JPH067342A - Computational x-ray tomography system - Google Patents

Abstract

PURPOSE:To obtain images at continuous slice positions at high speed by calculating a tomographic image at one slice position by preserving projection data for two times of rotation, and calculating the tomographic image at the adjacent slice position based on the difference of projection data between both of slice positions. CONSTITUTION:In a third generation X-ray CT system of a helical scan type helically scanning an examinee body by continuously moving a bed 22 while rotating an X-ray tube 20 and an X-ray detector 24 around the bed 22, the output of the X-ray detector 24 is inputted to a reconstitution circuit 28 through a reprocessing circuit 26 for performing amplification, A/D conversion and various correction or the like. In this reconstitution circuit 28, the tomographic image at one slice position is calculated with normal reconstitution processing by interpolating the projection data for two times of rotation, the tomographic image at the adjacent slice position is calculated by reconstituting the difference of projection data between both of slice positions, multiplying a coefficient to the difference and adding them, and the result is displayed on a display device 30.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an X-ray computed tomography apparatus (hereinafter referred to as an X-ray CT apparatus), and particularly to an X-ray of a third generation which employs a helical scan method.
It relates to image reconstruction of a line CT device.

[0002]

2. Description of the Related Art FIG. 14 shows a schematic block diagram of a processing circuit relating to image reconstruction calculation of a conventional X-ray CT apparatus. Projection data from an X-ray detector (not shown) is supplied to the preprocessing circuit 2, and amplification, A / D conversion and the like are performed. A raw data storage device 4 and a central processing unit (hereinafter referred to as CPU) 6 are connected to the output of the preprocessing circuit 2. The output of the preprocessing circuit 2 is the convolution operation circuit (hereinafter, CO
It is stored in the image memory 12 as a reconstructed image through an NV circuit 8) and a back projection arithmetic circuit (hereinafter referred to as a BP circuit) 10. The image data in the image memory 12 is supplied to the contrast scale circuit 14 to be a CT image. The CT image is also stored in the image memory 12, and the CT image is supplied to the display device (not shown). Therefore, the number of the image memory 12 may be one, but when the reconstructed image and the CT image are stored in different memories, the number is two.
Provide one.

A CPU 16 is connected to the CONV circuit 8, the BP circuit 10, and the contrast scale circuit 14, and
The circuit 10 includes a constant setting unit 1 for setting various constants for back projection calculation under the control of the CPU 16.
8 is also connected. Although not shown, various buffers are provided in the CONV circuit 8, the BP circuit 10 and the like in order to perform the reconstruction operation at high speed.

In such a conventional example, basically, data concerning projection of 360 ° or 180 ° is calculated as one group (unit). Image reconstruction is performed independently for each slice position from a group of projection data of 360 ° or 180 °. for that reason,
When reconstructing images at a large number of slice positions, the reconstruction time must be extended in proportion to the number of images. Further, in consideration of storing the reconstructed image, when reconstructing the images at many slice positions one by one, there is a problem that a large amount of storage area is required.

Therefore, a method for efficiently reconstructing images at a large number of consecutive slice positions has been considered. An example of this is the method described in JP-A-58-32748 (US Pat. No. 4,495,645).

However, this conventional continuous slice reconstruction method relates to a continuous scan method at the same slice position, and a method for efficiently reconstructing images at consecutive slice positions in the helical scan method has not been proposed. .

[0007]

SUMMARY OF THE INVENTION The present invention has been made to address the above-mentioned circumstances, and its purpose is to efficiently obtain images of a large number of consecutive slice positions in an X-ray CT apparatus adopting a helical scan method. It is to reconfigure.

Another object of the present invention is to provide an X-ray CT apparatus adopting a helical scan system, which is capable of continuously observing the inside of the body by continuously displaying images of a large number of consecutive slice positions. Is to provide.

Another object of the present invention is to reduce the number of images to be stored by reconstructing and displaying the images at a large number of consecutive slice positions in an X-ray CT apparatus adopting the helical scan method. It is to provide an X-ray CT apparatus which can be used.

[0010]

An X-ray CT apparatus for collecting projection data by helical scanning according to the present invention has a predetermined projection data group for a predetermined projection angle before and after a first slice position. Means for performing a reconstruction calculation to obtain a tomographic image at the first slice position, and means necessary for reconstructing a tomographic image at a second slice position adjacent to the first slice position from the first projection data group Means for extracting a first projection data part that is not common to the second projection data group, and a second projection data part having the same projection angle as the first projection data part from the second projection data group Means to do
A means for performing a predetermined reconstruction operation on the first and second projection data portions and synthesizing the reconstruction result and the tomographic image at the first slice position to obtain a tomographic image at the second slice position. It is characterized by

[0011]

According to the X-ray CT apparatus of the present invention, images at successive slice positions can be obtained at high speed in helical scanning. Therefore, by continuously changing the slice position according to the operation of the command input device such as a mouse, the inside of the body can be continuously observed, and useful information for diagnosis is provided.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT An X-ray CT according to the present invention will be described below with reference to the drawings.
A first embodiment of the apparatus will be described. First, the principle of image reconstruction according to the present invention will be described. Hereinafter, the projection data according to the present invention is projection data collected by the helical scan method. If the image at a certain slice position is the m-th image (m-th image) mI, this m-th image is sliced by linearly interpolating the projection data in the projection angle range of 720 ° before and after the slice position. It can be obtained by obtaining the projection data of the position and reconstructing it.
This reconstruction may be the same as the normal helical scan reconstruction processing.

