JP3535248B2 - Computer tomography equipment - Google Patents

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 more particularly to image reconstruction of, for example, a third-generation X-ray CT apparatus adopting a helical scan method. .

[0002]

2. Description of the Related Art In recent years, in order to shorten the time required to collect projection data, an X-ray CT apparatus of a continuous rotation type with movement of a bed, a so-called helical scan type has been developed. In this apparatus, projection data necessary for image reconstruction centering on an arbitrary position in the trajectory of continuous rotation of the X-ray source can be cut out from the acquired data to obtain a slice image at an arbitrary position. By sequentially shifting arbitrary positions at fine intervals, a large number of slice images with a very fine pitch can be obtained.

The most basic reconstruction of the helical scan type X-ray CT apparatus is to interpolate projection data of the same projection angle before and after the designated position to project a 360 ° projection of the designated position (slice plane). There is a method in which data (virtual projection data) is created, reconstruction processing is performed on the data, and each slice image is created. This reconstruction method includes a full data reconstruction method that interpolates virtual projection data from projection data of a total of 720 ° each 360 ° before and after the specified slice position, and a total of 360 ° projection 180 ° before and after the slice position. There is a half data reconstruction method that interpolates from data.

Although the helical scan type apparatus can obtain an image at an arbitrary slice position, in any reconstruction method, if the slice position is slightly shifted, interpolation processing and re-processing are performed from the beginning one by one. Since the configuration calculation process is performed, there is a drawback that the reconstruction time becomes long in proportion to the number of slices to be reconstructed.

Therefore, as a method of efficiently reconstructing images at a large number of slice positions having continuous fine pitches, the first one is created by a normal reconstruction calculation process, and the images at adjacent positions are slices of the first slice. A method has been proposed in which the difference between the data required for the image reconstruction calculation and the data required for the adjacent position reconstruction calculation is calculated, only this difference is reconstructed, and the calculation result is added to the first tomographic image. . In this method, a large number of convolution backprojection operations are performed for the first slice, but one convolution backprojection operation is performed for the subsequent slices, and then simple operations such as a sum product operation are performed. All you have to do is do it. It can be said that this method is a very efficient method because the convolution back-projection operation has a larger amount of operation than other operations.

However, this method minimizes the image reconstruction time, but since the difference equivalent amount is added to the immediately preceding reconstructed image (reference image), the adjacent images are successively reconstructed. , Data processing errors tend to accumulate. Further, since the continuous slice images are reconstructed while sequentially updating the reference image, there is a drawback that an image other than a desired slice must be created.

[0007]

SUMMARY OF THE INVENTION The present invention has been made to address the above-mentioned circumstances, and is a computer of a helical scan system capable of efficiently reconstructing an image at an arbitrary slice position among a large number of consecutive slice positions. A tomography apparatus is provided.

[0008]

A computer tomography apparatus according to the present invention comprises means for collecting projection data by helical scanning and projection data of a partial projection angle range obtained by dividing a projection angle range, Virtual projection data is created by weighting the projection data for the partial projection angle range with an interpolation coefficient, and reconstruction calculation is performed using the virtual projection data to sequentially obtain a plurality of partial projection images in the partial projection angle range. and partial reconstruction unit to obtain a, a storage unit for storing a plurality of partial projection image obtained by the partial reconstruction unit, before
And a weighting addition unit that obtains a tomographic image at a desired slice position by weighting and adding the plurality of partial projection images corresponding to the projection angle range .

[0009]

According to the computer tomography apparatus of the present invention, partial projection images for each partial projection angle are created as intermediate data, and a tomographic image of a desired slice is obtained by weighted addition of these intermediate data. The desired tomographic image can be directly obtained by using only the projection data within the predetermined projection angle range necessary for the configuration, that is, the necessary and sufficient projection data as the original data, which increases the amount of calculation as compared with the conventional method. There is an advantage that errors do not accumulate without needing to create an unnecessary intermediate image.

[0010]

Embodiments of the computer tomography apparatus according to the present invention will be described below with reference to the drawings. Although full data reconstruction method in X-ray CT apparatus of the helical scan system is described in detail in JP Application Rights 4-168919,
The principle of full data reconstruction will be briefly described. Here, the X-ray used for scanning is a fan beam.

In order to create an image at a certain slice position, it is necessary to interpolate virtual projection data V (m) at a certain slice position from the preprocessed projection data D (m, n). FIG. 2 shows projection data D (m,
It is a figure which shows the weighting coefficient applied to n).

As will be described later, according to the present invention, reconstruction is considered as weighted back projection, and virtual projection data is created by linear interpolation, from which an image is reconstructed. Therefore, the weight coefficient is an abstraction of the linear interpolation coefficient.

The index n of the projection data in FIG. 2 indicates a channel (corresponding to each detector of the detector array), and n = 1 to N. The index m indicates the collection position of the projection data (slice position in the body axis direction), and here it is assumed that the projection data in the range of m = 1 to MM is collected. Projection data group (vertical lines in FIG. 2) D (m, 1) to D (m, corresponding to one fan beam
N) is called view V (m). Let M be the number of views per 360 ° (one rotation).

