JPH01232938A - Subtraction method of radiograph - Google Patents

Abstract

PURPOSE:To perform interpolation operation at a high speed, by temporarily storing the signal value relating to an actually present pixel already used in interpolation operation in a memory and reading the same from the memory when the interpolation operation of the signal value relating to the positioning pixel adjacent to said pixel is performed next to utilize the same. CONSTITUTION:The gathering of positioning pixels D' becomes one wherein the gathering of actually present pixels D is parallelly moved or rotated as it is and, therefore, for example, four actually present pixels Dn<1>, Dn<2>, Dn<3>, Dn<4>, surrounding the positioning pixel Dn become common to four actually present pixels Dn+1<1>, Dn+1<2>, Dn+1<3>, Dn+1<4> surrounding the positioning pixel Dn+1, by two pixels in extremely many cases. A plurality of image signals Sn<1>, Sn<2>, Sn<3>, Sn<4> used at the interpolation operation of the signal Sn' in the positioning pixel Dn' are temporarily stored in a memory and, when same of them are reutilized in the interpolation operation of the image signal Sn+1' in the positioning pixel Dn+1, the reading number of times of the image signal relating to the actually present pixel from an image film are reduced to a large extent.

Description

DETAILED DESCRIPTION OF THE INVENTION (Industrial Application Field) The present invention relates to a subtraction method for radiographic images, and more particularly, when there is a positional shift between two radiographic images to be subjected to subtraction, this positional shift is corrected by arithmetic processing. This invention relates to a subtraction method that enables high-speed correction.

(Prior Art) Subtraction processing of radiographic images has been known for some time. This radiographic image subtraction is a process in which two radiographic images taken under different conditions are photoelectrically read out to obtain digital image signals, and then these digital image signals are subtracted by making each pixel of both images correspond. ,
This is a method of obtaining a difference signal that extracts a specific structure in a radiation image, and by using the difference signal obtained in this way, it is possible to reproduce a radiation image in which only the specific structure has been extracted.

Broadly speaking, subtraction processing can be divided into two parts: From an image signal representing a radiographic image (live image) in which specific structures of the subject are emphasized by contrast agent injection, a radiographic image of the subject without contrast agent injection (mask image) is generated. So-called temporal subtraction, which extracts a specific structure by subtracting the image signals shown in the image, and temporal subtraction, which extracts a specific structure by subtracting the image signals shown, and a method that extracts a specific structure by subtracting the image signals shown in the image. It can be divided into so-called energy subtraction, which performs the above-mentioned subtraction.

The image signal used for the above-mentioned subtraction can be obtained by photoelectrically reading a radiation image recorded on an X-ray photographic film.
No. 429, No. 55-116340, No. 55-1634
No. 72, No. 56-11395, No. 56-104645
In the radiation image information recording and reproducing system using a stimulable phosphor sheet shown in the above issue, the image signal carrying the radiation image can be obtained directly without the need for processing such as film development. It is extremely convenient to record and read radiographic images for subtraction, that is, the mask image and the live image.

This system will be briefly explained below.

When certain types of phosphors are irradiated with radiation (X-rays, α-rays, β-rays, γ-rays, ultraviolet rays, electron beams, etc.), some of the energy of this radiation is stored in the phosphor, and then the phosphor is When irradiated with excitation light such as visible light, the phosphor exhibits stimulated luminescence depending on the accumulated energy. A phosphor exhibiting such properties is called a stimulable phosphor (stimulable phosphor).

The above system uses this stimulable phosphor to record radiographic image information of a subject such as a human body on a stimulable phosphor sheet (hereinafter referred to as stimulable phosphor sheet), and then uses excitation light to record the radiation image information of a subject such as a human body. Scanning is performed to cause stimulated luminescence, and this stimulated luminescence is read photoelectrically to obtain an image signal.

By the way, the two radiation images to be subjected to subtraction cannot necessarily be recorded at exactly the same position on the recording medium such as the above-mentioned stimulable phosphor sheet, and there may be a positional shift or misalignment between the two images.
That is, rotational deviations and translational deviations in vertical and horizontal directions often occur. If the aforementioned image signal subtraction is performed between corresponding pixels of the image recording medium while such positional deviation exists, subtraction will be performed between different parts of the object, and as a result, Problems arise, such as parts of the subtraction image that should be erased not being erased, or conversely, parts that should be extracted being erased.