An image at a slice position adjacent to this slice position is obtained by the following processing unique to the present invention, which is different from the normal processing described above. Here, it is assumed that the slice position number increases as the scan progresses. The image is reconstructed without interpolating the projection data of 360 ° collected in the temporal direction before the slice position of the m-th image (direction in which the slice position number decreases), and this image is set as the m-th B image. Similarly, the image is reconstructed without interpolating the projection data of 360 ° collected temporally after the slice position of the m-th image (direction in which the slice position number increases), and this image is converted into the m-th image.
Let it be an F image. In addition, the distance that the bed moves while the X-ray tube and the X-ray detector rotate by one projection angle in order to obtain the projection data of 360 ° is d, and the distance d from the slice position of the m-th image The image at the subsequent slice position is the (m + 1) th image. Hereinafter, in order to simplify the description, it is assumed that the slice position and the projection position (bed position) match. The following images are obtained to obtain the (m + 1) th image. (M + 1) BI = mBI + BP [CONV (a-b)] (1) (m + 1) FI = mFI + BP [CONV (c-a)] (2) Here, BP: Back projection operation CONV: Convolution operation a : Projection data at a slice position of the (m + 1) th image b: Projection data at a slice position that is 360 ° before the a c: Projection data at a slice position that is 360 ° after the a In the present invention, the (m + 1) th image is obtained as follows using the equations (1) and (2). (M + 1) I = mI + DW × {(m + 1) FI− (m + 1) BI} (3) where DW is the slope of linear interpolation. Similarly, the following images are obtained to obtain the (m-1) th image. (M-1) BI = mBI + BP [CONV (e-d)] (4) (m-1) FI = mFI + BP [CONV (d-f)] (5) where d: the (m-1) ) Projection data at slice position of image: projection data at slice position 360 ° before e: d f: projection data at slice position 360 ° after f: d. In the present invention, the formula (4) and the formula (5) are used to calculate the (m-1) th
Obtain the image as follows. (M−1) I = mI−DW × (mFI−mBI) (6)

Here, the projection data a, b, c have the same projection angle, and the projection data d, e, f also have the same projection angle. Therefore, if mFI-mBI = mFBI is defined, (1)
The following equation is obtained from the equation (2). (M + 1) FBI = (m + 1) FI- (m + 1) BI

= MFI + BP [CONV (c-a)]-mBI-BP [CONV (a-b)] = mFI-mBI + BP [CONV (c-a)]-BP [CONV (a-b)] = mFBI + BP [ CONV (c + b-2a)] (7) Similarly, the following equation is obtained from the equations (4) and (5). (M-1) FBI = (m-1) FI- (m-1) BI

= MFI + BP [CONV (d-f)]-mBI-BP [CONV (e-d)] = mFI-mBI + BP [CONV (d-f)]-BP [CONV (e-d)] = mFBI- BP [CONV (2d-fe)] (8) Substituting the equations (7) and (8) into the equations (3) and (6), the images of continuous rice positions are reconstructed by the following equations. It turns out that configuration is possible. (M + 1) I = mI + DW × (m + 1) FBI (9) (m-1) I = mI-DW × mFBI (10)

In this way, a tomographic image can be obtained at each slice position by only reconstructing only the difference in projection data from the slice position at which the tomographic image has already been obtained by reconstruction. Since it is not necessary to interpolate and reconstruct the projection data of 360 °, it is possible to quickly obtain images of consecutive slice positions in the helical scan with a small amount of calculation, and use the mouse or the like to determine the slice position of the m-th image. If changed continuously, the images at each slice position can be continuously displayed, and the inside of the body can be continuously observed.

An embodiment of the present invention based on such a principle will be described. FIG. 1 is a block diagram showing the schematic arrangement of the X-ray CT apparatus according to the first embodiment. Here, the third generation X-ray CT apparatus will be described as an example, but the present invention can also be applied to other generation apparatuses. In the third generation, imaging is performed while the X-ray tube 20 and the X-ray detector 24 rotate around the bed 22. In the third-generation CT apparatus, there are many methods in which the rotation direction is reversed in the clockwise direction and the counterclockwise direction for each scan, but recently, there is a method in which images are continuously rotated in the same direction. Normally, the X-ray tube 20 and the X-ray detector 24 are 360
° While rotating, the bed 22 is stopped and a picture is taken. However, in the present invention, an imaging method called helical scan in which the subject is spirally scanned by the bed 22 continuously moving in synchronization with the rotation of the X-ray tube 22 and the X-ray detector 24 is adopted. There is. According to this, a multi-slice tomographic image can be captured at high speed.

The preprocessing circuit 2 in which the output of the X-ray detector 24 performs amplification, A / D conversion, various corrections and the like on the detection signal.
It is input to the reconstruction circuit 28 via 6. Reconstruction circuit 2
An input device 32 including a key switch and a mouse is also connected to 8. The output of the reconstruction circuit 28 is displayed on the display device 30.

FIG. 2 is a detailed block diagram of the reconstruction circuit 28 of FIG. The raw data storage device 40 and the CPU 42 are connected to the output of the preprocessing circuit 26. Preprocessing circuit 26
Output from the CONV circuit 4 via the interpolation (difference) circuit 44.
6 is supplied. The interpolation circuit 44 sets the interpolation constants to -1, +1.
A differential circuit is obtained by setting CONV circuit 4
6 outputs the FB image memory 50 via the BP circuit 48,
It is supplied to the reconstructed image memory 54. The definition of the FB image is as described above. Difference circuit 44, CONV
The CPU 56 is also connected to the circuit 46 and the BP circuit 48.
Further, a constant setting unit 58 is connected to the BP circuit 48.
The outputs of the FB image memory 50 and the reconstructed image memory 54 are combined by the image composition circuit 60, and the composition result is supplied to the reconstructed image memory 54 again. The output of the reconstructed image memory 54 is supplied to the CT image memory 64 via the contrast scale circuit 62. The output of the CT image memory 64 is displayed on the display device 30. Input device 3 for CPU 56
2 are connected. Table 1 shows a set of data collected by the helical scan.

[0021]

[Table 1] Further, this data is schematically shown in FIG.

Here, d is the X-ray tube 20 and the X-ray detector 24.
Is the distance that the bed 22 moves while is rotated by one projection angle. Although the bed position and the projection angle are strictly continuous, they are represented by the central value of the data collection time.

Table 1 shows an example in which projection is performed for each rotation of the X-ray tube 20 and the X-ray detector 24. Therefore,
In one rotation, 360 sets of projection data can be obtained. Meanwhile, the bed moves by D = 360d. This is photographed 20 times in succession, for example. In that case, the scan length is 2
0D, the number of projections obtained is 7200.

