JP4692245B2 - Phase contrast X-ray imaging system for asbestos and phase contrast X-ray imaging method for asbestos - Google Patents

Description

The present invention relates to an asbestos phase-contrast X-ray imaging system and an asbestos phase-contrast X-ray imaging method for capturing an X-ray image by irradiating a subject with X-rays, performing image processing on the X-ray image, and outputting the processed image.

  In recent years, asbestos that has entered the human body has become a problem as it induces serious diseases such as pneumoconiosis and mesothelioma. For this reason, an examination for taking a chest X-ray image of a patient and detecting the presence or absence of asbestos in the human body from the X-ray image is performed. Detection of asbestos is expected to lead to rapid treatment (treatment, preventive measures) and suppress onset.

  However, in a general medical X-ray imaging system that images the chest, limbs, and the like, it is difficult to visualize a fine structure having a diameter of about 10 μm or less such as asbestos. Even if a visible image can be formed, the visibility is low and the doctor cannot often observe (discriminate). This is because the size of the structure to be imaged is smaller than the spatial resolution of the image detector used in the X-ray imaging system.

  In order to obtain an X-ray image capable of finely observing a minute object such as asbestos, enlargement photography is effective. In the magnified photographing, the X-ray image is enlarged more than the actual size of the subject by setting the focal diameter of the X-ray tube, the distance from the X-ray tube to the subject, and the distance from the subject to the image detector as a predetermined relationship. Is a shooting method.

  In order to further improve the visibility in magnified imaging, it is conceivable to increase the X-ray dose by increasing the magnification rate during magnified imaging or increasing the focal diameter of the X-ray tube. Conventionally, in order to capture a small animal as an object to be imaged, a method has been disclosed in which a magnified image with a magnification of 10 times is taken, and the obtained X-ray image is read at a reading sampling pitch of 25 μm to obtain digital image data (for example, , See Patent Document 1).

Also disclosed is a method of increasing the enlargement ratio by setting the distance from the subject to the image detector to 0.3 m or more (see, for example, Patent Document 2).
Japanese Patent Laid-Open No. 10-268450 Japanese National Patent Publication No. 11-502620

  However, in enlargement photographing, when the enlargement ratio and / or the focal diameter are increased, a phenomenon called blurring occurs in which the sharpness of the edge portion is reduced accordingly. Accordingly, as described above, a simple increase in the enlargement ratio aiming at improving the visibility causes blurring, which in turn causes a decrease in visibility. Also, if the focal spot diameter of the X-ray tube is large, the amount of X-ray irradiation per unit time increases, so that the density characteristics of the X-ray image can be improved, but the enlargement of the focal spot diameter also leads to the occurrence of blurring. As with the enlargement ratio, simple enlargement is not preferable.

An object of the present invention is to provide an asbestos phase contrast X-ray imaging system and an asbestos phase contrast X-ray imaging method capable of obtaining an X-ray image with high asbestos visibility.

The invention described in claim 1
An X-ray tube that irradiates the subject with X-rays;
An image detector that generates an X-ray image corresponding to an X-ray dose of X-rays transmitted through the subject;
An image processing apparatus that performs image processing on the generated X-ray image;
Have
The focal diameter of the X-ray tube is D (μm),
The distance from the focal point of the X-ray tube to the subject is R1 (m),
The distance from the subject to the image detector is R2 (m),
The distance from the focal point of the X-ray tube to the image detector is R3 (m),
Magnification factor M = R3 / R1,
Blur width B = D · R2 / R1 (μm),
And when
1 ≦ D ≦ 30,
3 ≦ R3 ≦ 5,
10 ≦ M ≦ 40,
The filling,
The image processing apparatus is a phase contrast X-ray imaging system for asbestos that performs frequency enhancement processing on the generated X-ray image in a frequency band of 500 / (10M + B) (lp / mm) or more. is there.

The invention described in claim 2
Irradiating a subject with X-rays from an X-ray tube;
Generating an X-ray image corresponding to an X-ray amount of X-rays transmitted through the subject by an image detector;
Applying image processing to the generated X-ray image by an image processing apparatus;
Including
The focal diameter of the X-ray tube is D (μm),
The distance from the focal point of the X-ray tube to the subject is R1 (m),
The distance from the subject to the image detector is R2 (m),
The distance from the focal point of the X-ray tube to the image detector is R3 (m),
Magnification factor M = R3 / R1,
Blur width B = D · R2 / R1 (μm),
And when
1 ≦ D ≦ 30,
3 ≦ R3 ≦ 5,
10 ≦ M ≦ 40,
The filling,
In the step of performing the image processing, the generated X-ray image is subjected to frequency enhancement processing in a frequency band of 500 / (10M + B) (lp / mm) or more, and phase contrast X-ray imaging for asbestos Is the method.

According to the invention described in claim 1, 2, it is possible to provide a highly magnified image sharpness corresponding to the detection of very small asbestos. Asbestos is very small, and it is difficult to visualize asbestos with a normal photographing method. Therefore, the focal diameter at the time of photographing is reduced to perform enlarged photographing with enhanced image sharpness, and further, the asbestos sharpness is enhanced by emphasizing the frequency of the spatial frequency 500 / (10M + B) or more necessary for visual recognition of asbestos. Thereby, the visibility of asbestos in an enlarged image can be improved.
Further, even in a limited shooting environment, it is possible to adjust to a relatively large enlargement ratio of 10 to 40.

