around the subject like the cross section A. Therefore, when the imaging range is wide in the body axis direction (Z-axis direction) of the subject, the imaging time becomes long when there are many target cross sections.

The second scanning method is a helical scan. As shown in FIG. 6 (b), the X-ray focal point and the detector are continuously rotated, and the bed is moved in the opposite axis direction of the subject in synchronization with the rotation to collect data. The locus of the X-ray focal point 101 scans around the subject in a spiral. This scanning method can scan a wide area at high speed.

The coordinate system is defined as shown in FIG. The XY plane corresponds to the cross sections A and B scanned by the conventional scan, the Z-axis direction is the direction opposite to the subject, and the direction called the slice direction in the single slice CT.

(3) Conventional Scan Image Reconstruction The image reconstruction of the X-ray CT apparatus will be briefly described. In the case of a conventional scan, it consists of the following three steps.

[1] Data collection and correction Collect data by conventional scanning. The rotation angle is usually 360 °, 180 ° + fan angle, or the like. The projection data is corrected in consideration of various factors such as the sensitivity of the detector 103 and the X-ray intensity, and raw data is obtained.

[2] Convolution calculation with reconstruction function The convolution of the raw data of each angle and the reconstruction function is performed.

[3] Backprojection calculation All pixels on the X-ray passing path when the convolution data is collected (
Pixel). When this back projection operation is repeated by a necessary angle, the original cross-sectional image is reconstructed.

(4) Helical Scan Image Reconstruction FIG. 8 is a side view of the two scan methods shown in FIG. 6, the conventional scan and the helical scan. Horizontal axis is slice (Z axis) direction, horizontal axis is rotation phase (angle)
And the sampling positions of each data are connected by arrows. Hereinafter, such a diagram is referred to as a scan diagram.

In the conventional scan shown in FIG. 8A, 360 ° data necessary for the target slice plane corresponding to the above-described [1] is collected, and [1] to [3] as described above.
The image can be reconstructed by the steps of

On the other hand, in the helical scan of FIG. 8B, only one view is collected on the target slice plane because of the helical scan. Therefore, instead of the above [1], the raw data obtained by correcting the collected projection data is interpolated in the Z-axis direction to obtain necessary data, and then the above [2] to [3] are performed. Typical interpolation methods in single slice CT are the following two types (a) and (b).

(A) 360 ° interpolation method As shown in FIG. 9A, the 360 ° interpolation method is the distance between the slice plane and the sampling position with the actual data of two views having the closest in-phase sandwiching the target slice position. It is a method of linear interpolation with the inverse ratio of.

For example, if the target slice position (Z coordinate of the slice plane) is Z = Z0, the data collected at this slice position is only one view at a phase of 0 °.

Therefore, for example, when obtaining the data of the phase θ, the actual data 1 on the upper side of the slice position and the actual data 2 on the lower side are selected, and the distance between the Z coordinate obtained by sampling each data and the target slice position Z0 (Z Linear interpolation is performed with the inverse ratio of coordinates) to obtain interpolation data. This is repeated for all necessary phases.

(B) Counter beam interpolation method This method uses a counter beam that is virtual data. As shown in FIG. 9C, the beam to each detection element of the actual data collected when the focus is at the position of the black circle is as indicated by a solid arrow. At this time, the left beam 1 and the dotted beam when the X-ray focal point is at the position of the white circle are beams that pass through the same path. The beam from the white circle is referred to as a counter beam. Similarly, the beam 2 and the dotted beam from light gray, and the beam 3 and the dotted beam from dark gray are a beam and an opposite beam that pass through the same path. Thus, all the beams in the black circle have opposite beams.

Therefore, a method of extracting virtual beams (referred to as opposed data) by extracting the opposed beams corresponding to each beam from the data of the focus positions of white circles, light gray and dark gray, and linearly interpolating between the actual data and the opposed data. Is the counter beam interpolation method.

