JPS63247870A - X-ray picture processing method - Google Patents

Abstract

PURPOSE:To improve the picture processing accuracy by using a quadratic to secure the approximation between the contents of two types of reference materials having the same intensity of the X rays transmitted through a subject and obtaining the equivalent contents of the reference material of the subject from an intersecting point of said approximation. CONSTITUTION:The picture data having the low and high effective energy levels are stored in the picture memories 21a and 21b respectively. The picture data reading devices 22a and 22b produce the equi-transmittance curves for both picture data and an intersecting point calculating device 23 obtains the root of a quadric equation showing the two equi-transmittance curves obtained for each point of a picture. Thus the intersecting point between said equi- transmittance curves. Then each coordinate value is corrected by the picture correcting devices 24a and 24b and sent to the picture output devices 26a and 26b via the picture memories 25a and 25b respectively.

Description

Detailed Description of the Invention (Industrial Application Field) The present invention processes an X-ray image taken using two types of X-rays with different energies to determine the equivalent content of two types of reference substances constituting an object. The present invention relates to an X-ray image processing method for determining the amount.

(Prior art) Dual energy X-ray imaging methods, which attempt to determine the composition of an object by utilizing differences in the X-ray transmission spectrum due to the elemental composition of the object, have been attracting attention as computer image processing has become commonplace. Ta. In general, when a substance is irradiated with two types of X-rays with different effective energies, the X-ray transmittance is the equivalent content of each component when the substance is equivalently equal to the sum of two types of reference substances. It can be expressed as the X-ray transmittance of a reference material at an energy of . Therefore, by image processing two X-ray images obtained by photographing an object with two different effective energies, we can calculate the The equivalent content of each component can be determined, and this method is called dual energy X-ray imaging.

For example, two X-ray images of a human body are taken using high-energy X-rays and low-energy X-rays, and
Methods for obtaining a bone image and a soft tissue, ie, meat, image from these two X-ray images include (a) a method of subtracting both X-ray images; and (b) a method of subtracting both X-ray images; A known method is to express each content as a high-order polynomial and use a computer to find a numerical solution to the simultaneous high-order polynomial. However, when using X-rays with a wide spectrum generated from a normal X-ray generator, the method (a) above does not take into account beam hardening of the X-rays, resulting in poor results. This method has drawbacks such as poor accuracy, poor separation of bone and soft tissue, and difficulty in quantifying the content of each component. On the other hand, in the method (b) above, the approximation accuracy can be increased as necessary by increasing the order of the high-order polynomial, and it is also possible to quantify the component content, but the calculation takes time, It wasn't practical.

Therefore, considering the shortcomings of the above method, we decided to use Keh-3l+in.
A method called the isotransparent straight line method was proposed by Chuang et al. This method connects points on a plane where the contents of two reference materials are the two coordinates so that the intensities of X-rays transmitted through materials with such components are equal (equal-transmittance straight lines). ), and this isotransparency line is the
It can be determined experimentally by examining the radiation transmission. When the intensity of the X-rays transmitted through the target part of the object is a certain value, the equivalent content of each component of the reference material in the target part is that X It is expressed as a single point somewhere on the isotransparent straight line of line intensity. Two iso-transparent lines are obtained from the X-ray images with two different effective energies, and the iso-transparent lines have different slopes.

The point that represents the equivalent content of each component of the reference material in that part of the object must be located on either of these two equal transmittance lines, so by finding the intersection of the two equal transmittance straight lines, The method is such that the coordinates become the transparent content of each component of the reference material in that part of the object.

However, the calculation accuracy of this equitransparent straight line method is better than the simple subtraction method in (a) above, but is not necessarily sufficient. In particular, we applied a single imaging method in which two xm images with two different effective energies are obtained in one X-ray exposure using an energy-selectable detector, for example, two sets of detectors sandwiching an energy separation filter. In this case, it did not work well when the difference in effective energy between the two images was small and there was a large change in the thickness of the object in the image.

(Objects and Structure of the Invention) The present invention has been made in view of the above points, and provides a dual energy For the purpose of providing an X-ray image processing method for a radiation image forming method,
In order to achieve this 11th objective, we approximate the relationship between the contents of two types of reference substances that make the intensity of the transmitted X-rays equal, and calculate the equivalent content of the reference substances of the subject from the intersection point. I asked for it.