The helical scan can scan a multi-slice tomographic image at high speed, but cannot obtain a mathematically strict tomographic image. This is because in order to reconstruct a tomographic image from a projected image, a projected image of 180 ° or more is required on the same tomographic plane, but in the case of helical scanning, the projected tomographic plane moves continuously. This is because it is not possible to obtain a projected image of 180 ° or more on the same tomographic plane. However, since the projected image has a constant slice thickness, a practically sufficient tomographic image can be obtained unless the bed moves quickly.

However, in order to reconstruct a tomographic image in the helical scan, it is necessary to artificially calculate a projection image lacking at the reconstruction position from projection images at other slice positions. There are two typical methods for this. The first is a method of interpolating two projection data of the same projection angle projected at the position closest to the slice position. The second is a method of creating reflection data by the so-called reflection method and performing interpolation in the same manner. The present invention provides a method for efficiently performing the first method when reconstructing continuous images.

The reconstruction method according to this embodiment will be described below with reference to Table 1 and FIG. Here, as an example, the case of reconstructing the second image of image number 2 will be described. The projection data position of the second image is 361d in the position of the bed. Since the projection data 361 is the projection at this position, the projection data for the angle 1 exists, but the projection data for the other angles does not exist. This nonexistent projection data has the same angle and is obtained by interpolation from two projection data whose slice positions are close to each other. For example, the interpolation projection data for the angle 2 can be obtained by interpolation from the projection data 2 and the projection data 362. Interpolation is often performed by primary interpolation of the distance between the position of projection data and the slice position of the image. Therefore, the angle 2 interpolation projection data for the second image is obtained as follows. 2PP ₂ = 2P × WB ₂ + 362P × WF ₂ (11) Here, 2PP ₂ : projection data of angle 2 for the second image 2P: projection data 2 362P: projection data 362 WB ₂ , WF ₂ : An interpolation coefficient. The interpolation coefficient is expressed as follows. WB ₂ = d / 360d = 1/360 (12) WF ₂ = (360d-d) / 360d = 359/360 (13) where m: image number n: interpolation projection data number (1 to 360).

Note that n is projection data at the same angle as the projection data at the slice position of the image, n = 1, that is, the first interpolation projection data. For example, in the case of image 2, the interpolation projection data of angle 1 becomes the first interpolation projection data, and in the case of image 3, the interpolation projection data of angle 2 becomes the first interpolation projection data. Based on that, n also increases in the direction in which the number of the projection data increases. Therefore, for example, in the case of the image 2, the interpolation projection data of the angle 2 is obtained.
If the data is the image 3, the interpolation projection data of the angle 3 is n
= 2, that is, the second interpolation projection data. mPP _n: n-th interpolation projection data of the m image - data Pj: the j projection de - data (where, Po = 0) Pk: k-th projection de - data (however, in the case of k> the number of projections Pk = 0 ) j = m + 358 + n -360 ... (14) k = j + 360 ... When (15), the _{mPP n = Pj × WB n +} Pk × WF n ... (16) ( provided that, n = 1 to 360). Where WB _n ,
WF _n is the interpolation coefficient of the n-th interpolation projection data, and these are expressed as follows. WB _n = _(n-1) is a _{/ 360 ... (17) WF n} = (361-n) / 360 ... (18). Here, CONV: convolution BP: back projection mI: m-th image MQQ: If the set of mPP _{n (n} = 1~360), the m image mI can be determined as follows. mI = BP [CONV (mQQ)] (19) However, convolution and back projection are performed for each projection angle. Next, in order to explain the reconstruction of continuous slice images according to the present invention, B image, F image, F
The image named B image is defined by: mBI = BP [CONV (mBPJ)] (20) mFI = BP [CONV (mFPK)] (21) mFBI = mFI-mBI (22) where mFBI: FB image relating to mth image mBI: mth B image for image mFI: F image for mth image mBPj: Set of Pj (j = (m-1) to (m + 35)
8)) mFPk: A set of Pk (k = j + 360). By substituting the expressions (20) and (21) into the expression (22), the following expressions are obtained. mFBI = BP [CONV (mFPk)]-BP [CONV (mBPj)] = BP [CONV (mFPk)] + BP [CONV {(-1) * mBPj}] (23)

Next, the mth FB image mFBI and the mth FB image
The Bth image mBI, the mth F image mFI to the (m + 1) th FB image (m + 1) FBI, the (m +) th
A method for obtaining the 1) th B image (m + 1) BI and the (m + 1) th F image (m + 1) FI will be described.

MBI, mFI and (m + 1) BI, (m
+1) FI is an image reconstructed from projection data of 360 °, and only one projection is different. Therefore, according to the method of the above-mentioned Japanese Patent Laid-Open No. 58-32748, the following can be obtained. (M + 1) BI = mBI + BP [CONV (SBP _{m + 1} )] (24) (m + 1) FI = mFI + BP [CONV (SFP _{m + 1} )] (25) where SBP _{m + 1} = P (m + 359) -P (m-1) ... (26) SFPm _{+ 1} = P (m + 719) -P (m + 359) ... (27). Similarly, (m-1) BI and (m-1) FI are obtained from mBI and mFI as follows. (M-1) BI = mBI + BP [CONV (SMBP _m-1 )] (28) (m-1) FI = mFI + BP [CONV (SMFP _m-1 )] (29) Here, SMBP _m-1 = P (m−2) −P (m + 358) (30) SMFP _m−1 = P (m + 358) −P (m + 718) (31) Therefore, the following equation is obtained from the equation (22). (M + 1) FBI = (m + 1) FI- (m + 1) BI = mFI + BP [CONV (SFP _{m + 1} )]-mBI-BP [CONV (SBP _{m + 1} )]

= MFBI + BP [CONV (SFP _{m + 1} )] − BP [CONV (SBP _{m + 1} )] (32) where P (m−2), P (m + 358), P (m + 7)
Since the projection angle of 18) is the same, equation (32) can be rewritten as follows. (M + 1) FBI = mFBI + BP [CONV (SFP m -SBP m + 1)] = mFBI + BP [CONV {P (m + 719) + P (m-1) -2 × P (m + 359)}] ... (33) Similarly, ( m-1) FBI = mFBI + BP [CONV {2 * P (m + 358) -P (m + 718) -P (m-2)}] ... (34) is obtained.