Human structures are mainly composed of elements such as C, H, O, and Ca, while asbestos is mainly composed of elements such as Si, O, Fe, and Mg. The relationship between each constituent element and its X-ray absorption rate is as shown in Table 1 below.

  Asbestos is often inhaled through the mouth and present in the alveoli. The alveolar constituent elements are mainly C, H, O, N and the like, and the difference is 10 times or more when the X-ray absorption rate is compared with Si which is the main component of asbestos. Therefore, when the same X-ray dose is applied to the human body, the X-ray dose difference reaching the image detector is also large, and the contrast (density difference) on the X-ray image is also large.

  Further, since the bone part of the human body has a high Ca content with a high X-ray absorption rate, it is difficult to penetrate the human body and the X-ray dose reaching the image detector is reduced. Compared to the constituent element Si of asbestos, the X-ray absorption rate of Si is 1/3 of Ca, and it can be seen that the contrast between asbestos and alveoli is smaller than the contrast between bones and alveoli.

  In conventional X-ray images of the chest, the tone characteristics on the X-ray image are adjusted so that the bones, alveoli, and pulmonary blood vessels with large differences in X-ray absorption rates are easily read at the same time. is doing. However, under the gradation characteristics adjusted in this way, sufficient contrast cannot be obtained in an image portion having a relatively small difference in the reached X-ray dose, such as asbestos and alveoli, and the boundary portion between asbestos and alveoli It becomes impossible to distinguish.

Therefore, in the present invention, in order to obtain an X-ray image in which minute asbestos can be visually recognized, the X-ray tube used at the time of imaging is controlled to control the focal diameter and magnification, and the digital X obtained by the magnification imaging is used. An X-ray image with high asbestos visibility is generated by subjecting the line image to gradation conversion processing and / or frequency enhancement processing, which are image processing unique to digital processing according to detection of asbestos.
Hereinafter, an embodiment to which the present invention is applied will be described.

FIG. 1 shows an X-ray imaging system 100 in the present embodiment.
As shown in FIG. 1, the X-ray imaging system 100 includes an imaging device 10, an image processing device 20, a film output device 30, an image server 40 having an image DB 40a, and a display device 50. The devices 10 to 50 are connected via a network N so that they can communicate with each other. The network N is a LAN (Local Area Network) to which DICOM (Digital Imaging and Communication in Medicine) standards are applied.

  In the X-ray image system 100, when an X-ray image is captured by irradiating a subject with X-rays in the imaging apparatus 10 and digital data of the X-ray image is generated, various image processing on the X-ray image is performed on the image processing apparatus. 20 is applied. The processed image output from the image processing device 20 is stored in the image server 40 and output to the film output device 30 or output to the display device 50 in response to a request from the display device 50.

Hereinafter, each component apparatus will be described in detail.
As shown in FIG. 2, the imaging apparatus 10 includes an X-ray tube 2 and an image generation device 3, and X-rays emitted from the X-ray tube 2 toward the subject W are included in the image generation device 3. It is detected by the image detector 31, and digital data of an X-ray image corresponding to the X-ray dose is generated. At the time of photographing, enlargement photographing with an enlargement ratio M is performed by adjusting the distances R1 and R2 between the subject W and the X-ray tube 2 and between the subject W and the image detector 31.

  The X-ray tube 2 generates X-rays having a focal diameter D (μm) and irradiates the subject W. In the X-ray tube 2, the larger the focal diameter D, the larger the X-ray dose irradiated within a certain time.

The image generation device 3 includes an imaging unit 32 including an image detector 31, a main body unit 33 for performing imaging control, and the like.
The imaging unit 32 includes an image detector 31 and is configured to be able to adjust the height position according to the imaging region.

  The image detector 31 detects irradiated X-rays. As the image detector 31, a volatile phosphor plate capable of absorbing and storing X-ray energy, an FPD (Flat Panel Detector), or the like can be applied. When a volatile phosphor plate is applied, a reading unit that irradiates the luminescent phosphor plate with excitation light such as laser light and photoelectrically converts the stimulated light emitted from the phosphor plate into an image signal is photographed. It is provided in the part 32. The image signal (analog signal) generated by the reading unit is output to the main body unit 33.

  When a cassette in which a phosphor plate is accommodated in the housing is used as the image detector 31, image signal reading processing and digitization are performed using a cassette-dedicated reading device. .

  On the other hand, the FPD has a matrix of conversion elements that generate an electrical signal according to the incident X-ray dose, and differs from the phosphor plate in that an electrical signal (analog) is directly generated in the FPD. . When the FPD is applied, an electrical signal generated in the FPD is converted into a digital signal by sampling and output to the main body 33.

  The main body 33 is connected to the X-ray tube 2, an operation unit for performing a control operation of imaging operations of the X-ray tube 2 and the imaging unit 32, various signal processing such as converting an image signal into digital data, A processing unit that performs data processing, a control unit that centrally controls each unit of the image generation apparatus 3, a communication unit that communicates with other external devices, and the like are provided.

  In the main body 33, it is possible to instruct the X-ray irradiation conditions such as tube voltage and tube current in the X-ray tube 2 and the irradiation timing through the operation unit, and the control unit responds to this instruction operation. Centralized control of the operation of each unit such as the X-ray tube 2 and the imaging unit 32 is performed.