(Example) The present invention will be explained below based on the drawings.

Prior to the explanation, the principle of the X-ray image processing method according to the present invention will be explained. The formula expressing the relationship between the contents of two types of reference substances such that the intensity of the X-rays transmitted through the object is equal is: It is expressed as a line on a plane when the content is taken as two coordinates, and this straight line is called the iso-permeability line. The 0-equipermeance line theoretically differs depending on the spectrum of x & a, so Is this a curve that cannot be expressed analytically?One final IJ is to approximate this curve with a quadratic curve where XA=AXB2+BXB+C (XS is the content of one reference substance It was found that the accuracy is sufficient for practical use and calculation is easy.

Therefore, a combination of the contents of the two types of reference substances constituting the object is created, and the coefficients A, B, and C of the above-mentioned quadratic equation are determined for each thickness by the method of least squares or the like. This process is performed for a predetermined number of combinations of different thicknesses (in the case of an X-ray photograph, corresponds to density), and the coefficient A obtained for each thickness is
Memorize B and C. On the other hand, by photographing the subject using two different effective energies (using the already calculated coefficients, two approximation formulas corresponding to the X-ray intensity at the same point on the subject for the two transferred X-ray images) Solve to find the content of the two types of reference substances, the content of the reference substance XA and x8.The content ff1xA thus obtained for all the points constituting the subject.

If xn is used separately, the content of each reference substance can be calculated by CRT.
It can be displayed on a display or printed on film.

Next, the X-ray image processing method according to the present invention will be explained.

First, a method for obtaining two types of X-ray images with different effective energies used in the dual-energy X-ray image forming method, which is also used in the present invention, will be explained.

In principle, the same subject can be photographed twice by changing the tube voltage of the X-ray generator, but if the subject is a human body, such as in the case of diagnostic purposes, the subject may change between the two photographs. It may move and you won't get good results, so to avoid this, do x! It is possible to improve the tube voltage of the a generator so that it can be changed rapidly. Specifically, if the detector is an image intensifier/TV camera system, it is possible to change the tube voltage by scanning the TV camera and switching the tube voltage. If the detector is of a type that once forms a latent image, such as a screen/film system or one that forms a latent image on a stimulable phosphor and then reads the latent image by emitting stimulable light with a laser, it is necessary to switch the tube voltage. It is necessary to synchronize the detectors so that they can be replaced quickly. In addition, in the method of obtaining images by scanning a fan-shaped X-ray beam and a line-shaped detector, in addition to the method of switching the tube voltage for each image, it is also possible to switch the tube voltage for each line.In this way, two images can be obtained. The disadvantages are that the effective energy difference between the images becomes large, that the movement of the subject cannot be completely eliminated, and that the equipment is complex and expensive.

Therefore, as shown in FIG. 1, an energy separation film 3 is interposed between the two detectors 1.2, and X-rays are projected from the X-ray source 4 onto the subject 5.

Two energy images with two different effective energies can be obtained by one X-ray exposure, and this method is called a single imaging method. In this method, a curved detector 1 detects a low-energy part of the X-rays, an intermediate energy separation filter 3 cuts off the low-energy X-rays, and a rear detector 2 detects a high-energy part of the X-rays. By detecting a single exposure, the front detector 1 produces a low-energy
Both line images are obtained. As the detector 1.2, a screen/film system, a device that records a latent image on a stimulable phosphor and reads the latent image by causing the photoluminescence to emit light with a laser, a semiconductor detector with a scintillator, etc. are used. In this method, there is no movement of the subject in principle, and the apparatus is simple, but it has the disadvantage that the difference in effective energy between the two images is small.

Figure 2 shows the approximate quadratic formula XA=AXB2 used in the method of the present invention.
An example of a calibration step wedge used to obtain the coefficients of +BXB+C is shown.

Step wedges 6 and 7 with stepwise changes in thickness as shown are made by selecting an easy-to-handle material that has an effective atomic number and density close to those of the two reference materials that make up the subject. Step wedges arranged so that the longitudinal directions of the steps are perpendicular to each other are used as step wedges for calibration. If the subject is a human body, consider the bones and flesh that make up the human body as reference materials, and select aluminum and acrylic resin as easy-to-handle materials with effective atomic numbers and densities close to those of bones and flesh to construct the step wedge. .