Next, the procedure of this embodiment will be described in detail with reference to the flow chart shown in FIG. The following steps are controlled by the programs of the CPU 42 and the CPU 56. In step # 10, projection data is collected by helical scanning. In step # 12, the image number (slice position) m to be displayed as the current image is appropriately determined and input from the input device 32.

In step # 14, the current image (mth image) is reconstructed and displayed by the equation (19). Step # 1
The current FB image (m-th FB image) is calculated by the equation (23) in 6
Reconfigure. A value required to generate constants (reconstructed image, FB image) necessary for back projection,
Also, the convolution function and other constants necessary for the reconstruction are transferred from the CPU 56 to the interpolation circuit 44 and the CONV circuit 4.
6, transferred to the BP circuit 48 and stored.

At step # 20, it is determined whether or not the direction indicating switch of the input device 32 is operated. That is, it is determined whether or not other images adjacent to the m-th image are continuously reconstructed. If the direction switch has not been operated, the process waits in step # 20. Step #
If it is determined in step 22 that the direction indicating switch has been operated, in step # 24, the images of consecutive slice positions are displayed in the direction of continuous reconstruction, that is, in the order of increasing image number m (ascending order), or It is determined whether or not images at consecutive slice positions are displayed in the decreasing order (decreasing order) of the image number m. When displaying images at consecutive slice positions in the ascending direction, the process proceeds to step # 24, and when displaying images at consecutive slice positions in the descending direction, the process proceeds to step # 34.

At step # 24, the (m + 1) th FB image is reconstructed by the equation (33). In step # 28, the (m + 1) th image at the slice position (m + 1) is reconstructed by combining mI and (m + 1) FBI as described below. (M + 1) I = mI + DW × (m + 1) FBI (35) Here, DW is a coefficient of linear interpolation and is expressed as follows. _{DW = WB n + 1 -WB n} = - (WF n + 1 -WF n) = 1/360 ... (36)

In step # 30, m ← m + 1 is set, and in step # 32, the CT image obtained by performing the contrast scale conversion on the (m + 1) th image obtained by the equation (35) is displayed, and the process returns to step # 20. In this way, by sequentially increasing m, it is possible to continuously observe images whose slice positions are sequentially shifted in the forward direction. On the other hand, in step # 34, the (m-th) slice at the slice position (m-1).
1) The image is reconstructed by combining mI and mFBI as follows. (M−1) I = mI−DW × mFBI (37)

In step # 36, the (m-th)
-1) Reconstruct the FB image. In step # 40, m ← m
-1, the CT image obtained by subjecting the (m-1) th image obtained by the equation (37) to the contrast scale conversion is displayed in step # 42, and the process returns to step # 20. In this way, by sequentially reducing m, images whose slice positions are sequentially shifted in the backward direction can be continuously observed. Next, the formula (37) is proved. Formula (19): mI =
In BP [CONV (mQQ)], if m ← m−1, the following equation is obtained. (M-1) I = BP [CONV {(m-1) QQ}] (38) From equations (19) and (38), mI- (m-1) I = BP [CONV {(m- 1), mRR}] ... (39)

Is obtained. Where (m-1) and mRR are
It is the difference set between the set mQQ and the set (m-1) QQ. m
QQ is mPP_nSince it is a set of (n = 1 to 360),
It is expressed as. Similarly, the set (m-1) QQ is expressed as follows.Therefore, the difference set (m-1), mRR is expressed as
Be done. From formula (36), WB_n-WB_{n + 1} = WF_n-WF_{n + 1} = -DW, we substitute this into the above set and
When the projection data are summarized, the difference set (m-1), mRR is
It looks like this:

Here, (WF ₁ −WB ₃₆₀ ) = 1/36
Since 0 = DW, WB ₁ = 0, and WF ₃₆₀ = 1/360 = DW, the difference sets (m−1) and mRR are as follows when these are substituted into the above set and rearranged. −DW {P (m−1), ... P (m + 358)} + DW {P (m + 359), ... P (m + 718)} = − DW × mBPj + DW × mFPk (40) The equation (40) is converted into the equation (39). Substituting gives the following equation: mI- (m-1) I = BP [CONV {(m-1), mRR}] = BP [CONV (-DW * mBPj + DW * mFPk)] = DW * BP [CONV (mFPk)]-DW * BP [ CONV (mBPj)] = DW × [mFI−mBI] = DW × mFBI (41) When this is modified, (m−1) I = mI−DW × mFB
It becomes I. Therefore, the equation (37) is obtained. Equation (35) can be similarly obtained by transforming equation (37) as follows. mI = (m−1) I + DW × mFBI (37 ′) In equation (37 ′), if m ← m + 1, then (35)
The formula is obtained. The details of each step in FIG. 4 will be described below.

Details of the reconstruction of the m-th image in step # 14 of FIG. 4 will be described with reference to the flowchart shown in FIG. The number of interpolated projection images required for reconstructing the reconstructed image is RP
JN. In step # 141, the CPU 56 notifies the interpolation circuit 44, the CONV circuit 46, and the BP circuit 48 that the reconstruction is the first reconstructed image and the CP.
U56 initializes the reconstructed image memory 54 and sets the reconstructed image data to zero.

In step # 142, n = 1, in step # 143 j = m + 358 + n-360, and in step # 144 k = j + 360. In step # 145, the CPU 42 reads the jth projection data Pj (P0 = 0) from the raw data storage device 40 and transfers it to the interpolation circuit 44.
In step # 146, the CPU 42 causes the raw data storage device 40.
The k-th projection data Pk (here, if k> total projection number, Pk = 0) is read out and transferred to the interpolation circuit 44. In step # 147, the nth interpolation projection data PP of the mth image is obtained as follows. _{PP = Pj × WB n + Pk} × WF n ... (42) where, WB _n, WF _n n-th interpolation projection de - an interpolation coefficient data.