In the case of the helical scan, the sampling position of the opposite data is as shown in FIG.
Although it is different for each beam (each channel), in the following, it is represented by the sampling position of the central channel and displayed by a dotted line as shown in FIG. In addition, several other methods for interpolating the helical scan have been proposed, such as a method using a nonlinear function for interpolation.

As shown in FIG. 9, the distance between the sampling positions of two data used for interpolation is called an interpolation interval. The interpolation interval is equivalent to a helical pitch in the 360 ° interpolation method, and is ½ of the helical pitch in the counter beam interpolation method. Therefore, the opposed beam interpolation method is narrower. Since the effective slice thickness in the helical scan becomes thinner as the interpolation interval becomes narrower, the counter beam interpolation method becomes thinner.

(6) Multi-slice CT
Multi-slice CT having a plurality of detector rows, such as two rows, four rows, and eight rows, has been proposed in order to capture high-definition and high-speed images over a wide range. FIG. 7 (a) shows them as Z
As viewed from the axial direction, the circle in the figure is the effective field of view FOV (Field of View). FIG. 7B is an observation of a four-row multi-slice CT including a Z-axis from a direction perpendicular to the Z-axis. When X-rays that enter the detector element from the X-ray focal point pass through the rotation center ( The thickness of the beam in the Z-axis direction (distance FCD from the X-ray focal point) is defined as a basic slice thickness T.

(7) Helical scan in multi-slice CT The helical scan in multi-slice CT is described in JP-A-4-224736. The helical pitch HP in the multi-slice CT extends the concept of the basic pitch in the single slice CT described above, and is the product of the number of detector rows N and the basic slice thickness T, HP = N × T, that is, the total slice at the rotation center It is the same as the thickness.

Hereinafter, the helical pitch is expressed by a value obtained by dividing the helical pitch by the basic slice thickness. In the case of 4-row multi-slice, a helical scan with a pitch of 4 is used.

Furthermore, the amount of bed feeding per rotation is made smaller than the total slice thickness at the center of rotation, and the detector trajectory in the previous scan and the detector trajectory in the new scan do not overlap each other. You can also scan. Such a scan is called an oversampling scan.

Next, a conventional filter interpolation method in the third generation multi-slice X-ray CT apparatus will be described.

(8) Geometry definition As shown in FIG. 7, the view angle β, the channel angle γ, the fan angle 2 × γm, and the basic slice thickness T are defined as the geometry of the third generation multi-slice X-ray CT apparatus. Note that the definitions shown in FIG. 6B are used for the helical pitch HP and the pitch P.

(9) Expression expression of sampling position of real data and counter data In single slice CT, the actual data and counter data have the relationship shown in FIG.
r (β, γ): beam of actual data at view angle β and channel angle γ zr (β): slice position of actual data at view angle β and channel angle γ
(Z coordinate)
t (β, γ): Opposing beam facing the beam with the viewing angle β and channel angle γ zt (β, γ): Opposing beam facing the beam with the viewing angle β and channel angle γ
Slice position (z coordinate)
If zs (β): Z coordinate of the X-ray focal point at the view angle β HP (= T × P): helical pitch, the sampling positions zr and zt are expressed by the following equations.

zr (β) = zs (β) = zs (0) + β / (2π) × HP
Therefore, the slice position of the actual data does not depend on the channel angle. On the other hand, the opposite data is
(Equation 1)
t (β, γ) = r (β + π + 2γ, −γ)
∴zt (β, γ) = zr (β + π + 2γ) = zs (β + π + 2γ)
= Zr (β) + (π + 2γ) / (2π) × HP
Thus, the sampling position of the opposing data is a function of the channel angle γ.

FIG. 10 is a diagram showing how a certain view angle β is sampled, where the horizontal axis represents the channel angle γ and the vertical axis represents the slice position (Z coordinate). In order to interpolate data at a certain slice position Z = Z0, interpolation is performed at an inverse ratio of distance using two data in the vicinity of the slice position Z = Z0. For example, in the case of FIG. 10, the distance dz1 between the actual data and the opposing data is
(Equation 2)
dz1 (γ) = zt (β, γ) −zr (β) = (π + 2γ) / (2π) × HP
= Dz1 (−γm) + (2γ + γm) / (2π) × HP
And is a function of the channel angle.