Backprojection is performed by multiplying (2M + 1) views, 360 ° each before and after the specified position (slice center) m ₀ , for a total of 720 °, with a weighting factor for interpolation.
When a convolution calculation (reconstruction calculation) is performed, a tomographic image is obtained. This is shown in the following equation.

[0015] Here, I (m ₀ ): tomographic image at position m ₀ τ _M : weighting coefficient τ _M (k) = (M− | k |) / M, | k | ≦ M, = 0, otherwise (That is, the weighting coefficient does not depend on the channel number n) Φ: Convolution back projection operation Since the convolution back projection operation Φ is a linear operation, the operation Φ is a constant multiple (τ _M (m− for each m. m ₀ ) times is a constant times), the order can be changed (commutation is possible). Therefore,
Equation (1) can be modified as follows.

[0016] When performing reconstruction based on the equation (1 ′), M =
M '× B, 360 ° view in B blocks,
In other words, the 720 ° view is divided into (2B + 1) blocks. M ′ (= M / B) is the number of views per block, and one scale on the axis of the view number in FIG. 3 is the view of M ′.

Then, the weighting factor τ _M (m−m) of (1 ′)
₀ ) is a weighting coefficient τ _{M ′} (m−
It can be expressed as follows using (m ₀ + bM ′)). Here, b =-(B-1),-(B-2), ...,
0, ..., B-1.

[0018] By substituting the equation (2), the equation (1 ') can be expressed as follows.

[0019]

[Equation 1]

On the right side of equation (3) It is a partial projection image using the view V (m) in this range.
This is referred to as P ^b _{M ′} (m ₀ ), abbreviated as P ^b . Then,
Equation (3) can be expressed as follows.

[0021] That is, it can be seen from the equation (4) that the image I (m ₀ ) at a certain slice position m ₀ can be calculated by weighted addition of the partial projection images P ^b _{M ′} (m ₀ ).

FIG. 1 shows a block diagram of the first embodiment according to the above principle. The fan beam X-rays that have passed through the subject (not shown) are detected by an array detector (not shown), and the measurement data indicating the X-ray transmittance output from the detector is after A / D conversion. It is supplied to the preprocessing circuit 10 and preprocessing such as log conversion is performed. The output of the preprocessing circuit 10 is stored in the projection data memory 12. Here, the projection data is assumed to be collected by helical scanning. The projection data memory 12 is connected to the system bus 26. The slice center setting unit 20 is also connected to the system bus 26,
The center position of the slice to be reconstructed is set, and projection data at projection angles of 360 ° before and after the slice center are supplied from the projection data memory 12 to the partial projection image reconstruction unit 14.

The partial projection image reconstruction unit 14 is supplied with 72.
As shown in FIG. 3, the 0 ° projection data is divided into projection data blocks for each predetermined projection angle range, weighted and reconstructed for each block to obtain a partial projection image. The weighting coefficient τ _{M ′} (m− (m ₀ + b
M ′)) is stored in the weighting coefficient memory 18.

The partial projection image is stored in the partial projection image memory 16. The output of the partial projection image reconstruction unit 14 and the output of the partial projection image memory 16 are connected to the system bus 26. The partial projection image weighting addition circuit 22 is also connected to the system bus 26, and the partial projection images are weighted and added according to the equation (4) to obtain a tomographic image having the set position as the slice center.

The actual calculation procedure will be described. The projection data in the range shown in FIG. 2 is collected by the helical scan and pre-processed. The projection data D (m, n) is stored in the projection data memory 12. The number of divisions B and the number of views M ′ for one block are determined in advance. As described below,
Since the slice center can be arbitrarily shifted with this M'as a unit, it is preferable for the block to be small in terms of accuracy, but in reality it is not necessary to update the image for each view, so the number of divisions B For example, it may be decided to be about 10. If the number of divisions is too large, the partial projection image memory 16
Has the drawback of increasing the capacity of the.

[0026] Determine the slice center m ₀ of the tomographic image to be reconstructed by the slice center setting unit 20, the front and rear each 360 ° of the slice center m _0, a total of 720 ° (2M + 1) number of views partial projection from the projection data memory 12 It is supplied to the image reconstruction unit 14. (2M + 1) views are divided into B blocks every M '. The partial projection image reconstruction unit 14 performs weighting calculation and reconstruction calculation on the view of each block, obtains a partial projection image, and stores it in the memory 16. Here, when the partial projection image of the block is already obtained, the creation of the partial projection image is omitted. The partial projection image weighting addition circuit 22 obtains a tomographic image at the slice position m0 by weighting addition of the partial projection images of all blocks according to the equation (4). Therefore, even when the tomographic images at a large number of slice positions are continuously reconstructed, if a partial projection image is created, it can be obtained only by a simple sum-product operation.
There is an advantage that the amount of calculation is small.