Therefore, as shown in, for example, Japanese Patent Application Laid-Open No. 58-163338, a method has been considered in which the above-mentioned positional deviation is corrected by arithmetic processing. In this method, when there is a positional shift as described above, for one set of image signals, an alignment pixel Dn' that matches the pixel position corresponding to the other set of image signals is determined, and this alignment pixel Dn'' The image signal Sn in

For example, four real pixels D surrounding this pixel Dn°
jl' s D Q 2, D jl 3, D n'
The image signals Sn' are calculated by interpolation based on the image signals Sn1, So2, So3, and So4 for .

(Problem to be Solved by the Invention) By the way, when subtraction is performed, two sets of image signals are usually stored in respective image files, so that the image signal S at the alignment pixel Dn' is
In performing interpolation calculations for n°, it is necessary to read out a predetermined number of image signals per pixel (four in the above example) one by one from the image file, and therefore the time required for calculation processing to correct this positional shift is inevitably long. I tended to.

SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide a subtraction method that can perform the arithmetic processing for correcting the positional deviation at high speed.

(Means for Solving the Problems) The radiation image subtraction method according to the present invention obtains the image signal Sn° at the alignment pixel Dn° by the above-mentioned interpolation calculation, and combines this image signal Sn° with the other set of image signals Sn°. In a subtraction method in which subtraction is performed between image signals, a plurality of image signals used in interpolating the signal Sn' at the alignment pixel Dn' are temporarily stored in a memory, If some of the plurality of real pixels surrounding the alignment pixel Do or electric field adjacent to the alignment pixel Dn' are common to some of the real pixels surrounding the alignment pixel Dn', the common real pixels The image signal stored for the alignment pixel D n+I° is read out from the memory, and the image signal S n+ at the alignment pixel D n+I° is read out from the memory.
1' is reused when performing interpolation calculations.

(Function) As shown in Fig. 4, the set of alignment pixels D' is a set of real pixels that has been completely translated or rotated, so for example, 4 pixels surrounding the alignment pixel Dn' are formed.
Two real pixels Dn'', Dn 2, Dn 3, Dn
' is very often the alignment pixel D 1
Four real pixels Dn++'s D surrounding 1+I°
n++2X Dn++3s Din' and two each are common. In particular, if there is no rotational deviation and only the translational deviation mentioned above, the actual pixel Dn ” s Dn 2, Dn
3. Dn' and real pixel Dn++' s Dnn2
, D11+13, and D11+1' always have two in common. Furthermore, when there is a rotational shift, in most cases, two of each of the actual pixels are common as described above, other four of each are common, and sometimes one of each is common.

Therefore, the plurality of image signals Sn1 and Sn2 used when performing interpolation calculation on the signal Sn' at the alignment pixel Dn'
, Sn3, and Sn4 are temporarily stored in memory,
If some of them are reused for the interpolation calculation of the image signal S n+1° at the alignment pixel 0 and 1°, the number of times the image signal for the actual pixel is read out from the image file can be significantly reduced. For example, when there is no rotational shift and only translational shift as described above, the number of times this readout is approximately 1/
It becomes 2. Of course, as the number of times the image signal is read from the image file is reduced, it becomes necessary to read the image signal from the memory, but since this latter readout can be performed in a shorter time, overall positional deviation is reduced. The time required for correction calculation processing becomes shorter.

(Example) Hereinafter, the present invention will be described in detail based on an example shown in the drawings.

FIG. 1 shows an example of a radiation image information recording and reproducing system for obtaining subtraction images by the method of the present invention. This radiation image information recording and reproducing system basically includes a radiation image capturing section 20, a pre-reading reading section 30, a main reading reading section 40, and an image reproducing section 50. In the radiographic imaging unit 20, radiation 102 is emitted from a radiation [100] such as an X-ray tube toward a subject (subject) 101 on an imaging table 105. In this case, first, a contrast medium is not injected into the subject lot.

As described above, the stimulable phosphor sheet 103M that accumulates radiation energy is placed at the position where the radiation 102 that has passed through the subject 101 is irradiated, and the stimulable phosphor sheet 103M has the stimulable phosphor sheet 103M A radiation image (mask image) is accumulated and recorded.