On the other hand, the distance dz2 between the actual data and the slice position Z = Z0 is
(Equation 3)
dz2 (β) = Z0−zr (β)
And does not depend on the channel angle. The interpolation formula is
(Equation 4)
data (Z0, β, γ)
= T (β, γ) × w (β, γ) + r (β, γ) × (1-w (β, γ))
here,
w (β, γ) = dz2 (β) / dz1 (β, γ)
It becomes. Therefore, in the two-point interpolation between the opposing data and the actual data, it is necessary to calculate dz1 for each channel.

(10) Oversampling in helical scan of multi-slice CT In the helical scan by the oversampling scan method in which the helical pitch is smaller than the total slice thickness, the data sampling pitch in the slice direction is not uniform. For example, FIG. 11 shows a scan diagram showing sampling of the center channel when a helical scan with a pitch of 2.5 is performed with 4-row multi-slice CT.

At this time, similarly to FIG. 10, when the state of sampling at a certain view angle β is represented, it can be seen that a state in which data of a plurality of detector arrays are combined in a complex manner as shown in FIG. 12 is repeated. Therefore, the data sampling state changes not only in accordance with the channel angle γ but also in accordance with the target slice position and the view angle β.

(11) Three Realization Methods of Filter Interpolation Method Next, the concept of the filter interpolation method will be described with reference to FIG. FIG. 13 is a scan diagram of a helical scan with Pitch = 2.5 in 4-row multi-slice CT. This is a range of data d (1) assumed in the vicinity of the target slice position Z = Z0 in the phase θ.
, D (2),... Are extracted and shown according to the sampling position. Here, the number N of resampling points is N = 10.

First, N resampling points are considered in a certain range near the target slice position Z0, and the resampling data V-date (i) at each resampling point is converted to each resampling point using the counter beam interpolation method. Obtained by linear interpolation of two sandwiched data d (j) and d (j + 1).

Next, the resampling data V-date (i) is converted into the normalized weight WU (i).
To determine the data date (θ) of the phase θ at the target slice position Z0.

In this method, instead of increasing the number of interpolation calculations for obtaining resampled data, the relative position between the target slice position and the sampled position of the resampled data is fixed. It is possible to Further, the spatial resolution in the slice direction can be freely changed by changing the filter shape used as shown in the lower diagram of FIG.

Three types of filter interpolation processing methods in the slice direction will be described. A case where data of a channel angle γ having a certain view angle β is filtered will be described. In the following description, for simplicity, the actual data and the opposite data are not distinguished, and d in the slice direction
N sampling data (1), d (2), d (3),..., D (k),..., D (n) are z (1), z (2), z (3), .., Z (k),..., Z (n) are obtained at slice positions (Z coordinates). This is called original data.

(A) Re-sampling method In this method, a plurality of (for example, about 10 points) equally-sampled resampling data is obtained from the original data by a two-point interpolation method, and weighted addition by a predetermined filter is performed on the resampling data. How to do it.

First, from two adjacent original data d (k) and d (k + 1) and their slice positions (Z coordinates) z (k) and z (k + 1), a re-interval between the Z coordinates of these original data. The resampling point data dr2 (j) at the sampling point zr2 (j) (z (k) ≦ zr2 (j) ≦ z (k + 1)) is
(Equation 5)
dr2 (j) = (zr2 (j) -z (k)) / (z (k + 1) -z (k))
Xd (k + 1) + (z (k + 1) -zr2 (j))
/ (Z (k + 1) −z (k)) × d (k)
= W1 (j) * d (k + 1) + (1-w1 (j)) * d (k)
here,
w1 (j) = (zr2 (j) -z (k)) / (z (k + 1) -z (k))
The re-sampling point data is weighted and added according to the following equation to obtain the target filter processing result.

