JPH0814622B2 - Camera rotation type ECT device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION A. Field of Industrial Application The present invention relates to a camera rotation type ECT device, and more particularly to a camera rotation type EC equipped with a correction means for performing scattered ray correction and absorption correction.
Regarding T equipment.

B. Prior Art In a camera-rotating ECT device, gamma rays that are Compton scattered inside the subject (patient) or collimator provide false spatial information, which causes deterioration of the image quality of the projection image and eventually the reconstructed image.

In order to eliminate this inconvenience, generally, an energy signal is subjected to wave height analysis, and only an event signal within a certain energy window range is processed as valid. This is because the Compton scattered rays have lost some energy as compared with the original gamma rays.

Ie a single energy level, eg ^99m Tc
In the case of 140 keV gamma rays (technium in which the nuclei are in an excited state), the relationship between the wave height of the energy signal and the count number (spectrum) in the same projection image can be illustrated by Compton scattering or the like. As shown in the figure, dispersion occurs in the energy signal wave height.

By the way, since the energy resolution of the gamma camera is about 10% in full width at half maximum (FWHM), about 20% is required as the energy window width w in order to detect the dispersed energy signal with a high degree of accuracy. .
This window width w is 10 on the low level side and the high level side with respect to the energy signal peak value E _P of the count number peak value.
% Is taken (symmetrical window method). The effect of Compton scattering appears on the lower level side than the peak crest value E _P , and appears on the lower level as the scattering angle increases.

Therefore, as described above, the window width w is about 20%.
If there is no choice but to set a very large range, there is a problem that low-angle Compton scattered rays enter the window width w and cannot be removed.

In order to improve this problem, the following measures have been taken or proposed conventionally.

(A) Asymmetric window method This is, as shown in FIG. 7, shifting the center of the window width w by a certain amount to a higher level side than the peak peak value E _P ,
The window width w is asymmetric with respect to the peak crest value E _P. As a result, the count number of Compton scattered rays that enter the window width w can be reduced.

(B) Energy weighted collection method This is focused on the fact that the variance of the energy signal wave height changes depending on the type of nuclide and collimator, and the weighting function corresponding to these types is obtained in advance, and the detection obtained at the time of imaging is detected. The signal is corrected by its weighting function.

(C) Correction Method According to Scatterer Thickness For example, as described in JP-A-62-167491, the dispersion of the energy signal wave height is calculated for each pixel unit,
The correction function obtained based on this is used to perform correction according to the scatterer thickness.

C. Problems to be Solved by the Invention However, each of the above improvement measures has the following problems.

As shown in FIGS. 2A and 2B, when the rotation position of the gamma camera (scintillation camera) 1 is different, the dispersion state of the energy signal wave height is as shown by the solid line and the broken line in FIG. Changes to. In the figure, 2 is a collimator, m is a subject, m ₁ is a target organ, and RI is accumulated in this target organ m ₁ .

In the case of (A) of FIG. 2, since the thickness from the target organ m ₁ to the body surface of the subject m in the direction perpendicular to the camera surface of the gamma camera 1, that is, the scatterer thickness d ₁ is small, Compton scattering The influence of the line is small and the dispersion of the energy signal wave height becomes small, and it becomes like the solid line in FIG. 7, but in the case of FIG. 2B, the scatterer thickness d ₂ is large, so the Compton scattered line Is greatly affected and the dispersion of the energy signal wave height becomes large, as shown by the broken line in FIG.

The change in the scatterer thickness d depends not only on the rotation position of the gamma camera 1 but also on the difference in the position of the target organ m ₁ and the subject m.
It is also caused by the difference in shape (whether the body shape is fat or thin).

However, according to the asymmetric window method of the above (a), since the width w of the window and the level range in the energy signal wave height are always fixed, the correction for the change in the dispersion of the energy signal wave height is incomplete. In addition, there is a problem that the larger the variance is, the more the count loss occurs.

Further, in the energy weighted collection method of (b), since the rotation position of the gamma camera, the shape of the subject, and the change of the target organ are not taken into consideration in the weighting function, the energy signal caused by the change of the scatterer thickness is taken into consideration. Unable to compensate for changes in wave height dispersion.