The stimulable phosphor sheet 103M on which the radiation image information of the subject 101 is recorded in this manner is sent to the pre-reading reading unit 30 by sheet transport means 110 such as a transport roller. The laser beam 202 emitted from the pre-reading laser light source 201 in the pre-reading reading unit 30 passes through a filter 203 that cuts the wavelength range of stimulated luminescence light emitted from the stimulable phosphor sheet 103M due to excitation of the laser beam 202. After that, the light is linearly deflected by a light deflector 204 such as a galvanometer mirror, and then reflected by a plane reflecting mirror 205.
The light is incident on the stimulable phosphor sheet 103M through the stimulable phosphor sheet 103M. The phosphor sheet 103M is transported in the direction of the arrow 20B by a sheet transport means 210 such as a transport roller to perform sub-scanning, and as a result, the entire surface of the phosphor sheet 103M is irradiated with laser light 202. Here, laser light source 2
01, the beam diameter of the laser beam 202, the scanning speed of the laser beam 202, and the transport speed of the stimulable phosphor sheet 103M. It has been selected to be smaller than that of the main reading conducted in .

When the laser beam 202 is irradiated as described above, the stimulable phosphor sheet 103M emits stimulated luminescence light with an amount of light corresponding to the radiation energy stored therein, and this luminescent light is transmitted to the pre-reading light guide. 207. The stimulated luminescence light is guided through the light guide 207, exits from the exit surface, and is received by a photodetector 208 such as a photomultiplier. The detected stimulated luminescence light is converted into an electrical signal carrying accumulated recording information and amplified by an amplifier 209.

The signal output from the amplifier 209 is sent to the A/D converter 211
The pre-read image signal Sp is digitized and input to the main reading control circuit 314 of the main reading reading section 40. This main reading control circuit 314 determines the reading gain setting value Sk, the recording scale factor setting value Gp1 and the reproduction image processing condition setting value C, based on the accumulated recording information indicated by the pre-read image signal Sp, for example, by histogram analysis or the like.

The stimulable phosphor sheet 103M, which has undergone pre-reading as described above, is transferred to the main reading reading section 40.

In the main reading reading section 40, the main reading laser light source 301
The laser beam 302 emitted from this laser beam 302
A filter 3 that cuts the wavelength range of stimulated luminescence light emitted from the stimulable phosphor sheet 103M by excitation of the stimulable phosphor sheet 103M.
03, the beam diameter is strictly adjusted by a beam expander 304, linearly deflected by an optical deflector 305 such as a galvanometer mirror, and passed through a flat reflector 30G to a stimulable phosphor sheet. 103M
incident on the top. An fθ lens 307 is disposed between the optical deflector 305 and the plane reflecting mirror 306, so that the beam diameter of the laser beam 302 scanning the stimulable phosphor sheet 103M is made uniform. On the other hand, stimulable phosphor sheet 1
03M is transported in the direction of arrow 308 by a sheet transporting means 320 such as a transport roller to perform sub-scanning, and as a result, the entire surface of the stimulable phosphor sheet 103M is irradiated with laser light. When irradiated with the laser beam 302 in this manner, the stimulable phosphor sheet 103M emits stimulated luminescence light in an amount corresponding to the radiation energy stored and recorded therein, and this luminescent light is transmitted to the main reading light guide 309. incident. The stimulated luminescence light guided through the main reading light guide 309 while repeating total reflection is emitted from its exit surface,
The light is received by a photodetector 310 such as a photomultiplier.

A photodetector 310 that photoelectrically detects the stimulated luminescent light indicating the radiation image recorded on the stimulable phosphor sheet 103M.
The output is amplified into an electrical signal at an appropriate level by the amplifier att whose read gain is set according to the read gain setting value Sk determined by the control circuit 314. The amplified electrical signal is input to the A/D converter 312, and based on the recording scale factor setting value Gp, it is converted into a digital signal sM with a recording scale factor suitable for the signal fluctuation range, and is input to the subtraction processing section 500. Ru.