Next, enlargement photographing in the photographing apparatus 10 will be described.
FIG. 3 is a diagram for explaining the outline of the enlarged photographing.
As shown in FIG. 3, in the case of a normal imaging method, an image detector 31 is arranged at a position in contact with a subject (close contact imaging position in FIG. 3) and configured to receive X-rays emitted from the X-ray tube 2. Has been. In this case, the X-ray image is approximately the same size as the life size (which means the same size as the subject W).

  On the other hand, in the magnified shooting, the image detector 31 is arranged with a distance between the subject W and the image detector 31. The life size is increased by the X-rays irradiated from the X-ray tube 2 in a cone beam shape. An X-ray image (hereinafter referred to as “enlarged image”) that is enlarged with respect to FIG.

Here, the enlargement ratio M with respect to the life size of the enlarged image is determined by R1 as the distance from the focus a of the X-ray tube 2 to the subject W, R2 as the distance from the subject W to the image detector 31, and the focus a of the X-ray tube 2. If the distance from the image detector 31 to the image detector 31 is R3 (R3 = R1 + R2), it can be obtained by the following equation (1).
M = R3 / R1 (1)

  In the enlarged image, as shown in FIG. 4, X-rays refracted by passing through the edge of the subject W overlap with X-rays that have passed without passing through the subject W on the image detector 31, and X The line strength increases. On the other hand, a phenomenon occurs in which the X-ray intensity is weakened in the inner portion of the edge of the subject W by the amount of refracted X-rays. Therefore, an edge enhancement function (also referred to as an edge effect) in which the X-ray intensity difference is widened at the border of the subject W works, and an X-ray image with high visibility in which the border portion is sharply depicted can be obtained. .

When it is assumed that the X-ray source is a point source (that is, the focal point a is a point), the X-ray intensity at the edge portion is as shown by a solid line in FIG. E shown in FIG. 5 represents the half-value width of edge enhancement, and can be obtained by the following equation (2). The half-value width E indicates the distance between the peaks and valleys of the edge.

  However, a cooling ridge X-ray tube (also referred to as a thermoelectron X-ray tube) is widely used in medical sites and non-destructive inspection facilities, and this cooling ridge X-ray tube has a focal diameter D as shown in FIG. Since it becomes large, it cannot be regarded as an ideal point source. In this case, as shown in FIG. 6, the half-width E for edge enhancement is widened and the strength is lowered, resulting in a geometrical sharpness. This geometrical sharpness is called blur.

式中、δ及びｒの定義は、式２と同じである。
また、ＥＢはボケが無い場合のエッジ強調半値幅Ｅにボケの大きさを示すＢを加え、ＥＢ＝Ｅ＋Ｂで示される。 When the blur occurs, the X-ray intensity at the edge portion is as shown by a dotted line in FIG. The half width of edge emphasis when blurring occurs is wider than the edge emphasis width E when an ideal point source is assumed due to geometrical sharpness. If the half width of edge emphasis when this blur occurs is EB, EB can be obtained from the following equation (3).

In the formula, the definitions of δ and r are the same as those in Formula 2.
EB is represented by EB = E + B by adding B indicating the size of the blur to the edge emphasis half width E when there is no blur.

  Here, in order to improve the visibility of a minute object having a diameter of 10 μm or less, such as asbestos, it is necessary to increase the enlargement factor M. In order to increase the enlargement ratio M, the distance R2 may be increased from Equation 1, but an increase in the distance R2 causes an increase in the half-value width EB of the blur. Here, assuming that the diameter of asbestos is s (μm), the diameter s is a deformed body such as a thread-like elongated body whose diameter is a circumscribed circle when the asbestos is not a deformed body such as a substantially spherical shape or a substantially cubic shape. In this case, it means the diameter of the cross section in the direction orthogonal to the extending direction (elongated direction) of the deformed body.

  When adjusting the enlargement factor M, the enlargement factor M can be varied by fixing the distance R1 and increasing / decreasing the distance R2. However, if the distance R1 is set too large, it is inappropriate for actual photographing. It may be a distance setting. For example, when the magnification ratio M = 20 (times), if the distance R1 = 1 (m), the distance R2 must be set to 19 (m), but such a setting is not practical in a normal shooting room. .

  On the other hand, if the setting of the distance R1 is made small, the irradiation field becomes small and only a part of the subject W can be photographed. In general, an irradiation field stop or casing for preventing excessive exposure is often installed between the subject W and the X-ray tube 2, and there is a limit to reducing the distance R1.

  Therefore, when there is a limit on the setting of the distance R3 such as in the photographing room, it is preferable to fix the distance R3 and change the ratio of the distances R1 and R2 within the fixed distance R3. For example, when R3 = 3.5 (m) is determined, R1 = 0.7 (m) and R2 = 2.8 (m) are set for this distance R3. Considering the size of a general shooting room, the distance R is set to a range of 3 ≦ R3 ≦ 5, and the relationship between the magnification rate M and the visibility of the enlarged image is observed within this range, empirically and experimentally. What is necessary is just to determine the optimal distances R3, R1, and R2.