The correction method according to the scatterer thickness of (c) above is effective in the theoretical calculation with a sufficiently large number of counts, but in actual clinical practice, the correction function described above is accurately calculated for each pixel unit. It is difficult to obtain a sufficient number of counts (usually dozens or hundreds of counts per pixel), and statistical errors appear significantly, and when creating a correction function, and based on that correction function However, there is a problem that the amount of processing data when performing the correction calculation becomes enormous and it is not practical.

Since the gamma ray incident on the gamma camera is absorbed in the subject, it is necessary to perform absorption correction in order to obtain quantitative data.

The present invention has been made in view of such circumstances, and even if the count number is small as in the case of actual clinical practice, the present invention can be performed according to the rotational position of the gamma camera, the shape of the subject, and changes in the target organ. It is an object of the present invention to provide a camera rotation type ECT device capable of always performing appropriate scatter ray correction and absorption correction at the same time.

D. Means for Solving the Problems The present invention has the following configuration in order to achieve such an object.

That is, the scattered radiation correction device of the camera rotation type ECT device of the present invention has the energy level of the gamma ray to be measured which is 1
An energy signal height analyzer with multiple energy spectrum windows for each individual, a means for collecting image data separately for each window, a means for obtaining the total number of counts of the images obtained for each window, and a peak Means for calculating the ratio of the total count number for the windows on the higher level side of the crest value to the sum of the total count numbers for each window, and means for pre-storing the weighting factor for each window using this ratio as a parameter, Weighted addition processing means for multiplying the image data for each window by a weighting coefficient for each window read from the storage means based on the ratio in units of projection image and adding the result to obtain a final projection image. And for each window stored in the storage means See coefficients, the sum of the corrected count after addition process is characterized by being determined to be constant.

E. Operation In the present invention, within the same projection image, the scatterer thickness for forward scattering in the subject is substantially constant regardless of the pixel within the range of the target organ or a region of interest that is a part thereof. It is possible to treat it as if there is some, and that the count number in the range of the other region is generally smaller than the count number in the range of the region of interest that is the target organ or a part thereof. It was used.

As the count in the range of the other region, gamma rays from gamma-ray emitting nuclides distributed in parts other than the target organ in the subject, gamma rays emitted from the ground, buildings, etc., background radiation such as cosmic rays. There is a count etc.

 The operation of the configuration of the present invention is as follows.

Considering that there is almost no effect of Compton scattered radiation in the windows higher than the peak wave height, the total number of counts in the windows higher than the peak wave height is the number of total counts of all windows in advance. The ratio to the sum is calculated, and the weighting factor for each window is stored based on this ratio.

At the time of actual measurement, the energy signal obtained by the gamma camera is divided into each window by the energy signal wave height analyzer having a plurality of windows, and the total number of counts in the windows on the higher level side than the peak wave height value. Then, the total count number of the divided images for each window is obtained, and based on them, the ratio of the total count number in the high-level side window to the sum of the total count numbers at the time of actual measurement is obtained.

Then, for each projection image, the weighting factor for each window at the projection angle is read from the storage means based on the ratio at the time of actual measurement, and the weighting factor for each window is read as an image for each window. Multiply the data and add the results together to get the final projection image.

The weighting factor is common to all pixels at the same projection angle. Therefore,
As long as the count of each pixel is known, no matter how small the count is, the weighted addition process described above makes it possible to enhance the scattered radiation correction regardless of the rotation position of the gamma camera, the shape of the subject, and changes in the target organ. Can be done with precision.

Further, since the weighting coefficient for each window is stored in the storage means so that the total sum of the corrected counts after the addition processing is substantially constant, absorption correction can be performed at the same time.

F. Embodiment Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.

2A shows a state in which the gamma camera 1 is positioned right above the subject m, and FIG. 3A shows the gamma ray energy spectrum of the projection image in this state. In this case, since the scatterer thickness d ₁ is small, the number of counts at the peak crest value E _P is large.

2B shows a state in which the gamma camera 1 is positioned right beside the subject m, and the energy spectrum of the projection image in this state is shown in FIG. 3B. In this case, since the scatterer thickness d ₂ is large,
The number of counts at the peak crest value E _P is low. In FIGS. 2A and 2B, the nuclide and collimator 2
Of course, the subject m, the target organ m _{1 and} other conditions are all the same, and the nuclide used is a single-energy gamma-ray emitting nuclide (for example, ^{99 m} T with an energy of 140 keV only).
c) shall be used.