After recording a radiographic image (mass image) of the subject 101 on the above-mentioned stimulable phosphor sheet 103M in the radiographic image capturing unit 20, a contrast medium is injected into a specific structure (for example, a blood vessel) of the subject 101. In the radiation image capturing section 20, the image (live image) is recorded on another stimulable phosphor sheet 103L in the same manner as described above. Note that when photographing (recording) the mask image and the live image, the radiation source 10
The tube voltages at 0 are made equal. Further, as shown in detail in FIG. 3, markers 2A and 2B made of a radiation shielding material are fixed on the imaging table 105 at appropriate intervals, and markers 2A and 2B made of a radiation shielding material are fixed on the photographing table 105, and markers 2A and 2B made of a radiation shielding material are fixed on the photographing table 105 at appropriate intervals. is the subject image 101°, and these markers 2A, 2
B images 2A' and 2B' are also recorded. When photographing a mask image and a live image, the subject 101 is placed on the photographing table 105 so as to maintain a predetermined positional relationship with respect to the markers 2A and 2B.

Therefore, due to reasons related to the mechanical precision of the mechanism that holds the stimulable phosphor sheets 103M and 103L below the imaging platform 105, the stimulable phosphor sheets 103M and 103L are
3L are held in slightly different positional relationships with respect to the photographing table 105, each sheet 1 of the subject image 101°
Although the positions on 03M and 103L are different from each other, the subject image 101° and the marker images 2A' and 2B
'The relative positional relationship with both sheets 103M5103
They are equal to each other at L.

The stimulable phosphor sheet 103L on which the live image of the subject 101 is stored and recorded as described above is also read by the pre-reading reading section 30.
Then, the information is sequentially sent to the main reading reading unit 40, and the radiation image information recorded on the sheet 103L is read in exactly the same way as for the stimulable phosphor sheet 103M. The read image signal s obtained by this reading process
L is also sent to the subtraction processing section 500.

As described above, both sheets 103 M, 103
Although a positional shift occurs between the subject images 101' actually recorded on L, the subtraction processing is performed on both sheets 103M.
, 103 Each image signal s obtained by reading processing from L
Since the reading is performed between M and sL, the main reading reading section 40 scans each sheet 103 for the scanning range of the laser beam 302.
If the positional relationship of M and 103 L is different between both sheets,
This is also a mask image when performing subtraction processing.
This appears as a positional shift in the live image. The method of the present invention is capable of correcting positional deviations caused by the two causes described above, and its mechanism will be explained in detail below. FIG. 2 shows the subtraction processing section 500 in detail. Read image signal sM
and sL are the image files 50 of the processing unit 500, respectively.
1M, recorded in 501L. During subtraction processing, these image signals sM and sL are read out from the image files 501M and 501L, and first sent to the positional deviation detection unit 502. This positional deviation detection section 5
As an example, 02 is constituted by a known computer system together with a positional deviation correction section 503 and a subtraction calculation section 504, which will be described later.
, 2B' are located in the mask image and the live image, respectively. Marker images 2A', 2B
If marker images 2A' and 2B' are at the same position in the mask image and live image, there will be no positional deviation between the mask image and the live image, but if marker images 2A' and 2B' are in different positions in the mask image and live image, If it exists, there is a positional shift between the mask image and the live image.

As mentioned above, the relative positional relationship between the marker images 2A', 2B' and the subject image 101° is constant in both the mask image and the live image, so the marker image 2A' between the mask image and the live image is , 2B' directly becomes a positional shift between the subject images 101'.

Note that marker images 2A' are obtained from image signals SM and SL, respectively.
, 2B' and the method of detecting the positional shift between the two images based on the detected positions are described in detail in the above-mentioned Japanese Patent Application Laid-Open No. 163338/1982, and the method of the present invention The positional deviation can also be detected using these methods. In this embodiment, as an example, positional deviation is defined in terms of how the live image deviates from the mask image, and information Q indicating this positional deviation is sent to the positional deviation correction unit 503.