  As can be seen from Equation 3, the degree of blur B greatly depends on the focal diameter D. When observing minute asbestos of 0.05 ≦ s ≦ 10 on an enlarged image, increasing the focal diameter D increases the amount of X-ray irradiation and improves the visibility of the enlarged image. As a result, the final observed image has a reduced edge enhancement effect and, in some cases, has no edge enhancement effect.

  Therefore, when a small asbestos of 0.05 ≦ s ≦ 10 (μm) is to be imaged, the focal diameter D of the X-ray tube 2 is made as small as 1 ≦ D ≦ 30 to be as close to the point source as possible. It is preferable that the ratio M is set as relatively large as 10 ≦ M ≦ 40 to enlarge the image portion of minute asbestos. When determining the distance R1, R2, and R3 according to the focal diameter D, the enlargement ratio M, and the enlargement ratio M that are actually applied within this range, the extent of the blur B, that is, the asbestos edge affected by the blur B In addition to the degree of visibility degradation of the image, the degree of attenuation of the edge enhancement effect and the degree of improvement in visibility by the image processing (gradation conversion processing, frequency enhancement processing) performed later on the generated X-ray image are taken into consideration And may be determined as appropriate.

Next, image generation in the photographing apparatus 10 will be described.
When magnified shooting is performed, the magnified image detected by the image detector 31 is read, and the analog image signal of the magnified image is sampled in the main body 33 (FPD when FPD is applied as the image detector 31). The image data is generated by being converted into a digital image signal by (reading).
At this time, it is preferable that the read pixel size P (μm), which is a generation unit of the enlarged image data, satisfies P ≦ s × M. Under this condition, there is an increased possibility that the read data (digital signal value) for at least one pixel (pixel) in the enlarged image area of the asbestos image area (area s × M) becomes a signal value level corresponding to asbestos, A minute asbestos signal can be detected.

  More preferably, reading is performed with a reading pixel size P that satisfies 2P ≦ s × M. That is, at least two pixels are read in the asbestos image area (area s × M) of the enlarged image. In the case of P ≦ s × M, depending on the imaging phase at the time of reading, a signal of one asbestos image region may be detected over a plurality of pixels, and the signal intensity of the asbestos region may be impaired. However, when 2P ≦ s × M, the read data for at least one pixel (pixel) within the asbestos diameter has a signal value level corresponding to asbestos regardless of the imaging phase at the time of reading. In the case of detecting a signal in the asbestos region with the signal intensity, it is preferable that 2P ≦ s × M is satisfied.

  When digital data of an enlarged image is generated in the main body 33 by the above-described enlarged photographing, the enlarged image data is output from the main body 33 to the image processing apparatus 20.

Next, the image processing apparatus 20 will be described.
The image processing apparatus 20 performs various types of image processing using the enlarged image data input from the photographing apparatus 10, and as illustrated in FIG. 7, a control unit 21, an operation unit 22, a display unit 23, and a communication unit 24. A storage unit 25, an image memory 26, and an image processing unit 27.

  The control unit 21 performs various calculations or performs centralized control of the operations of the units 22 to 26 in accordance with a control program stored in the storage unit 25.

  The operation unit 22 includes a mouse, a keyboard, and the like. When these are operated, an operation signal corresponding to the operation is generated and output to the control unit 21.

  The display unit 23 includes a display such as an LCD (Liquid Crystal Display), and displays various operation screens, enlarged images, processed images, and the like according to the control of the control unit 21.

  The communication unit 24 includes a communication interface and communicates with each device on the network N. For example, enlarged image data is received from the imaging device 10, or processed image data generated by the image processing unit 27 is transmitted to the image server 40.

  The storage unit 25 stores various control programs, an image processing program in the image processing unit 26, parameters necessary for executing the program, data, and the like.

  The image memory 26 is a memory for temporarily storing the enlarged image to be processed and the processed image data after the image processing.

As shown in FIG. 8, the image processing unit 27 performs various image processing such as gradation conversion processing and frequency enhancement processing on the enlarged image to generate a processed image.
Hereinafter, the contents and flow of each image processing will be described with reference to FIG.

<Irradiation field recognition processing>
The image processing unit 27 first executes an irradiation field recognition process on the input enlarged image as a premise of the gradation conversion process, the frequency enhancement process, and the like. The irradiation field refers to an area where X-rays have reached through the subject. In the irradiation field recognition processing, the irradiation field area is distinguished from an irradiation field outside area (an area other than the irradiation field). This is because appropriate processing cannot be performed if gradation conversion processing or the like is performed including an image in an irradiation field region having a biased signal value (digital signal value).

  Any method of irradiation field recognition may be adopted. For example, as disclosed in Japanese Patent Laid-Open No. 5-7579, an enlarged image is divided into a plurality of small areas, a dispersion value is obtained for each of the divided areas, and an edge of the irradiation field area is detected based on the obtained dispersion value. The irradiation field area may be determined. Usually, since the reaching X-ray dose is substantially uniform in the irradiation field region, the dispersion value of the small region becomes small. On the other hand, in a small region including the edge of the irradiation field region, a portion having a large arrival X-ray dose (outside the irradiation field region) and a portion (irradiation field region) in which the arrival X-ray dose is somewhat reduced by the subject are mixed, and thus the variance value is growing. Therefore, assuming that an edge is included in a small region having a variance value greater than a certain value, the region surrounded by such a small region is determined as an irradiation field region.