(Equation 6)
data (z0) = (Σ (wa (j) × dr2 (j))) / Σ (wa (j))
FIG. 14 shows the relationship between the Z coordinate z (k) of the original data and the Z coordinate zr2 (j) of the resampling point in the resampling method described above.

(B) Direct Filter Method In this method, the original data within the filter range is weighted and added with the weight coefficient of the direct filter. Although the calculation is simple, it does not coincide mathematically with the resampling method of (a), and the filter effect depends on the Z coordinate: z (k) of the original data.

(Equation 7)
data (z0) = (Σ (wb (z (k) −z0) × OD (k)))
/ Σ (wb (z (k) −z0))
: Wb is a function of only z (k) and z0 (c) Direct filter method based on resampling method The original data is weighted and added directly by a weighting factor wc, but this wc is a resampling method whose calculation result is (a) In order to be mathematically equivalent to the above, the filter shape wa is modified according to the Z coordinate distribution of OD.

That is, the resampling data obtained by the two-point interpolation based on the original data d (k) and d (k−1) is the m1-th to m-th resampling data, and the original data d (k + 1) and d (k When the re-sampling data obtained by the two-point interpolation based on () is the n-th to n-th re-sampling data, each re-sampling data is as shown below.

(Equation 8)
dr2 (m1) = w1 (m1) * d (k) + (1-w1 (m1)) * d (k-1)
dr2 (m1 + 1) = w1 (m1 + 1) × d (k) + (1-w1 (m1 + 1))
Xd (k-1)
...
dr2 (m2) = w1 (m2) * d (k) + (1-w1 (m2)) * d (k-1)
dr2 (n1) = w1 (n1) * d (k + 1) + (1-w1 (n1)) * d (k)
dr2 (n1 + 1) = w1 (n1 + 1) × d (k + 1) + (1-w1 (n1 + 1))
Xd (k)
...
dr2 (n2) = w1 (n2) * d (k + 1) + (1-w1 (n2)) * d (k)
Therefore, in order to obtain a result equivalent to the resampling method of (a) by directly weighting the data d (k),
(Equation 9)
wc (k) = w1 (m1) × wa (m1) + w1 (m1 + 1) × wa (m1 + 1)
+ ... + w1 (m2) × wa (m2)
+ (1-w1 (n1)) × wa (n1) + (1-w1 (n1 + 1))
Xwa (n1 + 1) + .. + (1-w1 (n2)) * wa (n2)
= Σ (w1 (m1) × wa (m1)) + Σ ((1-w1 (n1))
× wa (n1))
Thus, wc (k) may be obtained for all original data d (k) and directly weighted and added using the following equation.

(Equation 10)
data (z0) = wc (k, dz) × d (k) / Σ (wc (k))
This calculation formula becomes a complicated calculation formula according to the shape of the filter (function wa).

Even if the filter shape is a simple rectangle, assuming a finite re-sampling point, each weighting factor w1 varies depending on the position of the re-sampling point, so that the complicated calculation formula remains unchanged. .

As described above, the conventional resampling method has a problem that the amount of calculation increases and it takes a long time to obtain an image with a desired filter.

In this resampling method, the distance between the actual data, which is the denominator of two-point interpolation, and the opposite data is
Since it changes depending on the channel and the view, it is necessary to execute a division every time when calculating the weighting coefficient w1, which is a factor for increasing the calculation time.

In addition, the direct filtering method requires a small amount of calculation and can perform calculation at high speed. However, the filter effect is completely different depending on the relative positional relationship between the position of the original data and the position of the data to be obtained after filtering. There was a problem of incorrect filtering.

For example, as shown in FIG. 15, in the case where the sampling pitch at which the original data is collected is uneven and the number of samplings in the region A is large, the specific gravity of the data in the region A with respect to the filter result becomes large. .