The positional deviation correction unit 503 reads the image signal sM regarding the mask image from the image file 501M, and if the pixel position of the mask image matches that of the live image,
How this image signal sM changes is determined. That is, the correction unit 503 first, based on the information Q,
A pixel whose coordinates match the pixel coordinates of the live image, that is, an alignment pixel Dn' is determined. Here, as shown in Figure 4,
When the coordinates of one standard real pixel Dn in the X-Y orthogonal coordinate system are (Xn, Yn), this real pixel D
The coordinates (Xn, Vn) of the alignment pixel Dn° corresponding to n in the x-y orthogonal coordinate system are xn-Xn+XFn Vn-yn+YFn, and the values of XFn and YFn can be determined from the information Q above. If the distance between pixels is 1, then when there is a rotational deviation of θ between the mask image and the live image,
Alignment pixel Dn+1' adjacent to alignment pixel Dn°
The coordinates (x n++, y h@) are Xn++"
"xn + cosθ...(1)yn@""
yn +sln θ −・−−・・
(2) becomes. Therefore, if the alignment pixel Dn' corresponding to one reference real pixel Dn as described above is found, other alignment pixels can be easily found one after another based on the relationships (1) and (2) above. be able to.

Next, the signal value Sn° assumed at the alignment pixel Dn′ is divided into four real pixels Dn° surrounding the pixel Dn°.
An interpolation calculation is performed from lSDn2, Dn3, and Dn4. As an example, this interpolation calculation is performed for the pixel Dn' as shown in FIG.
and each real pixel Dn l 5Dn2, Dn3, Dn4 as two opposite corner points wn1, Vn2, V
Area of n3, xn4 (wn' +Wnz+Wn3+Wn
'-1) as a weighting coefficient, each real pixel Dn''
, Dn 2, Dn3, Dn4 signal values Sn1, Sn2,
This is done by weighted averaging of Sn3 and Sn4. That is, S'l' - Wfi ''Sn' +wn2Sn
2+ yes. 3Sl 3+w. 4 sn 4.

Here, as is clear from FIG. 4, the alignment pixel Dn"
Of the four real pixels surrounding the , the X coordinates of the two pixels on the left side in the figure are XIn, and similarly the Y coordinates of the two pixels on the top side in the figure are Yln, and when x144>XIn+1 yn++<Y 1n+ 1 = -(3) are the signal values Sn1, Sn2, Sn
3. S03 and S04 of Sn4 can be used. Moreover, as shown in FIG. 6, xr, I<XIn+1 yr1@< Y I O+ 1 --(4)
In this case, all of the above signal values Sn1 to Sn4 can be reused as the signal values of the four actual pixels to be weighted and averaged when determining the signal value S at the alignment pixel Drl@'.

Therefore, the positional deviation correction unit 503 corrects the positioning pixel Dnl.
The signal values Sn 1 to S04 of the four actual pixels used in determining the signal value Sn° of are stored one by one in the memory 505, and the signal value Sn+1' for the next alignment pixel Dn+1' is stored one by one in the memory 505.
When performing an interpolation calculation on the pixel Dn@', the above (
3) If the following relationship holds true, the image file 501M
, only the signal values Sn++3 and Sy1+1' are read out, while the signal values S03 and Sn4 read out from the memory 505 are used as the signal values S+i and Sr++12, and are subjected to the weighted averaging described above. In addition, the positional deviation correction unit 503
If the relationship (4) above holds, no signal is read out from the image file 501M, and the signal value Sn ”s Sn stored in the memory 505 is read out.
2, Syl 3, and Sn' and use them as the signal value Snn''s 5nn2%5in3s Snn' to perform the above-mentioned interpolation operation.Of course, the signal value stored in the memory 505 is only one position. It is updated sequentially each time an interpolation calculation is performed on the combined pixels.

As described above, the image signal s about the mask image
M is converted to correspond to the positional shift of the live image, and the converted image signal SM' is sent to the subtraction calculation unit 504 together with the image signal sL regarding the live image. The subtraction calculation unit 504
Upon receiving both image signals SM' and sL, subtraction of 5sub = SH' - SL is performed for each signal for corresponding pixels. The difference signal S obtained in this way
sub is sent to the image processing circuit 313, where it undergoes processing such as gradation processing and frequency processing, and then is sent to the image reproduction section 50. Note that the conditions for this image processing are determined by the setting value C determined by the control circuit 314.

The difference signal S sub is transmitted to the optical modulator 4 of the image reproducing section 50.
01 is input. In the image reproducing unit 50, a laser beam 403 from a recording laser light source 402 is modulated by an optical modulator 401 based on a difference signal Ssub input from the signal processing circuit 313, and the scanning mirror 403
A photosensitive material 405 such as photographic film is deflected by
Scan above. The photosensitive material 405 is then transported in a direction perpendicular to the scanning direction (arrow 406 direction) in synchronization with the scanning, and a radiation image based on the difference signal 5sub is output onto the photosensitive material 405. In addition to this method, various other methods can be used to reproduce the radiographic image, such as the above-mentioned CRT display.