<Region of interest>
When the irradiation field region is determined, a region of interest (hereinafter referred to as ROI: Region Of Interest) is set from the irradiation field region. At this time, the reference signal value is set together with the ROI setting. Hereinafter, an example in which a lung field where asbestos exists is set as an ROI will be described.

  First, the horizontal direction and the vertical direction of the enlarged image data are sequentially scanned to create a signal value profile in each direction. Since the lung field shows a higher value than the peripheral organs such as the trachea and the thoracic vertebra, the inflection point is detected in the profile, and the region of the lung field is specified by the position of the inflection point. Note that the lung field may be detected by pattern matching, and any method may be applied.

  Then, a histogram of the specified lung field region is created, and values at predetermined ratios from the maximum value side and the minimum value side in the histogram are determined as the maximum reference value H and the minimum reference value L, respectively. The maximum reference value H and the minimum reference value L are used as reference values for converting the signal value range of the enlarged image immediately after generation into the signal value range (maximum value SH, minimum value SL) in the output image. .

<Tone conversion processing>
When the preprocessing is completed as described above, gradation conversion processing is performed.
The gradation conversion process is a process for adjusting density and contrast at the time of image output. When a doctor diagnoses the presence of a human body disease (for example, lung cancer in the chest) by interpretation of an X-ray image, the presence or absence of the disease is determined based on the density and contrast (gradation) of the structure on the X-ray image. The Therefore, by adjusting the density and contrast suitable for interpretation, a doctor's disease detection operation can be supported.

  The gradation conversion processing is performed in two stages, (1) normalization processing and (2) conversion processing using a basic LUT (lookup table), so that the desired signal value range and gradation characteristics are finally obtained. Tone conversion is performed.

  From the background that the screen / film method has been conventionally used for photographing, the doctor's interpretation ability (diagnosis) has been adopted even now that the digital processing method using the image detector 31 such as a volatile phosphor plate or FPD has been adopted. In order to maintain (performance), conversion processing of an input signal (read signal) is performed with a target of gradation characteristics (contrast) cultivated by the screen / film method.

  The gradation characteristic obtained by the screen / film system is an S-shaped curve as shown in FIG. In the gradation conversion process, an LUT indicating this gradation characteristic is prepared as a basic LUT, and after individual signal adjustment is performed on the target image by the normalization process, signal values are converted using the basic LUT. .

FIG. 10 shows the relationship between the X-ray dose detected by the image detector 31 (in the case of the phosphor plate) and the signal value of the X-ray image finally output according to the X-ray dose.
In the coordinate system of FIG. 10, the first quadrant indicates the reading characteristic, and indicates the relationship between the X-ray dose reaching the image detector 31 and the reading signal value (analog signal value). The second quadrant shows normalization characteristics, and shows the relationship between the read signal value and the normalized signal value (digital signal value) after the normalization process is performed. The third quadrant shows tone conversion characteristics and shows the relationship between the normalized signal value and the output density value (digital density signal value) converted by the basic LUT. Here, the output density value is a 12-bit resolution of 0 to 4095.

  In the second quadrant, the straight line indicating the normalization characteristic can adjust the output value range (size between SH and SL) by changing the slope thereof, and can also change the contrast of the entire image. This slope is defined as a G value. Further, by changing the intercept of the straight line indicating the gradation conversion characteristic, the height of the entire output value range (SH-SL movement) can be adjusted, and thereby the density of the entire image can be changed. Let this intercept be an S value.

For example, comparing the case where normalization is performed using the straight line h2 and the straight line h3 shown in FIG. 10, by increasing the G value, the normalized signal value corresponding to asbestos corresponds to the straight line region of the basic LUT. By improving the contrast, the visibility of the asbestos part is improved. At this time, the region where the transmitted X-ray dose is large is saturated, but there is no problem because there is no interpretation target in this portion.

  That is, the density range and contrast of the output image can be adjusted by controlling the gradient G value and intercept S value of the straight line indicating the gradation conversion characteristics as gradation conversion parameters.

The G value is determined by the following formula (4) for obtaining the gradient of the gradation characteristic curve in the screen / film system shown in FIG.
G = (D2-D1) / (logE2-logE1) (4)
here,
D1 = 0.25 + Fog, D2 = 2.0 + Fog, Fog = 0.2,
E1 and E2 are incident X-ray doses corresponding to D2 and D1, respectively.
In the case where each part of the human body such as the chest and breast is to be observed, the G value is generally about 2.5 to 5.0 in many cases.

Moreover, S value is calculated | required by following formula (5).
S = QR × P1 / P2 (5)
here,
QR is the quantization domain value,
P1 is an ultimate X-ray dose that results in a signal value 1535 (QR = 200, output density 1.2), and P2 is an actual ultimate X-ray dose of a pixel that has an output density of 1.2 in the image after gradation conversion.
The value of P1 is uniquely determined by the setting of the quantization region QR value before photographing.

The target image is normalized by performing signal conversion of the target image from a straight line determined by the obtained G value and S value. Then, gradation conversion is performed on the normalized image using the basic LUT to obtain a processed image having desired gradation characteristics.
That is, by performing normalization using the G value and the S value, gradation conversion can be performed in accordance with variations in X-ray dose and output characteristics of the output means even when the same LUT is used.