At step # 148, the CONV circuit 46 performs a convolution operation on the output PP of the difference circuit 44 and the convolution function. In step # 149, the BP circuit 48 back-projects (adds) the convolution result PPCV to the image mI of the reconstructed image memory 54.
To do. The back projection result is the reconstruction memory 54.
Stored in.

In step # 150, n ← n + 1, j ← j +
1, k ← k + 1, and in step # 151, n ≦ RPJN
It is determined whether or not. If n ≦ RPJN, step # 1
Returning to 45, if n> RPJN, next step # 15
Go to 2. At this time, the reconstructed image is stored in the reconstructed image memory 54.

In step # 152, the contrast scale circuit 62 performs a contrast scale conversion for every pixel of all the pixels of the reconstructed image, creates a CT image, and stores it in the CT image memory 64. The contrast scale conversion is a known technique, and is generally performed by the following formula. PV = a × RPV + b (43) where PV: pixel value of CT image RPV: pixel value of reconstructed image a, b: constants, where a and b are generally 0 for water and −1000 for air. Is determined according to the device and the X-ray tube voltage. In step # 153, the CT image is transferred to the image display device 30, and the image is displayed. As a result, the current image (mth image) is displayed. Details of step # 16 for reconstructing the mFB image in FIG. 4 will be described with reference to the flowchart shown in FIG.

First, in step # 162, the CPU 56 informs each of the circuits 44, 46 and 48 that the reconstruction of the first FB image is made, and the FB image memory 50 is made.
Is initialized and the FB image data is set to 0.

In step # 164, n = 1, in step # 166 k = m-359, and in step # 168 j.
= M-1. In step # 170, the CPU 42 reads the kth projection data Pk from the raw data storage device 40 and transfers it to the interpolation circuit 44. Here, interpolation coefficients 1 and 0 are set in the interpolation circuit 44, and step # 172
Then, the interpolation circuit 44 does nothing and simply passes the data and outputs the projection data Pk as PP.

At step # 174, the CONV circuit 46 performs a convolution operation on the output PP of the interpolation circuit 44 and the convolution function. In step # 176, the BP circuit 48 back-projects (adds) the convolution result PPCV to the image in the FB image memory 50.
Back projection result FBI is FB image memory 5
Stored in 0.

In step # 178, the CPU 42 reads the j-th projection data Pj from the raw data memory 40 and sends it to the interpolation circuit 44. In step # 180, the interpolation coefficient is set to -1 and 0, and the interpolation circuit multiplies the projection data Pj by -1 and outputs it as PP.

At step # 182, the CONV circuit 46 convolves PP and outputs it as PPCV. In step # 184, PPCV is backprojected to the FB image and stored in the FB image memory 50.

In step # 186, n ← n + 1, in step # 188 k ← k + 1, and in step # 190, j
← J + 1. In step # 192, it is determined whether n ≦ RPJN. If n ≦ RPJN, the process returns to step # 170. If n> RPJN, the process ends. At this time, the current (m-th) FB image mFBI is reconstructed and FB
It is stored in the image memory 50.

The (m + 1) th FB image (m +
1) A specific flowchart of step # 24 for reconfiguring the FBI is shown in FIG. Here, mI, mFBI
Are the reconstructed image memory 54 and the BF image memory 5, respectively.
It is assumed to be stored in 0.

Although not shown in the drawing, first, information that reconstruction of a continuous image in the direction in which m increases is performed, slope DW of linear interpolation, values necessary for generating constants necessary for back projection of an FB image, And the convolution function and other necessary information from the CPU 56 to the interpolation circuit 4
4, transferred to the CONV circuit 46 and the BP circuit 48 and stored therein. It is not necessary to transfer the information already stored in the reconfiguration arithmetic unit.

In step # 242, the CPU 42 sends the (m + 359) th projection data and the (mth) projection data from the raw data storage device 40.
−1) Three pieces of projection data of the projection data and the (m + 719) th projection data are read out and supplied to the interpolation (difference) circuit 44. Here, each projection image data is converted into PC1, PC2, PC3.
And

In step # 244, the interpolation circuit 44 sets the PC2
Calculate + PC3 and store the result as SBPCA.
In step # 246, the interpolation circuit 44 uses SBPCA-2 × P.
C1 is calculated, the calculation result is set to SBPCB, and the result is transferred to the CONV circuit 46. In step # 248, the CONV circuit 4
Reference numeral 6 performs a convolution operation on the difference result SBPCB using a convolution operation function. A known method can be appropriately used for the convolution calculation. The convolution result is SBPCV. In step # 250, the BP circuit 48 sends the convolution result SBPCV to FB.
Back projection is performed by adding to the mFBth image mBFI in the image memory 50. A known method can be appropriately used for the back projection. The constant necessary for the back projection of the FB image is created by the constant setting unit 58 under the control of the CPU 56 and transferred to the BP circuit 48.

Incidentally, step # 248 and step # 25
Part or all of 0 may be multiplexed and implemented by different circuits, printed circuit boards, or different devices.

A concrete flow chart of step # 28 in FIG. 4 is shown in FIG. Here, the matrix size of the image is Mx (pixels) × My (pixels). Step # 2
At 81, j = 1 is set, and at step # 282, i = 1 is set.
In step # 283, the pixel value FBI (i, j) at the address (i, j) of the FB image memory 50 is read out, and this is FBP.
V. In step # 285, the pixel value I (i, j) at the address (i, j) of the reconstructed image memory 54 is read out and is set as RPV. At step # 286, the following calculation is performed. RPV = RPV + DW × FBPV (44) Here, DW is the slope (difference) of the primary interpolation coefficient of the helical scan.