Since the scatterer thickness is larger in the case of FIG. 2 (B) than in the case of FIG. 2 (A), the gamma rays that are Compton scattered at a low angle in the subject m are gamma cameras 1.
There is a high probability of being incident on the projection image, and the proportion of false information included in the projection image is high accordingly.

As a window in the energy spectrum, conventionally, only one window is provided for one energy level, but in the present invention, the window is divided into a plurality of windows. In the case of FIG. 3, the case where the number of divisions is 3 is illustrated, and the first window W ₁ , the second window W ₂ , and the third window W ₃ are set in order from the lowest energy level.
The third window W ₃ is determined to exist on the higher level side than the peak crest value E _P. In addition, each window W ₁ , W ₂ , W ₃
The window widths may be the same or different, but in FIG. 3, they are the same.

The third window on the higher level side than the peak value E _P
At W ₃ , there is almost no effect due to the incidence of Compton scattered rays. The influence appears in the first window W ₁ and the second window W ₂ .

The difference in the incident probability of Compton scattered rays between (A) and (B) in FIG. ₃ is that the total counts N ₁ , N ₂ , N ₃ of the images in the individual windows W ₁ , W ₂ , W ₃ are different. Can be adapted to. More specifically, it is as follows.

Although two states of the gamma camera 1 have been described with reference to FIGS. 2 and 3, the actual number n of projection angles is several tens to hundreds of tens. Therefore, i-th (i = 1,2 ...
n) projection angle, each window W ₁ , W
The total number of counts of each image for each ₂ and W ₃ N _1i , N _2i , N _3i
Considering the ratios r _1i , r _2i , r _3i of the total number of counts to N _si , r _1i = N _1i / N _si r _2i = N _2i / N _si r _3i = N _3i / N _si . However, N _si = N _1i + N _2i + N _3i .

Ratio of first window W ₁ and second window W ₂ r _1i , r _2i
Is while significantly affected by Compton scattered radiation, the ratio r _3i of the third window W ₃ being hardly affected, the proportion r _3i of the third window W _3, i.e., Based on the weighting factors k _1i , k _2i , for correction at the i-th projection angle for each window W ₁ , W ₂ , W ₃ .
_Find k _3i in advance by experiments. The weighting factors K _1i , K _2i and K _3i are the same as those described in Japanese Patent Application Laid-Open No. 62-167 of the same applicant.
This corresponds to the scatter correction coefficient f described in Japanese Patent No. 491, and the weighting coefficients K _1i , K _2i and K _3i are calculated by the same method as the scatter correction coefficient f described in the same publication. However, since the method described in Japanese Patent Laid-Open No. 62-167491 is correction for each pixel, the weighting factor K _1i ,
When calculating K _2i and K _3i , projection data,
And the total number of counts N _1i , N _2i , of each window W ₁ , W ₂ , W ₃ .
It is necessary to replace with N _3i , calculate the weighting coefficient for each window, and store them. The weighting factors k _1i , k _2i , k _3i are stored in advance in the RAM 7 (see FIG. 1) described later in a state corresponding to the address created by the ratio r _3i of the third window W ₃ .

The weighting factors k _1i , k _2i , k _3i are shown in FIG. 4, for example. In this case, in order to be able to perform the absorption correction as well as the scattered radiation correction at the same time, the sum of the corrected count numbers C _i (x, y) in the arithmetic expression of the weighted addition processing in pixel units described later, that is, ∬ Each weighting factor k _1i , k _2i , k _3i is determined so that C _i (x, y) dxdy = k _1i · N _1i + k _2i · N _2i + k _3i · N _3i is constant.

Regardless of the projection angle number i, the weighting factor k _3i of the third window W ₃ is simply
This is because if the other weighting factors k _1i and k _2i are set in the state where is always “1”, the absorption correction cannot be performed even if the scattered radiation correction can be performed in the correction by the equation described later. First, the third window W ₃ is said to be hardly affected by the incidence of Compton scattered rays, but it is affected by the photoelectric effect and absorption by Compton scattering, so the weighting factor k _3i is always "1".
In that case, the absorption correction cannot be performed, and it becomes necessary to perform the absorption correction separately from the scattered radiation correction.

Next, weighted addition processing in pixel units will be described.