[0027] Here, it is according to the reconstruction operation method of the present _{^{invention, P b M '(m 0}} + b'M') = P b + b 'M' (m 0).

Because,

[0029] Therefore, the partial projection image P ^b _M (m ₀ ) is b = −B +
1, -B + 2, ..., 0, ..., B-1, B, B + 1, ... Sequentially, the image I is sequentially calculated every m ′ from m _{0 by} only the weighted addition shown in the above equation. I see what I can do.

Now, Φ (V (m)) is sequentially calculated and P ^b
_The step of sequentially obtaining _{M '} (m ₀ ) will be described. When m is expressed as m = m ₀ + bM ′ + C (0 ≦ C <M ′), Φ (V (m)) obtained by the convolution back projection is two partial projection images P
^It is used to calculate ^b _{M ′} (m ₀ ) and P ^{b + 1} _{M ′} (m ₀ ).
That is, P ^b _{M ′} (m ₀ ) is weighted by τ _{M ′} (c) and added, and P ^{b + 1} _{M ′} (m) is weighted by τ _{M ′} (c−M ′). And add.

After all, the convolution back-projection calculation need only be performed once for one view V (m). As described above, according to this embodiment, (2M + 1) views V (m) of 720 ° before and after the slice position are divided into (2B + 1) blocks,
For each block, convolution back projection calculation and weighted addition processing within the block are performed to obtain partial projection images, and tomographic images are obtained by weighted addition of partial projection images. Here, each partial projection image is stored in the memory. Therefore, when the slice position shifts, only a new partial projection image of the block corresponding to the shift needs to be newly obtained. As described above, in the present embodiment, when obtaining the tomographic image, the calculation amount of the calculation amount is changed because the partial projection image that is made into a block and the weighted addition of the partial projection image are formed instead of the serial data update type calculation process. It is not necessary to repeatedly perform many convolution / back-projection calculations, and it is possible to continuously reconstruct tomographic images at slice positions slightly shifted with a small amount of calculation. In addition, the partial projection images are weighted and added for each slice image, instead of reconstructing the adjacent images sequentially by adding the difference equivalent amount to the immediately preceding reconstructed image (reference image). It is not necessary to sequentially reconstruct images of continuous slices, only images of necessary slices can be easily created, and errors in data processing do not accumulate.

The present invention is not limited to the above-described embodiments, but can be implemented with various modifications. For example, in the above description,
During blocking, but by dividing the weighting coefficient tau _M to (2B-1) pieces of weighting coefficients tau _{M ',} and further subdivided this, for example, as shown in FIG. 4, the first half of the triangular weighting function tau
_It may be divided into ₁ and the latter half τ ₂ .

[0033]

As described above, according to the present invention, a partial projection image as intermediate data is created, and by weighting and adding these intermediate data, an image at a constant position interval is obtained, so that necessary and sufficient projection data can be obtained. Direct images are obtained using only as raw data. Therefore, a tomographic image is created without increasing the amount of calculation and in a state where an error does not easily occur.
Can be displayed. Moreover, it is not necessary to create an unnecessary intermediate image.

[Brief description of drawings]

FIG. 1 is a first computer tomography apparatus according to the present invention.
The block diagram which shows the structure of an Example.

FIG. 2 is a diagram showing a relationship between projection data and weighting factors upon reconstruction.

FIG. 3 is a diagram showing an example of weighting coefficients when a partial projection image is created.

FIG. 4 is a diagram showing another example of weighting factors when creating a partial projection image.

[Explanation of symbols]

10 ... Preprocessing circuit, 12 ... Projection data memory, 14 ... Partial projection image reconstruction unit, 16 ... Partial projection image memory, 18 ... Weight coefficient memory, 20 ... Slice center setting unit, 22 ... Partial projection image weighting addition circuit, 24 ... Display section.

Claims (2)

(57) [Claims] 1. A unit for collecting projection data by helical scanning, and each time projection data of a partial projection angle range obtained by dividing a projection angle range is obtained, projection data for the partial projection angle range is interpolated by an interpolation coefficient. Create virtual projection data by weighting, perform reconstruction operation using this virtual projection data,
A partial reconstructing unit that sequentially obtains a plurality of partial projection images in the partial projection angle range ; a storage unit that stores the plurality of partial projection images obtained by the partial reconstructing unit; and a storage unit that corresponds to the projection angle range. A weighted addition unit that weights and adds a plurality of partial projection images to obtain a tomographic image at a desired slice position. 2. The weighted addition unit obtains a new partial projection image in addition to the plurality of projection images forming the tomographic image of the desired slice position stored in the storage unit in the partial reconstruction unit. 2. The computer according to claim 1 , wherein the plurality of partial projection images and the new partial projection image are weighted and added to obtain a tomographic image with a slice position shifted from the tomographic image at the desired slice position. Tomography equipment.