At step # 287, the RPV is written in the address (i, j) of the reconstructed image memory 54. Step # 28
In step 8, i ← i + 1 is set, and in step # 289, it is determined whether i ≦ Mx. If i ≦ Mx, the process returns to step # 283. If i> Mx, j ← j + 1 is set in step # 290, and it is determined in step # 291 whether j ≦ My. j ≦
If it is My, the process returns to step # 282. If j> My, in step # 292, the contrast scale device 62 performs a contrast scale conversion for each pixel on all the pixels of the reconstructed image to create a CT image, and the CT image is generated.
It is stored in the image memory 64.

The details of step # 34 in FIG. 4 are the same as those in the flow chart shown in FIG. 8, but step # 286:
The only difference is that RPV ← RPV + DW × FBPV is changed as follows. RPV = RPV-DW × FBPV (45)

(M-1) th FB image (m-1) F
A detailed flowchart of step # 36 of FIG. 4 for reconfiguring the BI is shown in FIG. Although not shown in the drawing, first, information that image reconstruction is performed at slice positions consecutive in the direction in which m decreases, a slope DW of primary interpolation, a value necessary to generate a constant necessary for back projection of an FB image,
And the convolution function and other necessary information from the CPU 56 to the interpolation circuit 44, the CONV circuit 46, and the BP.
Transfer to circuit 48 for storage. It is not necessary to transfer the information already stored in the reconfiguration arithmetic unit.

In step # 362, the CPU 42 reads the (m + 358) th projection data and the (mth) projection data from the raw data storage device 40.
-2) The three projection images of the projection data and the (m + 718) th projection data are read out and supplied to the interpolation (difference) circuit 44.
Here, each projection image data is PC1, PC2, and PC3.

In step # 364, the interpolation circuit 44 sets 2 × P.
C1-PC3 is calculated, and the calculation result is stored as SBPCA. In step # 366, the interpolation circuit 44 sets SBPCA.
-Calculate PC2 and set the calculation result to SBPCB.
Transfer to the V circuit 46. At step # 368, the CONV circuit 46 performs the convolution operation on the difference result SBPCB by the convolution operation function. A known method can be appropriately used for the convolution calculation. The convolution result is SBPCV. Step # 37
At 0, the BP circuit 48 back-projects SBPCV and adds it to the FB image in the FB image memory 50. A known method can be appropriately used for the back projection. The constant necessary for the back projection of the FB image is created by the constant setting unit 58 under the control of the CPU 56 and transferred to the BP circuit 48.

Incidentally, step # 368 and step # 37
Part or all of 0 may be multiplexed and implemented by different circuits, printed circuit boards, or different devices.

As described above, according to the first embodiment, a tomographic image at one slice position is obtained by normal reconstruction processing by interpolating projection data for two rotations, and slice images at adjacent slice positions are obtained. The tomographic image is obtained by reconstructing only the difference between the projection data at both slice positions, and multiplying and combining the coefficients. Therefore, 3 for each slice position
Since it is not necessary to interpolate and reconstruct 60 ° projection data, it is possible to obtain images at consecutive slice positions in a helical scan at high speed, and to change adjacent slice positions continuously using a mouse or the like. , Images at each slice position can be continuously displayed, and the inside of the body can be continuously observed.

Another embodiment of the present invention will be described below. Since the device configuration of the other embodiments is the same as that of the first embodiment, the description thereof will be omitted and only the operation will be described. Also, in the operation, the description of the same parts as those in the first embodiment will be omitted.

In the first embodiment, adjacent images are sequentially reconstructed,
Although it is sequentially displayed, the display may be performed for each of a plurality of slice positions. A second embodiment for doing this will now be described. FIG. 10 is a flowchart thereof. Step #
Since steps from 10 to step # 16 for reconstructing the mFB image are the same as those in the first embodiment, their illustration is omitted. In step # 50, the display slice interval n (n ≧ 1) is designated.

When the direction of increasing m is instructed in step # 22, i = 0 is set in step # 52, and steps # 24 and # 28 are executed as in the first embodiment. Next, in step # 54, i ← i + 1 is set, and in step # 3
At 0, m ← m + 1 is set, and at step # 56, it is determined whether i <n. If i <n, the process returns to step # 24, and i
If not n, the (m + 1) th CT image is displayed in step # 32, and the process returns to step # 20.

When the direction of decreasing m is instructed in step # 22, i = 0 is set in step # 58, and steps # 34 and # 36 are executed as in the first embodiment. next,
In step # 60, i ← i + 1 is set, in step # 40, m ← m−1, and in step # 62, it is determined whether i <n. If i <n, the process returns to step # 34, and i <
If not n, the (m-1) th CT image is displayed in step # 42, and the process returns to step # 20.

In the first and second embodiments, the tomographic image is obtained for each slice position regardless of the presence or absence of display. Next, a third embodiment for obtaining a tomographic image for each of a plurality of slice positions will be described. This embodiment is applicable to the first and second embodiments, but here, the embodiment adapted to the second embodiment is shown in FIG.
This will be described with reference to the flowchart shown in FIG. 1 up to step # 22 of determining the changing direction of the slice position.
It is the same as 0. If the direction in which m is increased is instructed in step # 22, the process proceeds to step # 64, and if the direction in which m is decreased is instructed, the process proceeds to step # 83.

When the slice positions are changed in ascending order, i = 1 is set in step # 64, and (3
The (m + 1) FB image is reconstructed by the equation (3). In step # 67, the (m + 1) th FB image is registered in the register SIGF.
Store in B. In step # 69, m ← m + 1 is set, and in step # 70, it is determined whether or not n = 1. n is initialized in step # 50 shown in FIG.

If n = 1, the process proceeds to step # 80. n =
If it is not 1, the (m + 1) th FB image is reconstructed by the equation (33) in step # 71. In step # 73, the sum of the data in the register SIGFB and the (m + 1) th FB image is newly stored in the register SIGFB. In step # 76, m ← m + 1 is set, and in step # 77, i ←
When i + 1, it is determined in step # 78 whether i <n.