The number of pixels (x, y) in the image corresponding to the first window W ₁ of the energy spectrum at the i-th projection angle is defined as (x, y), which is the pixel in the two-dimensional direction on the camera surface of the gamma camera 1. To C _1i (x,
y), the count number of the pixel (x, y) in the second window W ₂ is C _2i (x, y), and the count number of the pixel (x, y) in the third window W ₃ is C _3i (x , y).

The x value and the y value take a large number of consecutive values within the effective visual field of the gamma camera 1. For example, gamma camera 1
Assuming that the two-dimensional pattern data obtained by has 6 bits in each of the X and Y directions,
1,2 …… 64 values are taken.

The weighting factors k _1i , k previously obtained as described above
_2i and k _3i , in the first window W ₁ , the count number C _1i (x, y) for each pixel (x, y) is multiplied by the weighting coefficient k _1i to obtain k _1i · C _1i (x , y) is calculated. Similarly, in the second window W ₂ , k _2i · C _2i (x, y) is calculated, and in the third window W ₃ , k _3i · C _3i (x, y) is calculated.

Then, finally by adding as in the following equation, each pixel (x,
Calculate the corrected count number C _i (x, y) in y).

C _i (x, y) ＝ k _1i・ C _1i (x, y) ＋ k _2i・ C _2i (x, y) ＋ k _3i・ C _3i (x, y) ………………… Weighting coefficient k _1i , k _2i and k _3i are different for each projection angle, but at the same projection angle,
It is common to all pixels regardless of the x value and the y value of the pixel (x, y). This approximates that within the same projection image, the scatterer thickness for forward scattering in the subject is almost constant regardless of the pixel within the range of the target organ or a region of interest that is a part thereof. It can be handled as a whole, and the fact that the count number in the range of other regions is generally smaller than the count number in the range of the region of interest that is the target organ or a part thereof is utilized. is there.

Therefore, if you know the count number of each pixel,
No matter how small the number of counts is, the weighted addition process of the above formula enables highly accurate scatter correction and absorption correction regardless of the rotation position of the gamma camera, the shape of the subject, and changes in the target organ. By performing image reconstruction by back projection based on the projection image data corrected in this way, a high-quality tomographic image with high spatial resolution and good contrast can be obtained.

Regarding the weighting factors k _1i , k _2i , k _3i , by applying this not only to the pixels in the range of the region of interest that is the target organ or a part thereof, but also to the pixels in other peripheral ranges, In the clinical image, the image quality (spatial resolution, contrast) of the region of interest, which is the target organ or a part thereof, and the vicinity thereof is high, and the quantification is high. There is no particular problem because it is not necessary to give much importance to other surrounding images.

The total count numbers N _1i , N _2i , N _3i can be expressed as follows.

N _1i = ∬C _1i (x, y) dxdy N _2i = ∬C _2i (x, y) dxdy N _3i = ∬C _3i (x, y) dxdy Next, scattered rays from the camera rotation type single photon ECT device The configuration of the correction device will be described with reference to FIG.

The gamma camera 1 outputs an x-coordinate signal X on which gamma rays are incident, a y-coordinate signal Y, an energy signal E, and an unblank signal (UNBLANK signal: timing signal) T. Gamma rays are randomly incident on each pixel on the two-dimensional matrix on the camera surface.

As shown in FIG. 5, when one gamma ray is incident on a pixel, an x coordinate signal X and ay coordinate signal Y of that pixel are output. The unblank signal T is output when the x coordinate signal X and the y coordinate signal Y become stable, and the energy signal E of one incident gamma ray is output in synchronization with the unblank signal T.

The energy signal E and the unblank signal T are
It is input to the _first to third wave height analyzers 3 ₁ , 3 ₂ , 3 ₃ provided corresponding to the _three windows W ₁ , W ₂ , W ₃ . The x-coordinate signal X and the y-coordinate signal Y are the first to third image memories 4 ₁ , 4 provided corresponding to the _three windows W ₁ , W ₂ , W _{3.
Entered in 2} , 4 ₃ . These image memories 4 ₁ , 4 ₂ , 4
Reference numeral ₃ is a two-dimensional image storage memory that adds 1 to the count number to be stored in the storage area designated by the x coordinate signal X and the y coordinate signal Y.