  In FIG. 10, a symbol h indicates a histogram of X-ray dose for the lung field where asbestos is present. The signal value of the asbestos part is biased toward the low signal value side in the lung field part, and the signal width is narrow.

<Frequency enhancement processing>
When gradation conversion is completed, frequency enhancement processing is performed on the processed image.
As a method of frequency enhancement processing, there are a method of performing non-sharp mask processing, a method of performing multi-resolution decomposition, and the like. Here, a method of performing multi-resolution decomposition described in JP-A-9-44645 and the like Will be described as an example.
In the method of performing multi-resolution decomposition, a processing target image is decomposed into a plurality of frequency band signals, and a signal in a desired frequency band is emphasized among the decomposed signals.

  First, every other pixel is sampled from the processing target image, and a pixel having a signal value “0” is interpolated between the sampled pixels. That is, “0” pixels are inserted every other row and every other row of the pixels arranged in a matrix in the sampled image.

  Next, the interpolated interpolated image is subjected to filter processing using a low-pass filter to obtain a low resolution image g1. This low resolution image g1 is obtained by extracting a low frequency band having a spatial frequency lower than half compared to the original image to be processed. This is because an image in a frequency band with a spatial frequency higher than half is lost by interpolating a pixel having a pixel value of “0” by sampling and a pixel value of “0”.

Then subtracted from the target image to low resolution image g _1, to obtain a high resolution image j _1. The high resolution image j ₁ are those spatial frequency image having high high frequency band than half are extracted, among the Nyquist frequency N of the original image is an image showing the frequency band of N / 2 to N.

Next, with respect to the low-resolution image g _1, filtering by the low-pass filter, it is subjected to interpolation processing to obtain a low-resolution image g _2, from the low-resolution image g ₁ by subtracting the low-resolution image g ₂ High-resolution images j ₂ is obtained. The high resolution image j ₂ Of the Nyquist frequency N of the original image, an image of the frequency band of N / 4 to N / 2 only.

By repeating the filtering process and the interpolation process in this way, a high resolution image j _k (k = 1, 2,...) Indicating the frequency band of N / 2 ^{k + 1 to} N / 2 ^k is ^obtained from the low resolution image gk. Obtainable.

When the high-resolution image j _k in the desired frequency band is obtained, the high-resolution image j _k in the frequency band to be emphasized is multiplied by the enhancement coefficient. Then, in order to restore the multi-resolution separated images, performing an inverse transformation of each image j _k containing high-resolution images the enhancement coefficient is multiplied.

First, subjected to interpolation processing between pixels with respect to the low resolution image g _k, and generates an inverse transformed image gg _k 4 times the size of the low resolution image g _k. Next, the inversely transformed image gg _k and the high resolution image j _k are added to obtain an added image gj _k−1 . An inversely transformed image gg _k-2 is generated by interpolation processing for this added image gj _k-1 , and a high resolution image j _k- 1 is added thereto to obtain an added image gj _k-2 .

Such processing is repeated, and an added image jk ₁ obtained by addition with the high resolution image j ₁ having the highest resolution is output as a processed image by frequency enhancement processing.

Here, the spatial frequency F (lp / mm) necessary for visually recognizing a structure having a width A (μm) or less must satisfy F ≧ 500 / A. In the magnified shooting, the subject width A on the image detector 31 is A = r × M + B obtained by multiplying the subject size r by the enlargement factor M and adding the blur width B. Asbestos having a diameter of 0.05 ≦ s ≦ 10 has a maximum size of 10 (μm), and is represented by A = 10M + B. Therefore, the spatial frequency F necessary for visual recognition of asbestos is F = 500 / ( 10M + B).
Therefore, to enhance the visibility of asbestos in the output image, when performing the frequency emphasis performs enhancement processing for the spatial frequency F = 500 / (10M + B ) or higher-resolution images j _k.

  As described above, the obtained processed image is transmitted from the image processing apparatus 20 to the image server 40, and is stored in the image DB 40a in the image server 40. The processed image stored in the image DB 40 a may be output to the display device 50 or output to the film output device 30 in response to a request from the display device 50.

In the X-ray image system, enlarged photographing was performed under the following experimental conditions, and visual evaluation was performed on an output image obtained by outputting the obtained enlarged image on a film.
<Experimental conditions>
The subject was a glass wool having a diameter of 5 (μm) attached to a 5 cm thick acrylic plate and used as a simulated phantom of asbestos in the chest.
The X-ray tube used was a prototype manufactured by Konica Minolta, and the one having a focal diameter of D = 10 (μm). The company also used a prototype manufactured by the company.
The image detector used was a Regius plate RP-5PM and a Regius cassette RC-110M, which are phosphor plates manufactured by the same company.
The enlarged image detected by the image detector was read by Regius model 190 manufactured by Konica Minolta. This reading apparatus can switch the reading pixel size to 43.75 or 87.5 (μm).
The film output device used DRYPRO model 793 made by the same company, and output to the film with a writing pixel size of 25 (μm). At this time, each pixel of the output target image and each pixel of the output image were made to correspond to 1: 1 and output without performing interpolation processing.