If i <n, the process returns to step # 71 and i
If not n, in step # 80, the image (m-th image) mI to be displayed next is calculated according to the following equation. Where (m
-N) I is the image currently displayed. mI = (m−n) I + DW × SIGFB (46)

In step # 82, the CT image obtained by converting mI into the contrast scale is displayed, and in step # 2
0 (determine whether or not there is an instruction to reconstruct and display continuous slices:
Return to FIG. 4).

When the descending order is instructed in step # 22, i = 1 is set in step # 83, and the mFBth image is stored in the register SIGFB in step # 84. At step # 86, the (m-1) th FB image is reconstructed by the equation (34). In step # 88, m ← m−1 is set, and in step # 89, it is determined whether or not n = 1.

If n = 1, the process proceeds to step # 97, where n
Otherwise, in step # 90, register SIGFB
The data obtained by adding the mFB image to the data of
It is newly stored in GFB. At step # 92, the (m-1) th FB image is reconstructed by the equation (34). Step #
In step 94, m ← m-1 is set, in step 95, i ← i + 1 is set, and in step # 96, it is determined whether i <n.

If i <n, the process returns to step # 90 and i
If not n, the image to be displayed next in step # 97
(Mth image) mI is calculated by the following equation. Where (m +
n) I is the image currently displayed. mI = (m + n) I-DW × SIGFB (47) In step # 98, the m-th image is converted into a contrast scale.
Display the CT image obtained by
It Next, the formula (46) is proved. In equation (35)
If m ← m + n−1, then equation (46) is obtained as follows.
It Similarly, in the equation (47), m ← m−n + in the equation (37)
If it is set to 1 and is similarly expanded, it is required.

Although the slice position is sequentially changed in the front-back direction by the switch in the above-mentioned embodiment, a desired slice position may be designated by interlocking with a pointer such as a mouse or a trackball. In this case, instead of steps # 20 and # 22 of each embodiment, a signal from a pointer such as a mouse is read, and if there is no change in the signal, the process returns to step # 20. The number of slices n to be changed is obtained by multiplying the coefficient. If n> 0, the process in the front direction is executed, and if n <0, the process in the backward direction is executed.

In the above description, the slice position is changed in association with the pointer such as the mouse or the track ball, but a fourth embodiment in which this is further modified will be described. Normally, a display device of an X-ray CT apparatus displays a scanogram (a projected image obtained by scanning only a bed without rotating the X-ray tube) to determine a slice position, and a vertical line ROI is displayed on the scanogram. The cursor is overlaid and displayed, and the display position of the ROI cursor is moved with a mouse or the like. The fourth embodiment is the vertical line ROI on this scanogram.
Align the reconstructed slice to the position of. FIG. 12 is a flowchart thereof.

The scanogram is scanned in step # 101, and the helical scan is performed in step # 102.
Collect the projection data. In step # 103, the scanogram is displayed on a part of the image display device. Step # 10
In step 4, the vertical ROI cursor is superimposed on the scanogram and displayed. The position RP of the vertical line ROI cursor is stored in step # 105, and the cursor position RP is stored in step # 106.
The image number m corresponding to is calculated.

At step # 107, the current image (m-th image) is reconstructed by the equation (19) and the CT image is displayed. At step # 108, the current FB image (m-th FB image) is reconstructed by the equation (23). In step # 110, the display position of the vertical line ROI cursor on the scanogram is moved using the mouse or the like.

At step # 111, the cursor position RP is updated, the change amount is obtained, and the change amount is multiplied by a constant coefficient to obtain the changed slice number n. Step # 11
At 2, it is determined whether n> 0. If n> 0, then step # 24 or step # 52 (see FIG.
0) is executed, and if n> 0 is not satisfied, the same as in the case where the descending order is instructed in step # 22 of the above-mentioned embodiment,
34 (FIG. 4) or step # 58 (FIG. 10). The helical scan may be used as a prescan for positioning, and the function of this embodiment may be used for a scan plan on the scanogram. Further, although the reconstructed image is not stored in the above-described embodiment, the designated image may be stored in the image storage device as needed. Further, the image at the slice position of the designated image is changed by another method, for example, the reflection data.
Data may be used to reconstruct, display, and store.

Next, a fifth embodiment will be described. In the fifth embodiment, with respect to a plurality of axial images, only a part of one image is reconstructed, and from this image, a coronal image,
Create sagittal and oblique images (Multi-planar
construction) It is intended to be displayed. In the conventional MPR display, since the interval between the axial images is larger than that of the pixels in the axial image, there is a problem that a step is visible on the vertical axis (horizontal axis) of the MPR image. However, in this embodiment, since the interval between the axial images can be shortened, it is possible to reconstruct an MPR image having no step. FIG. 14 is a flowchart of the fifth embodiment for reconstructing a coronal image. The same can be done for sagittal and oblique images by changing the direction of the line cursor.

In step # 116, an appropriate axial image is displayed according to any of the above embodiments. Step #
To create a coronal image at 117, specify the slice interval n of the axial image to be reconstructed. Step #
At 118, a horizontal ROI cursor is displayed at the position of the coronal image to be reconstructed on the axial display image. In step # 119, m = 1 is set. In the following operation, the back projection and the storage of the image are executed only for the pixels of one line in which the horizontal line ROI cursor is displayed.

In step # 120, the image (first image) is reconstructed by the equation (19) and stored. Step # 121
Then, the FB image (first FB image) is reconstructed by the equation (23). In step # 123, it is determined whether (m + n) is greater than or equal to the maximum slice number taken.

If (m + n) is not equal to or larger than the maximum slice number taken, in step # 123, the steps (64) to # 80 of the third embodiment shown in FIG.
+ N) Reconstruct image, store, and return to step # 122. However, also in this case, the reconstruction is performed only for the pixels of one line of the horizontal ROI cursor.

If it is determined in step # 122 that m + n is greater than or equal to the maximum slice number taken, step #
At 124, the coronal image is reconstructed and displayed from the previously stored images by known methods.