By the wave height analysis in each wave height analyzer 3 ₁ , 3 ₂ , 3 ₃ , a window W _j (j = 1, 2,
The unblank signal T _j is output from the wave height analyzer 3 _j corresponding to 3) to the corresponding image memory 4 _j . For example, at the first projection angle, pixel (x, y) (x
Value, y value gamma rays are incident on any), the crest of the energy signal E is when to have entered a first window W _1, unblank signal T ₁ from the first pulse height analyzer 3 ₁ a first is input to the image memory 4 _1, this in the image memory 4 ₁ x coordinate signal X, data at the address specified by the y-coordinate signal Y is plus 1.

At the i-th projection angle, when data collection for a predetermined time (for example, 30 sec: several tens to several hundreds in terms of the number of counts per pixel) is completed, each pixel of each image memory 4 ₁ , 4 ₂ , 4 ₃ ( x, y) Counts at each window W ₁ , W ₂ , W ₃ stored in (x, y) C _1i (x, y), C _2i (x, y), C _3i
(X, y) is fixed.

On the other hand, in parallel with this data collection, the first wave height analyzer
3 ₁ unblank signal T ₁ of the from is input to the first adder 5 _1, the number of counts in the first window W ₁ is gradually being sequentially accumulated, total count number at the time the data collection is complete N
_1i becomes N _1i = ∬C _1i (x, y) dxdy. Similarly, the total number of counts N _2i at the completion of data collection by the second adder 5 ₂ is N
_2i = ∬C _2i (x, y) dxdy, and the total count number N _3i in the third adder 5 ₃ becomes N _3i = ∬C _3i (x, y) dxdy.

When the data collection is completed, the first to third adders
The signals of the total counts N _1i , N _2i and N _3i of 5 ₁ , 5 ₂ and 5 ₃ are output to the _fourth adder 5 ₄ and the sum of the total counts N _Si = N _1i + N _2i + N
_3i is calculated, and the sum of the total counts N _Si and the third
Total count number N of the _third adder 5 ₃ corresponding to the window W ₃
Each signal of _3i is input to the divider 6, and in the divider 6,
The ratio r _3i is Is calculated by The signal at ratio r _3i specifies the address in RAM 7.

The RAM 7, previously, the weighting factors for each window W _1, W _2, W ₃
k _1i , k _2i , k _3i are stored, the weighting factors k _1i , k _2i , k _3i are read out by the address signal r _3i and transferred to the weighted addition processing unit 8.

On the other hand, in each of the image memories 4 ₁ , 4 ₂ and 4 ₃ , the address is sequentially updated by the address generating circuit 9, and each pixel (x,
The data of the count numbers C _1i (x, y), C _2i (x, y) and C _3i (x, y) of y) are transferred to the weighted addition processing unit 8.

The weighted addition processing unit 8 receives the data of the count numbers C _1i (x, y), C _2i (x, y), C _3i (x, y), which are sequentially transmitted, and the RAM 7
Based on the weighting factors k _1i , k _2i , k _3i read from
The formula is calculated to calculate the corrected count number C _i (x, y) for the pixel (x, y), and the weighted addition processing unit 8
The data is stored in the address corresponding to the pixel (x, y) in the RAM 8a built in.

By performing such processing for all pixels (x, y), the internal RAM 8a of the weighted addition processing unit 8 stores i
The data of the final projection image at the second (first is the first) projection angle is stored.

When this storage is completed, each image memory 4 ₁ , 4 ₂ , 4 ₃ ,
The adders 5 ₁ , 5 ₂ , 5 ₃ , 5 ₄ and the divider 6 are cleared, and the same processing as above is performed for the next projection angle (i + 1). By repeating such processing, the data of the corrected count number C _i (x, y) for each pixel (x, y) is added to the weighted addition processing unit for all projection angles i = 1, 2, ... 8 are stored in the built-in RAM 8a.

When this is completed, the data of all the corrected count numbers C _i (x, y) in the built-in RAM 8a are transferred to the image reconstruction processing unit 1.
The image is transferred to 0, and the image reconstruction processing unit 10 reconstructs a tomographic image based on each projection image data.

In FIG. 1, reference numerals 12 ₁ and 12 ₂ are A / D converters.

The number of windows, the lower and upper limits of each window, and the value of the weighting factor for each window must be set appropriately according to the nuclide used and the collimator.

The present invention can be variously modified without departing from the scope of the claims.

For example, in the above embodiment, the number of windows is three for one energy level, but the present invention is not limited to this, and the number of windows may be any number of two or more. Further, the window widths do not necessarily have to be the same.