<Shooting conditions>
The tube voltage of the X-ray tube at the time of imaging is 65 (kVp) and 80 (kVp), and the tube current is 1 (mA).
The distance R3 from the focal point of the X-ray tube to the image detector is fixed at R3 = R1 + R2 = 3.5 (m), and the distances R1 and R2 are 2 ≧ R1 ≧ 0.07, 3.43 ≧ R2 respectively. Shooting was performed in a range where the magnification rate M was 1 ≦ M ≦ 50 by varying the range of ≧ 1.5.
Table 2 below shows the photographing conditions at this time and the measurement result of the blur B caused by the conditions. In order to identify each imaging condition, the imaging condition No. Is granted. Each of these photographing conditions is such that the degree of improvement in visibility due to enlargement, the degree of deterioration in visibility of asbestos edges due to increased blur, and the degree of edge enhancement effect are different.

<Image processing conditions>
A processed image subjected to gradation conversion processing and frequency enhancement processing and an unprocessed enlarged image were created for the enlarged image obtained by photographing, and these were output by the film output device. For the processed image, the G value is changed in the gradation conversion process, and the frequency band to be emphasized is changed in the frequency enhancement process.

<Evaluation criteria>
The evaluation criteria for the enlarged image output and formed on the film are as follows.
A: Each fiber can be clearly recognized.
○: It is possible to recognize a group of several fibers.
(Triangle | delta): It turns out that there is a lump of fiber.
X: Fiber cannot be visually recognized.
In the above evaluation, the number of fibers that can be recognized is the same as that of the comparative example, but a symbol “+” is given when the appearance of the fibers is clearer.
In accordance with the above evaluation criteria, seven image evaluators observed the image on the film and evaluated the image of the glass wool fiber that was the subject.

<Evaluation results of Examples 1 to 7>
No. above. When photographing is performed under the photographing conditions 1 to 9 and only the gradation conversion process is performed on the enlarged image generated with the read pixel size 43.75 (μm) (Examples 1 to 3), only the frequency enhancement process is performed. Table 3 below shows the image evaluation results in the case of applying (Examples 4 to 7). In Examples 1 to 3, gradation conversion processing is performed with G values of 10, 20, and 50. Examples 4 to 7 are 0.1 or more, 1 or more, 1.28 or more, 2 or more, respectively. The frequency band is subjected to enhancement processing. On the other hand, as a comparative example, an image having only a tone conversion process with a G value of 4.89 was prepared.

  From the evaluation results shown in Table 3, when the conditions 4 to 8 satisfy P ≦ 5M and the enlargement ratio M satisfies 10 ≦ M ≦ 40 in common with Comparative Example 1 and Examples 1 to 7, the fiber of the subject is determined. It is a good result that can be visually recognized. Moreover, when the comparative example 1 and Examples 1-3 are compared, the visibility of a fiber will improve by making G value 20 or more, and also a fiber can be identified finely by making G value 20 or more. It can be seen that the visibility is improved. This is considered to be because the contrast was increased by adjusting the G value.

  Although there is no difference in image evaluation between Comparative Example 1 and Examples 4 and 5, in Examples 6 and 7 in which enhancement processing is performed on a frequency band of 500 / (10M + B) or more, conditions 5 The best result was obtained at 7 and the image quality was such that the fibers could be clearly observed.

  In addition, when Examples 2 and 3 where the best evaluation is performed are compared with Examples 6 and 7, it can be seen that an image in which the contrast is adjusted by controlling the G value can obtain a better overall evaluation. .

<Examples 8 to 12>
No. above. When photographing is performed under the photographing conditions 1 to 9 and only the gradation conversion process is performed on the enlarged image generated with the read pixel size 87.5 (μm) (Examples 8 and 9), only the frequency enhancement process Table 4 below shows the image evaluation results in the case of applying (Examples 10 to 12). In Examples 8 and 9, gradation conversion processing is performed with G values of 10 and 20, respectively. In Examples 10 to 12, enhancement processing is performed in frequency bands of 0.1 or more, 1.28 or more, and 2 or more, respectively. Is given. On the other hand, as a comparative example, an image having only a tone conversion process with a G value of 4.89 was prepared.

From the evaluation results shown in Table 4, when the conditions 6 to 8 satisfy 2 × P ≦ 5M and the enlargement ratio M satisfies 10 ≦ M ≦ 40 in common with Comparative Example 2 and Examples 8 to 12, It is a good result that the fiber is visible.
Further, when comparing Comparative Example 2 with Examples 8 and 9, Example 9 that performed gradation conversion processing with a G value of 20 or more gave better results. Further, in Examples 11 and 12 in which enhancement processing was performed on a frequency band of 500 / (10M + B) or higher, the evaluation was generally the same as in Comparative Example 2, but the fiber bundles were clear enough to be identified. It turns out that it is.

<Examples 13 to 18>
No. above. When photographing is performed under the photographing conditions of 1 to 9, and gradation conversion processing and frequency enhancement processing are performed in combination on an enlarged image generated with a read pixel size of 43.75 (μm) (Examples 13 to 18) The image evaluation results are shown in Table 5 below. Examples 13 to 15 have a G value of 10, and Examples 16 to 18 have been subjected to gradation conversion processing with a G value of 20. The frequency bands to be emphasized are 0.1 or more, 1.28 or more, 2 or more, respectively. It is made variable. On the other hand, as a comparative example, an image having only a tone conversion process with a G value of 4.89 was prepared.