Further, the direct filtering method based on the resampling method can obtain a mathematically correct filtering result at a relatively high speed, but has a problem that the calculation formula of the weighting factor w1 becomes very complicated. Further, if the number of resampling points is increased in order not to lose sampling information in the slice direction, the calculation becomes further complicated.

In view of the above problems, an object of the present invention is to perform X-ray CT capable of executing filter processing that can execute mathematically correct filter interpolation with a simple calculation formula and that does not lose sampling information in the slice direction. Is to provide a device.

To achieve the above object, the present invention provides an X-ray beam generation source that emits an X-ray beam toward a subject and a plurality of detections that detect an X-ray beam emitted from the X-ray beam generation source. X-ray detection means having an instrument array, and bed movement means for moving a bed on which the subject is placed in the body axis direction of the subject, wherein the X-ray beam generation source is connected to the subject. A plurality of data obtained by the helical scan in an X-ray CT apparatus that scans the subject in a spiral manner by exposing the X-ray beam while rotating around and moving the subject in the body axis direction. When performing weighted addition according to a predetermined filter shape for a first data group included in a certain range in the slice direction, a weighting factor for each data in the first data group is set to the size of the range and Of that data Based on the sampling position, an X-ray CT apparatus according to subject matter that determined to be the weighting factor that reflects the unequal pitch of the respective data.

In the present invention, the weighted addition includes resampling processing for obtaining a second data group from the first data group by a first weighted addition, and the predetermined filter shape for the second data group. Assuming that the second weighted addition is performed by the filtering process, the second
Can be calculated as an infinite number of data groups arranged in an infinitely small pitch in the slice direction.

  In the present invention, the predetermined filter shape may be a rectangle.

As described above, according to the present invention, the original data composed of actual data and / or opposite data is included in a certain range in the slice direction when performing the filter interpolation process for applying a filter having a predetermined shape in the slice direction. By determining the weighting factor for each data of the first data group based on the size of the range and the sampling position of the data, mathematically correct filtering can be performed at high speed, and the data collection can be resumed from the data collection. There is an effect that the time until the constituent image is displayed is greatly shortened.

  Next, embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a system configuration diagram showing a schematic configuration of an X-ray CT apparatus according to the present invention. In the figure, an X-ray CT apparatus 10 according to the first embodiment includes a system control unit 11 and a gantry / bed control unit 1.
3, bed movement unit 15, X-ray control device 17, high voltage generation device 19, X-ray beam generation source 21, detector 23, rotating frame 25, data collection unit 27, and collected data storage Device 2
9, an image reconstruction unit 31, and a display unit 41.

The system control unit 11 includes a gantry / bed control unit that uses the rotation speed, slice thickness, fan angle, etc. as a gantry / bed control signal among helical scan conditions such as slice thickness and rotation speed input using an input device (not shown). 13 is output. In addition, the system control unit 11 outputs an X-ray beam generation control signal for controlling X-ray beam generation to the X-ray control device 17.

In addition, the system control unit 11 outputs a detection control signal indicating the detection timing of the X-ray beam to the data collection unit 27. Further, the system control unit 11 outputs a data collection control signal for data collection to the data collection unit 27. Further, the system control unit 11 outputs a control signal indicating an interpolation method, a filter shape, and the like to the image reconstruction unit 31.

The gantry / bed control section 13 rotates the gantry 25 based on the gantry / bed control signal output by the system control section 11 and outputs a bed movement signal to the bed movement section 15. The couch moving unit 15 obtains the amount of movement of the couch 15a per rotation of the rotating gantry 25 based on the couch movement signal output by the gantry / couch control unit 13, and moves the couch 15a by this moving amount.

The X-ray controller 17 controls the timing of high voltage generation by the high voltage generator 19 based on the X-ray beam generation control signal output by the system controller 11. High voltage generator 1
9 designates a high voltage for exposing the X-ray beam in accordance with a control signal from the X-ray control unit 17.
A line beam generation source 21 is supplied.