In the fifth embodiment, one MPR image is created. However, the axial image (including the line ROI cursor) and the MPR image are displayed at the same time, and the line ROI cursor is moved by the mouse to move the line. The MPR image may be reconstructed and displayed according to the position of the ROI cursor. By doing so, MPR images can be continuously observed.

Next, reconstruct a plurality of axial images,
A sixth embodiment for creating and displaying a three-dimensional display image (surface display or the like) from these images will be described. The conventional three-dimensional display image has a problem that a step is visible in the slice thickness direction because the interval between the axial images is larger than that of the pixels of the axial image. According to this embodiment, the interval between the axial images can be shortened, so that a three-dimensional display image without steps can be reconstructed. Steps for reconstructing (m + n) images #
Steps 116 to # 124 are performed in the same manner as in the fifth embodiment. However, the reconstruction of steps # 120, # 121 and # 123 is performed for all the images. Therefore, in the flowchart shown in FIG. 13, instead of reconstructing the coronal image in step # 124, a three-dimensional display image may be created from the stored image by a known method and displayed.

In the above-described embodiment, the slice and the projected image are at the same position, but the present invention can be implemented by modifying the equation even when the positions are deviated.

Further, even when the interval between the slice positions is not an integral multiple of the interval between the projection positions, the present invention can be implemented by similarly modifying the mathematical expression. Therefore, in the fifth embodiment and the sixth embodiment, which is a modification thereof, it is possible to match the intervals of the slice positions with the pixel pitch of the axial image. The present invention is not limited to the above-mentioned embodiments, but can be implemented with various modifications.

[0089]

As described above, according to the present invention,
Images n slices apart can be reconstructed by only n convolutions and backprojections, and image addition and subtraction. Since the execution time of image addition and subtraction is short, the reconstruction time is the time required for convolution and back projection almost n times. On the other hand, the time required for normal reconstruction is 36
This is the time required for convolution and back projection for several times of 0 ° projection. Usually n is 360 °
Since it is far less than the number of projections of the image, the present invention can reconstruct images separated by n slices in a short time. Therefore, continuous images taken by the helical scan can be observed in almost real time. In addition, since only necessary images can be stored, the image storage capacity can be small. Further, it is possible to reconstruct an MPR image having a small step.

[Brief description of drawings]

FIG. 1 is a diagram showing a schematic configuration of a first embodiment of an X-ray CT apparatus according to the present invention.

2 is a detailed block diagram of the reconfiguration circuit of FIG.

FIG. 3 is a diagram illustrating a helical scan using projection data collected in the first embodiment.

FIG. 4 is a flowchart showing the operation of the first embodiment.

5 is a flowchart showing details of a step of reconstructing the m-th image in FIG.

6 is a flowchart showing details of a step of reconstructing the mFB image of FIG.

7 is a flowchart showing details of a reconstruction step of the (m + 1) th FB image in FIG.

FIG. 8 is a flowchart showing details of a (m + 1) th image reconstruction step of FIG. 4;

9 is a flowchart showing details of a reconstruction step of the (m−1) th FB image in FIG.

FIG. 10 is a flowchart showing the operation of the second embodiment of the present invention in which CT images are displayed every n slices.

FIG. 11 is a flowchart showing the operation of the third embodiment of the present invention for reconstructing and displaying a CT image every n slices.

FIG. 12 is a flowchart showing the operation of the fourth embodiment of the present invention in which the position of the vertical line ROI cursor on the scanogram is set as the slice position to be reconstructed.

FIG. 13 is a flowchart showing the operation of the fifth embodiment of the present invention for displaying an MPR image.

FIG. 14 is a schematic block diagram of a processing circuit relating to image reconstruction calculation of a conventional X-ray CT apparatus.

[Explanation of symbols]

20 ... X-ray tube, 24 ... Detector, 26 ... Preprocessing circuit, 28
... Reconfiguring circuit, 30 ... Display device, 32 ... Input device, 40
... raw data storage device, 44 ... interpolation circuit, 46 ... convolution circuit, 48 ... back projection circuit,
50 ... FB image memory, 54 ... Reconstruction memory, 60 ... Image combining circuit, 62 ... Contrast scale circuit, 64 ...
CT image memory.

Claims (3)

[Claims] 1. An X-ray computed tomography apparatus for collecting projection data by helical scanning, wherein a predetermined reconstruction calculation is performed on a first projection data group for a predetermined projection angle before and after the first slice position. A means for obtaining a tomographic image at the first slice position, and a second unit necessary for reconstructing a tomographic image at a second slice position adjacent to the first slice position from the first projection data group. Means for extracting a first projection data part that is not common to the second projection data group, and a second projection data part having the same projection angle as the first projection data part from the second projection data group. Means for performing the predetermined reconstruction calculation on the first and second projection data portions, and combining the reconstruction result and the tomographic image at the first slice position to obtain a tomographic image at the second slice position. Seeking a statue X-ray computed tomography apparatus characterized by comprising a that means. 2. An X-ray computed tomography apparatus for collecting projection data by helical scanning, wherein predetermined reconstruction is performed after interpolating projection data of 360 ° before and 360 ° after the first slice position. Calculate
Means for obtaining a tomographic image at the first slice position, projection data at the first slice position, projection data 360 ° apart from the projection data at the first slice position, and rear 3
Means for obtaining a first sub-image by performing a predetermined reconstruction operation on projection data obtained by combining projection data separated by 60 °, and multiplying the first sub-image by a predetermined coefficient, By adding to the tomographic image at the first slice position,
An X-ray computed tomographic imaging apparatus, comprising: means for obtaining a tomographic image of a second slice position which is adjacent to the first slice position and is temporally later than the first slice position. 3. 360 ° before the first slice position,
The means for obtaining a second sub-image by performing a predetermined reconstruction calculation without interpolating the projection data for the subsequent 360 °, and multiplying the second sub-image by a predetermined coefficient, and the multiplication result is It further comprises means for obtaining a tomographic image of a third slice position, which is adjacent to the first slice position and is temporally prior to the first slice position, by subtracting from the tomographic image of the first slice position. The X-ray computed tomographic imaging apparatus according to claim 2, wherein