The maximum thickness of each reference material is preferably thicker than the maximum thickness of each component of the object to be photographed, and the thickness interval may be approximately equal to 10 equal parts of the maximum thickness of each component of the object. For example, in the case of chest X-ray photography, 10 stages of aluminum 5■■, 10 stages of 50mm acrylic resin 30mm, a total of 3
It is best to set it to about 0c■.

In the step wedge manufactured in this way,
The content of the reference material is usually expressed as the thickness of the reference material, or may be expressed as the mass per unit cross-sectional area in the direction perpendicular to the X-ray flux (indicated by a white arrow in the figure).

A step wedge for calibration is placed at the position of the subject 5 in the fJS1 diagram, and a low-energy X-ray image is obtained on the front detector 1 and a high-energy X-ray image is obtained on the rear detector 2 using the single-shot method.

When a film is used as the detector 1.2, images with different densities can be obtained on the X-ray film that is the detector.

Next, the procedure for determining the coefficients A, B, and C of the approximate equation from the two types of X-ray images obtained in this manner will be explained using FIG.

In the figure, 10 has the configuration shown in Fig. 1, and a digital image input for digitizing the analog image obtained by exposing x6 once to the step wedge for calibration as shown in Fig. 2 and using the detector. Various methods of digitization are known, such as a method of reading the film density with a CCD using a film scanner, or a method of measuring the amount of transmitted light with a photomultiplier tube as a laser light source. In this case, the imaging conditions, such as the effective energy and dose of X-rays irradiated with the calibration step wedge, are the same as those for imaging an actual subject (for example, a human body). Figure 4 shows the density of the analog image obtained using an aluminum step wedge whose thickness was changed by 10 m5 as a calibration step wedge and an acrylic SL step wedge whose thickness was changed by 20 m5. As shown in the table, the image consists of a combination of the thicknesses of the two step wedges. The digitized image signal is stored in the image memory 11 as image data.

In the image analysis device 12, the image data stored in the image memory 11 is read out and analyzed to obtain image signal values for each thickness combination. If the density gradation required now is 100 gradations, the image data stored in the image memory 1.1 does not provide image data for each gradation, so the data interpolation device 13 interpolates image data for the missing density gradation. Required by law. For example, concentration 1
When attempting to find the combination of thicknesses of the reference materials (aluminum and acrylic resin) that result in a concentration of 1.5 by interpolation, as shown in the table of FIG. 4 and FIG.
7. Mark points A, B, C, D, E, F in 1.8,
Proportional distribution is used between adjacent points (for example, points A and B) to find a point P for a concentration of 1.5, and the thickness of the aluminum at this point P is 7.51 and the thickness of the acrylic resin is 40-1. Get a combination with. Thus, in addition to point P, Q
, R, S, and T.

Next, in the approximation curve creation device 14, each combination of thickness for a predetermined number of density gradations is used to calculate the coefficient A of the approximation formula XA=AXfi2+BX, +C of the equal transmittance line.

Determine B and C. Therefore, there are as many combinations of coefficients A, B, and C as there are density gradations. Coefficients A and B.

When determining C, it is preferable to use the least squares method,
For example, assuming that the acrylic resin content is XA and the aluminum content j4 is XR, points P and Q are obtained by interpolation for the above-mentioned concentration of 1.5.

Using the combination of thicknesses for R, S, and T, the coefficients A, B,
When calculating C, the coefficient A-0,0140, B=-2,60
, C=59.7. Coefficients A and B thus obtained
, C are stored in the coefficient storage bag Zt l 5 a.

The above-described operation is performed for the low effective energy X-ray image obtained by the front detector 1 and the high effective energy X-ray image obtained by the rear detector 2, and the coefficients A, B, and B for each density gradation are calculated. C is determined and stored in coefficient storage devices 15a and 15b, respectively. This completes the preparations for image processing.

Next, processing of an image obtained by irradiating an actual object with xm will be explained using FIG. 6.

In FIG. 6, 20 is the same digital image input device as the IO in FIG. 3, and by irradiating the subject with X-rays using the configuration shown in FIG. Obtain X-ray images, digitize each, and output image data. The image data thus obtained are stored in the image memory 21a for low effective energy and in the image memory 21b for high effective energy.