The X-ray beam generation source 21 exposes the X-ray beam with a high voltage supplied from the high voltage generator 19. The detector 23 is exposed from the X-ray beam generation source 21 and passes through the subject.
A line beam is detected.

The rotary mount 25 holds the X-ray beam generation source 21 and the detector 23. In addition, rotating platform 2
5 is rotated about a rotation axis passing through an intermediate point between the X-ray beam generation source 21 and the detector 23 by a gantry rotation mechanism (not shown).

The data collection unit 27 collects the X-ray beam (actually a detection signal) detected by the detector 23 in correspondence with the data collection control signal output by the system control unit 11.

The collected data storage device 29 stores the X-ray beam detection signal collected by the data collecting unit 27 and supplies it to the image reconstruction unit 31 as necessary.

The image reconstruction unit 31 calculates data of a desired slice position by interpolation based on the data stored in the collected data storage device 29, and reconstructs an image using this interpolation result.

For this reason, the image reconstruction unit 31 performs interpolation processing such as counter beam interpolation and filtering processing on data in the vicinity of the desired slice position to obtain data at the desired slice position, and the image reconstruction unit 31 of the desired slice position. A filter-corrected backprojection unit 35 for reconstructing an image of the data by the filter-corrected backprojection method, a memory 37 used in common for the interpolation processing unit 33 and the filter-corrected backprojection unit 35, and an information storage unit 39 for storing parameters And.

The display unit 41 displays the image reconstructed by the image reconstruction unit 31 on a monitor (not shown).

  Next, a first embodiment of the filter interpolation processing operation in the X-ray CT apparatus 10 will be described.

In the first embodiment, the shape of the filter is a rectangular filter in which the weighting coefficient in the section included in the filter width FW is uniform and the weighting coefficient in the section not included in the filter width FW is zero. Then, assuming innumerable resampling points with an infinitesimal sampling pitch with respect to the original data, the weighted addition (Σ) is converted into integral (∫) and calculated.

After data collection, helical interpolation and image reconstruction are performed according to the following procedures (1) to (5).

(1) The interpolation processing unit 33 inputs information such as the target slice position Zc, the filter shape (rectangular) used for the filtering process, and the filter width (FW) from the system control unit 11.

(2) The interpolation processing unit 33 includes a plurality of pieces of original data necessary for performing the filter processing with the filter width FW at the target slice position Zc, sampling position (Z coordinate) information of d1, d2,. .., z9 are determined, and data addresses adr1, adr2,..., adr9 whose data are stored in the collected data storage device 29 are generated. The sampling position information may be input from the information storage unit 39 such as a table. Although the number of data has been described as nine, it is not limited to this.

(3) The interpolation processing unit 33 reads data according to the data addresses adr1, adr2, ..., adr9. When the opposite data needs to be generated, the necessary actual data is read and the opposite data is generated by a predetermined weighted addition.

(4) The interpolation processing unit 33 performs interpolation data data (
Zc) is obtained, and the interpolation data is passed to the filter-corrected backprojector 35.

(5) The filter-corrected backprojection unit 35 performs image reconstruction by the filter-corrected backprojection method and outputs the image to the display unit 41.

Next, details of the interpolation process (4) will be described.
Sampling as shown in FIG. 2 is obtained in the slice direction. Data d1
˜d9, the data d1 and d9 outside the both ends of the filter width FW, and the filter width F
Data d2 and d8 closest to the boundary inside W and the data d3 to d7 inside each will be described.

(A) In the case of d3 to d7 For example, the weighting factor w3 for the third data d3 is between case-A where the virtual resampling point is between Z2 and Z3, and between Z3 and Z4. Case-B is divided into two cases. Depending on the variables t and s, the weighting factor for case-A two-point interpolation is t /
In the case of (Z3-Z2) and case-B, it is represented by [1-s / (Z4-Z3)]. Therefore, w3 is obtained by integrating each corresponding section and normalizing the result of integration by dividing by the filter width FW. This is shown in equation (1).