The radiation image reproduced and recorded on the photosensitive material 405 based on the signal S sub extracts and shows only the specific structure injected with the contrast agent of the subject IO1. In this method, the mask image signal sM is converted to correct the positional deviation between the mask image and the live image before being subjected to subtraction, so that the mask image signal sM is reproduced based on the difference signal Ssub. In the subtraction image, problems such as parts that should be erased are not erased, or conversely, parts that should be extracted are erased, do not occur.

As shown in FIG. 7, when Yh+l>y I n +1, all signals related to the four real pixels D h+1' to I) rt++' may be read out from the image file 501M, or In the y direction as well, signals that have already been used in interpolation calculations may be stored in memory, and some of them may be reused in new interpolation calculations.

The embodiments in which the present invention is applied to so-called time subtraction have been described above, but the method of the present invention can also be applied to the above-mentioned energy subtraction.

(Effects of the Invention) As explained in detail above, in the subtraction method of the present invention, when performing interpolation calculations on signal values at alignment pixels in order to correct the positional shift between two images that affects subtraction, this interpolation A signal value for an actual pixel that has already been used for calculation is temporarily stored in a memory, and then the signal value is read from the memory when performing an interpolation calculation for a signal value for an alignment pixel adjacent to that pixel. Therefore, interpolation calculations can be performed at a higher speed than when a predetermined number of signal values per alignment pixel are sequentially read out from a large-capacity image file. Therefore, according to the method of the present invention, compared to the conventional method,
Subtraction processing of radiographic images can be performed in a shorter time.

[Brief explanation of the drawing]

FIG. 1 is a schematic diagram showing an example of a radiographic image information recording and reproducing system that implements the method of the present invention, FIG. 2 is a block diagram showing a subtraction processing section of the system, and FIG. 3 is a radiographic image capturing system according to the present invention. 4 is an explanatory diagram showing the relationship between actual pixels and alignment pixels according to the present invention; FIG. 5 is an explanatory diagram illustrating interpolation calculation of signal values in the alignment pixels; 6 and 7 are explanatory diagrams showing other examples of the relationship between the actual pixels and alignment pixels. 2A, 2B...Markers 2A', 2B'...Marker image 101...Subject lot'...Subject image 102...Radiation 103M, 103L...Stormable phosphor sheet 1
05... Shooting stand 500... Subtraction processing unit 501 M, 501 L... Image file 50
2... Positional deviation detection unit 503... Positional deviation correction unit 504... Subtraction calculation unit 505... Memories Dn1, Dn2, Dn3, Dn4, DIIn', Dn+12, DH+13, Dn@'
'Real pixel Dn', Dn+1'... Alignment pixel sM... Mask image signal SM'... Mask image signal sL after positional deviation correction...
・Live image signal S sub...Difference signal Figure 3

Claims (1)

[Scope of Claims] Two radiation images of a common subject having different image information of a specific structure therein are recorded on separate recording media, and the radiation images are read from each recording medium to generate respective digital image signals. and subtracting these two sets of image signals between the signals for corresponding pixels to form a difference signal for extracting the specific structure, the method comprising: If there is a misalignment with each other,
For one set of image signals, an alignment pixel D_n' that matches the pixel position corresponding to the other set of image signals is determined, and the image signal S_n' at this alignment pixel D_n' is divided into a plurality of pixels surrounding the pixel D_n'. The interpolation calculation is performed based on each image signal for the actual pixels of , and the subtraction is performed between the image signal S_n' obtained in this way and the image signal of the other set, and each image signal used in the interpolation calculation is performed. is temporarily stored in the memory, and the alignment pixel D_n adjacent to the alignment pixel D_n' is
If some of the plurality of real pixels surrounding ___+_1' are common to some of the real pixels surrounding the alignment pixel D_n', reading out the image signals stored for the common real pixels from the memory; , image signal S_n_ at alignment pixel D_n_+_1′
A radiation image subtraction method characterized in that it is used for +_1' interpolation calculation.