Next, the image data reading devices 22a and 22b sequentially read out data corresponding to the same point on the subject from the density-corresponding image data stored in the image memories 21a and 21b, and calculate coefficients A, B, and C for each image data. memory bag′
i! 115a and 15b to create equal transmittance curves. As a result, two equal transmittance curves (approximate quadratic curves) as shown in FIG. 7 are obtained corresponding to the same point in each of the low effective energy X-ray image and the high effective energy X-ray image. In , L is an isotransmittance curve for low effective energy, and H is an isotransmittance curve for high effective energy.

The intersection calculation device 23 calculates the intersection of two trees of equal transmittance by finding the root of a quadratic equation representing two equal transmittance curves obtained for each point on the image. There are two roots of the quadratic equation, but in reality there is only one in the range of possible reference substance contents, and the coordinates of one of the roots and the root of the isopermeability curve are The other coordinate obtained by substitution corresponds to the equivalent content of the two types of reference substances in the subject. That is, in FIG. 7, the X coordinate at the intersection of two equal transmittance curves H corresponds to the thickness of the bone at a certain point, and the Y coordinate 2 corresponds to the thickness of the flesh at that point. All predetermined X coordinate values and all Y values of the X-ray image
The coordinate values are then appropriately corrected separately by the image correction devices t 24 a E and t 24 b so that the image is easily viewed.

This includes, for example, reversing black and white, or making thick bones appear whitish.

The equal image component images corrected in this way are stored in the image memories 25a and 25b separately for reference substances (bone and meat). The isometric component images are separately converted into analog values and then viewed by the image output devices 26a, 26b, such as by being displayed on a CRT display or printed on film.

By the way, when the size of the X-ray image obtained by photographing the subject is large, the coefficient A of the approximate quadratic curve is calculated for all points of the image using the method described above.

It takes time to calculate B and C.

Therefore, for all possible combinations of the two types of X-ray intensities with different effective energies transmitted through the subject, the equivalent contents of the two types of reference materials are determined in advance, and appropriate modifications are made to make them easier to recognize as images, such as black and white. If you calculate the results of inversion and gradation correction and create a table showing the relationship between the combinations of X-ray intensities and the results, you can use the table to determine the content of the actual subject. By searching for a combination of X-ray intensities corresponding to the same point in the X-ray images, two equal-area component images of the subject can be easily obtained. Figures 8 and 9 are examples considered from this perspective, with Figure 8 being a block diagram for creating the above-mentioned conversion table, and Figure 9 showing an example obtained by photographing an actual subject. FIG. 2 is a block diagram of an image processing device that processes two X-ray images obtained by the present invention. In both figures, the same reference numerals as in FIGS. 3 and 6 indicate rotating components.

To create the conversion table, create isotransmittance curves for each of a predetermined number of density gradations of the low effective energy and high effective energy □X-ray images of the calibration step wedge as described in connection with FIG. After determining the coefficients A, B, and C of the approximate quadratic curve represented, the intersection point calculation device 23 calculates the intersection points for all combinations of equal transmittance curves. The X coordinate of the intersection point calculated in this way,
The image correction device 24 performs image correction such as black and white inversion and gradation correction for the Y coordinate, and the corrected image data is
It is stored in the conversion table storage device 2127 as a conversion table in relation to the line intensity.

Next is the actual subject x! ! The procedure for searching the conversion table for coefficients A, B, and C for an image to obtain two equivalent component images is the same as that described with respect to FIG. 6, so the explanation will be omitted.

The above is the explanation of the image processing method according to the present invention. FIG. 10 shows a comparison of the calculation error of the content of the reference substance according to the method according to the present invention with the conventional subtraction method and the isotransparent straight line method. The example shown here uses aluminum and acrylic resin of known thickness at 70kVp and 100kVp.
A photograph was taken at Vp, and the content 1i) was calculated using this method and a known method using aluminum and acrylic resin as reference materials. The results are compared with the actual thickness of the subject, and the calculation error of the aluminum thickness is shown as the result of measuring the thickness of various acrylic resins attached when the thickness of the aluminum is Oc■. , From this figure, we can see that the method according to the present invention (solid line), which processes the iso-transmittance curve by approximating it with a quadratic curve, is compared with the conventional simple subtraction method (dashed line) and the iso-transmittance straight line method (point! 5I). It can be seen that the error is extremely small.