Similarly, since the weighting factors w4 to w7 for the data from d4 to d7 can be considered by shifting the reference points (zero points) of the variables t and s, the equation (2) is obtained.

(B) In the case of d1 and d9 Similarly to the above, the integration is performed for the range in which the data is related, and the expression (3) is obtained.

(C) In the case of d2 and d8 Similarly to the above, the integration is performed for the range in which the data is related, and Expression (4) is obtained.

Since the weighting factors w1, w2,..., W9 obtained here have already been normalized, the total sum is 1 as shown in the following equation (5).

Using the weighting coefficients obtained in Expression (1) to Expression (4), the interpolation data d (Zc) at the slice position Z = Zc in the filter processing when the filter shape is rectangular is expressed by the following Expression (6). Obtained by.

This is repeated for a necessary view angle (for example, 360 ° or 180 ° + 2γm) to obtain interpolation data. Using these interpolation data, the filtered back projection unit 3
5 performs image reconstruction by the filter-corrected back projection method, and outputs it to the display unit 41.

In the calculation of the above interpolation data, in the present invention, an infinite number of resampling points arranged at an infinitesimal pitch are assumed, and the weighting coefficient of the direct filter processing is determined by the integral formula. Therefore, the filter shape is mathematically correctly reflected. Interpolation can be performed at high speed, and a high-quality reconstructed image can be obtained in a short time.

In the first embodiment, the resampling method is assumed to be a two-point interpolation method. However, the resampling method is not limited to this, and may be a non-linear interpolation or a linear or non-linear interpolation of three or more points.

Further, the number of detector rows for detecting the X-ray beam is not limited to four, and may be one or any number of rows. Further, it is obvious that the present invention can be applied to a fourth generation X-ray CT apparatus and a fifth generation X-ray CT apparatus.

Next, the weighting coefficient when the number of data included in the filter width FW is small as shown in FIG. 3 will be described.

FIG. 3A shows a case where the filter width FW includes only the Z coordinates of the two data d2 and d3. At this time, d1 and d4 may be calculated according to equation (3) in the case of (b), and d2 and d3 may be calculated according to equation (4) in the case of (c).

FIG. 3B shows a case where only the Z coordinate of only one data d2 is included in the filter width FW. At this time, assuming that d1 and d3 are in the case of (b), according to the equation (3), d2 is the following equation (7
)by.

(D) In the case of d2

FIG. 3C shows a case where the Z coordinate of one data is not included in the filter width FW. In other words, the filter width FW is between the two data d1 and d2. At this time, it is the same as simple two-point interpolation, and is expressed by equations (8) and (9).

(E) Weight coefficient for d1 and d2

Next, a description will be given of a second embodiment in which the filter shape is symmetrical to a general shape other than a rectangle. Since only the interpolation reconstruction process (4) is different from the first embodiment, the overlapping description is omitted.

In the second embodiment, when the filter shape is an isosceles triangle shown in FIG. 4 as an example, the filter shape is expressed by the following equation (10), where dz is the displacement with respect to the center Z coordinate (Zc) of the filter. It is expressed by

When this filter is normalized and its area is 1, the filter width FW and the filter height A satisfy the relationship of the following equation (11).

Here, when a value expressed by the filter width FW is substituted for A, Equation (10) becomes Equation (12).

When data is sampled as shown in FIG. 4, the weighting factor for each data is
(A) In the case of d3, d4, and d5, the following equations (13) to (15) are obtained.

Although detailed equation modification is omitted, if the weighting factor is stored in the form of calculating the integral as shown in equation (14), high-speed filter interpolation processing can be performed. weighting factor w for dj
Similarly, the general formula (15) indicating j is also a function of the sampling position of each data.

Further, when the filter shape is defined by being normalized by Uflt (dz), generally, the weighting factor may be obtained by the following equations (16) to (20). For example, 9
When the original data is related to the filtering process, it is as follows.

(A) In the case of d3 to d7

(B) In the case of d1 and d2

(C) For d2 and d8

  The processing after obtaining the weighting coefficient by the above equations is the same as in the first embodiment.