In addition, in shooting with a single-energy nuclide,
Instead of the three wave height analyzers 3 ₁ , 3 ₂ , 3 ₃ of the above-mentioned embodiment, three wave height analyzers which are conventionally provided as standard equipment in the gamma camera 1 for multi-energy measurement may be used. . In this case, put the windows of each wave height analyzer adjacent to each other,
It is necessary to make an adjustment so that the upper level and the lower level of each other coincide with each other.

Further, in the above embodiment, a single energy nuclide was used as a gamma ray emitting nuclide, but the present invention is not limited to this, and when a multi-energy gamma ray emitting nuclide is used or when a plurality of nuclides having different energy levels are used simultaneously. Even when taking a picture, a wave height analyzer set, an image memory set, and a device group set for calculating the ratio r _3i are provided for each energy level, and two or more windows are provided for each energy level. Then, the same process as described above may be performed.

Further, as in the case of the conventional example, the gamma camera 1 may be of a type in which a wave height analyzer is included or may not be included.

G. Effects of the Invention According to this invention, the following effects are exhibited.

In advance, find the ratio of the total counts in the windows on the higher level side than the peak wave height to the sum of the total counts for all windows, and based on this ratio, store the weighting factor for each window. There is.

At the time of actual measurement, the energy signal obtained by the gamma camera is divided into each window by the energy signal wave height analyzer having a plurality of windows for each projection angle, and the total energy in the windows on the higher level side than the peak wave height value is divided. The ratio of the number of counts to the sum of the total number of counts is obtained, and the weighting factor for each window at the projection angle is read from the storage means.
The final projection image is obtained by multiplying the image data (count number of each pixel) of each window by the weighting coefficient of each window and adding the results.

And, since the weighting factor is common to all pixels at the same projection angle,
As long as the count of each pixel is known, no matter how small the count is, the weighted addition process described above makes it possible to enhance the scattered radiation correction regardless of the rotation position of the gamma camera, the shape of the subject, and changes in the target organ. Can be done with precision. Further, since the storage means stores the weighting coefficient for each window which is determined to be constant or the sum of the corrected counts after the addition processing, absorption correction can be performed at the same time. By performing image reconstruction based on the projection data corrected in this way, a high-quality tomographic image can be obtained.

[Brief description of drawings]

1 to 5 relate to an embodiment of the present invention.
FIG. 2 is a block diagram of a scattered radiation correction device of a camera rotation type single photon ECT device. FIG. 2A is an explanatory diagram of a state in which a gamma camera is positioned right above an object.
(B) of the figure is an explanatory view of a state in which the gamma camera is positioned directly beside the subject, (A) of FIG. 3 is an energy spectrum diagram of gamma rays corresponding to (A) of FIG. 2, and FIG. (B) is an energy spectrum diagram of gamma rays corresponding to (B) of FIG. 2, FIG. 4 is a characteristic curve diagram of the weighting coefficient of each window at each projection angle, and FIG. 5 is a gamma camera for each gamma ray. 3 is a time chart showing a state in which each signal from (1) changes. Further, FIG. 6 is a spectrum diagram of a nuclide having a single energy level in the symmetric window method, and FIG. 7 is a spectrum diagram in the asymmetric window method. W ₁ , W ₂ , W ₃ ...... Window 3 ₁ , 3 ₂ , 3 ₃ ...... Wave height analyzer 4 ₁ , 4 ₂ , 4 ₃ ...... Image memory 5 ₁ , 5 ₂ , 5 ₃ , 5 ₄ ...... Addition Unit 6 ... Divider 7 ... RAM for storing weighting factor 8 ... Weighted addition processing unit 10 ... Image reconstruction processing unit

Claims (1)

[Claims] 1. A gamma ray energy level of 1 to be measured.
An energy signal height analyzer with multiple energy spectrum windows for each individual, a means for collecting image data separately for each window, a means for obtaining the total number of counts of the images obtained for each window, and a peak Means for calculating the ratio of the total count number for the windows on the higher level side of the crest value to the sum of the total count numbers for each window, and means for pre-storing the weighting factor for each window using this ratio as a parameter, Weighted addition processing means for multiplying the image data for each window by a weighting coefficient for each window read from the storage means based on the ratio in units of projection image and adding the result to obtain a final projection image. And each window stored in the storage means Weighting factors, camera rotary ECT and wherein the total sum of the corrected count after addition process is defined to be constant.