In addition, when comparing the calculation means for determining the intersection between the method of the present invention and the conventional subtraction method and isotransparent straight line method, it is found that for two reference materials, the subtraction method performs two additions and subtractions, two multiplications and divisions, In the case of the equitransparent straight line method, addition and subtraction are performed three times and multiplication and division are performed two times, whereas in the case of the present invention, it is more complicated with seven additions and subtractions, eight multiplications and divisions, and one square root operation, but the numerical value of the equation is Unlike iterative calculations such as solving, this calculation is relatively easy, and there is no need to worry about convergence, so it is not a problem in practice. Furthermore, as explained in Figures 8 and 9, the relationship between all the combinations of the two types of X-ray intensities with different effective energies transmitted through the object and the equivalent content of the two types of reference substances is determined in advance. When calculating the actual content of an object, use the table to find the combination of xkm intensities of two X-ray images corresponding to the same point on the object. If we try to find the equivalent content of two types of reference substances,
Once a table is created, the same table can be used for X-ray images obtained with the same effective energy combination, so the difference in calculation time per subject compared to other methods is much smaller.

(Effects of the Invention) As explained above, in the present invention, the relationship between the content of two types of reference substances that makes the intensity of the X-rays transmitted through the object equal is approximated by a quadratic equation, and from the intersection point of the Since the equivalent content of the reference substance is determined, the dual energy X-ray imaging method can be used with relatively high accuracy even when using two X-ray images with a relatively small difference in effective energy, as in the single-shot method. Equivalent contents of two types of reference substances are determined.

Moreover, if two X-ray images with a large difference in effective energy obtained by a method other than the single imaging method are used, even higher precision of the equal content can be obtained.

[Brief explanation of the drawing]

Fig. 1 is a schematic configuration diagram for the single-shot method, which is a method of obtaining two types of X-ray images with different effective energies, and Fig. 2 is a perspective view showing an example of a step wedge for calibration used in the method of the present invention. , FIG. 3 is a block diagram showing the procedure for determining coefficients in the image processing method according to the present invention, FIG. 4 is a table showing density values of the step wedge for calibration, and FIG. 5 is a diagram showing the interpolation method used in the method of the present invention. TheoryIJJ
6 is a block diagram of an apparatus for carrying out the image processing method according to the present invention, FIG. 7 is a diagram explaining a method for determining the intersection of two types of equal transmittance curves, and FIG. 8 is a diagram according to Ill. A block diagram showing the procedure for creating a conversion table used in the image processing method, FIG. 9 is a block diagram of a device that implements the image processing method according to the present invention using the conversion table, and FIG. 1O shows calculation errors due to the method of the present invention. This is a graph showing a comparison between the conventional method and the conventional method. 10.20...Digital image input device, 11゜21a
, 21b... image memory, 12... image analysis device,
13...Data interpolation device, 14...Approximate curve creation device, 15a, 15b-coefficient storage device, 22a. 22b... Image data reading device, 23... Intersection calculation device, 24a, 24b... Image correction device, 25a
, 25b---Image memory, 26a, 26b--Image output device Patent applicant Konishi Roku Photo Industry Co., Ltd. Agent Patent attorney Hiroshi Suzuki Male No. 80 Figure 9 6a

Claims (3)

[Claims] (1) Image processing is performed on two X-ray images obtained by recording two types of X-rays with different effective energies that have passed through the object to determine the equivalent content of the two types of reference substances that make up the object. In the X-ray image processing method,
Coefficient when approximating the relationship between the contents of two types of reference substances such that the line intensities are equal by X_A=AX_B^2+BX_B+C (X_A is the content of one reference substance, X_B is the content of the other reference substance) A, B, and C are obtained for a plurality of transmitted X-ray intensities predetermined for the two different effective energies, and the same point of the subject in two X-ray images each having the same effective energy as the two effective energies of the subject is calculated. X
An X-ray image processing method, characterized in that the equivalent content of two types of reference substances at the point on the object is determined by solving two approximate equations corresponding to the ray intensity for X_A and X_B. (2) The X-rays according to claim 1, wherein the two X-ray images with two different effective energies are images obtained by one X-ray exposure by a detector capable of selecting energy. Image processing method. (3) Find in advance the relationship between all possible combinations of X-ray intensities of two different effective energies transmitted through the object and the equivalent content of two types of reference substances, create a table, and use the table to determine the relationship between the X according to claim 1, in which the equivalent content of two types of reference substances at a point on an object is determined by searching for a combination of X-ray intensities of two X-ray images corresponding to the same point. Line image processing method.