That is, interpolation data is repeatedly obtained for a range of a necessary view angle (for example, 360 ° or 180 ° + 2γm), and the filter correction backprojection unit 35 uses these interpolation data by the filter correction backprojection method. Image reconstruction is performed and output to the display unit 41.

FIG. 5 is a diagram conceptually comparing the conventional resampling method and the filter interpolation method using the X-ray CT apparatus of the present invention. In the present invention, the filter having a fixed shape Wt used in the resampling method is converted into a variable filter called W (N) reflecting the non-uniform pitch of the original data, and weighted addition is performed.

1 is a system configuration diagram showing a configuration of an X-ray CT apparatus according to the present invention. It is a figure explaining the parameter a, b, s, t, u while explaining the positional relationship on the Z-axis of the filter of the filter width FW and several data d1d9. It is a figure explaining the case where the number of data contained in filter width FW is small. It is a figure explaining embodiment in case a filter shape is a triangular filter. It is the figure which contrasted the resampling method and the filter interpolation method by the X-ray CT apparatus of this invention. It is a figure explaining each data collection method of conventional scan (a) and helical scan (b). It is a figure which shows the geometry of a 3rd generation multi-slice X-ray CT apparatus. It is a figure explaining the relationship between each data collection of conventional scan (a) and helical scan (b), and the slice position for the reconstruction objective. It is a figure explaining a 360 degree interpolation method and a counter beam interpolation method. It is the figure which showed the mode of sampling of a certain view angle (beta) on the sinogram which took the channel angle (gamma) on the horizontal axis and slice position Z on the vertical axis | shaft. It is a figure which shows the mode of the high-density sampling of the pitch 2.5 by a 4 row multi-slice X-ray CT apparatus. It is a figure which shows the mode of sampling of the view angle (beta) in the high-density sampling of the pitch 2.5 by a 4 row multi-slice X-ray CT apparatus. It is a figure explaining the concept of the filter interpolation method by a data and weighting process. It is a figure explaining the data sampling of the slice direction (Z-axis direction) in view angle β and channel angle γ. It is a figure explaining the problem in case a sampling pitch is uneven.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 ... X-ray CT apparatus, 11 ... System control part, 13 ... Stand, bed control part, 15 ... Bed movement part, 17 ... X-ray control apparatus, 19 ... High voltage generator, 21 ... X-ray beam generation source, 23 DESCRIPTION OF SYMBOLS ... Detector 25 ... Rotary mount 27 ... Data collection part 27, 29 ... Collected data storage device, 31 ... Image reconstruction part, 33 ... Interpolation processing part, 35 ... Filter correction back projection part, 37 ... Memory, 39 ... Information storage unit, 41... Display unit.

Claims (3)

An X-ray beam generation source that emits an X-ray beam toward the subject;
X having a plurality of detector rows for detecting an X-ray beam exposed from the X-ray beam generation source
Line detection means;
A bed moving means for moving the bed on which the subject is placed in the body axis direction of the subject,
In the X-ray CT apparatus for rotating the X-ray beam generation source around the subject, exposing the X-ray beam, moving the subject in the body axis direction, and scanning the subject in a spiral shape,
Among the plurality of data obtained by the helical scan, when performing weighted addition according to a predetermined filter shape for the first data group included in a certain range in the slice direction,
The weighting factor for each data of the first data group is determined based on the size of the range and the sampling position of the data so as to be a weighting factor reflecting an unequal pitch of each data. X-ray CT system. The weighted addition includes a resampling process for obtaining a second data group from the first data group by a first weighted addition, and a second weighted addition for the second data group by the predetermined filter shape. 2. The X-ray according to claim 1, wherein the weighting coefficient of the weighted addition is calculated as an innumerable data group in which the second data group is arranged at an infinitesimal pitch in the slice direction, assuming a filtering process to be performed. CT device. The X-ray CT apparatus according to claim 1, wherein the predetermined filter shape is a rectangle.