  From the evaluation results shown in Table 5, when the conditions 4 to 8 satisfy P ≦ 5M and the enlargement ratio M satisfies 10 ≦ M ≦ 40 in common with Comparative Example 3 and Examples 13 to 18, the fiber of the subject is determined. It turns out that it has become the favorable result which can be visually recognized. Further, from the comparison between Comparative Example 3 and Examples 13 to 15, the fiber was reduced even when the enlargement ratio was slightly small by setting the G value to 10 or more, or conversely, the enlargement ratio was large and the ratio of blur was large. It can be seen that the image quality is easy to see. In Examples 14 and 15 in which enhancement processing was performed for a frequency band of 500 / (10M + B) or higher in addition to gradation conversion processing, clear image quality was obtained for the fiber structure.

  In Examples 16-18, it turns out that the image quality which can identify one fiber is obtained by making contrast large as G value 20. In addition, as in Examples 14 and 15, the sharpness is improved in Examples 17 and 18 in which enhancement processing is performed on a frequency band of 500 / (10M + B) or more.

<Examples 19 to 24>
No. above. When photographing is performed under the photographing conditions of 1 to 9, and gradation conversion processing and frequency enhancement processing are performed in combination on an enlarged image generated with a read pixel size of 87.5 (μm) (Examples 19 to 24) The image evaluation results are shown in Table 6 below. Examples 19 to 21 have a G value of 10, and Examples 22 to 24 have been subjected to gradation conversion processing with a G value of 20. The frequency bands to be emphasized are 0.1 or more, 1.28 or more, 2 or more, respectively. It is made variable. On the other hand, as a comparative example, an image having only a tone conversion process with a G value of 4.89 was prepared.

From the evaluation results shown in Table 6, when the conditions 6 to 8 satisfy 2 × P ≦ 5M and the magnification ratio M satisfies 10 ≦ M ≦ 40 in common with Comparative Example 4 and Examples 19 to 24, It is a good result that the fiber is visible.
In addition, although there is no difference in evaluation between Comparative Example 4 and Example 19, in addition to the gradation conversion process, a frequency enhancement process is performed for a frequency band of 500 / (10M + B) or more from Examples 20 and 21. Thus, it can be seen that a clear image quality can be obtained from the fiber structure.

  In Examples 22 to 24 in which gradation conversion processing was performed with a G value of 20, evaluations better than those of Examples 19 to 21 with a G value of 10 were obtained. It is considered that the visibility of the fibers is improved by increasing the contrast. Further, it can be confirmed from Examples 23 and 24 that sharpness is enhanced by further enhancing the frequency band of 500 / (10M + B) or more.

Note that the same results as the evaluations shown in Tables 3 to 6 were also obtained when the same experiment was performed by changing the diameter of the glass wool fiber of the subject stepwise from 0.05 to 10 (μm) for each constant value. was gotten.
In the above evaluation, only the experiment in the case of the focal diameter D = 30 (μm) is shown. However, since the blur B becomes smaller as the X-ray source is closer to the point source, the focal diameter D should be as small as possible.

  As described above, according to the present embodiment, magnified imaging is performed in which the focal diameter D of the X-ray tube is 1 ≦ D ≦ 30 and the magnification rate M is 10 ≦ M ≦ 40. By applying frequency emphasis processing to a frequency band greater than / (10M + B), it is possible to obtain a high-quality enlarged image in which very small asbestos with 0.05 ≦ s ≦ 10 can be visually recognized.

  Furthermore, by selecting the generation unit P of the enlarged image data so as to satisfy P ≦ s × M, it is possible to improve the detectability of the signal of the asbestos portion, and improve the visibility of asbestos in the output image. Further, by setting 2P ≦ s × M, it is possible to improve the signal detection accuracy of the asbestos part, and it is possible to obtain an output image with higher visibility.

  In consideration of the size of the imaging room, the distance R3 from the focus of the X-ray tube to the image detector is fixed within a range of 3 ≦ R3 ≦ 5, and the distance R1 is within the fixed distance R3. , R2 are varied to adjust the enlargement factor M. Therefore, the enlargement factor M can be adjusted to a relatively large enlargement factor M such as 10 ≦ M ≦ 40 in a limited room.

  When signal processing is performed as in this case, human structures such as lung fields are uniformly crushed in black (corresponding to the maximum density Dmax) and bones are crushed in white (thinking Dmini). In the case of being swallowed, the pattern is a pattern in which asbestos parts having a relatively low concentration are scattered in the black background part. This is a pattern similar to a case in which a minute calcified cluster of about several hundred μm is present in a high-density breast portion as a background in a breast image. In the breast image, as described in Japanese Patent Laid-Open No. 10-91758, etc., a density gradient is calculated for such a pattern, and a microcalcification cluster portion is automatically detected from the calculation result. . Such a technique can also be applied to asbestos, and the asbestos part can be automatically detected as an abnormal shadow.

It is a figure which shows the structure of the X-ray-image system in this embodiment. It is a figure which shows the imaging device of FIG. It is a figure explaining expansion photography. It is a figure explaining the edge effect by expansion photography. It is a figure which shows the relationship between the edge intensity | strength in an edge effect, and a blur. It is a figure explaining the case where a blur arises in expansion photography. It is a figure which shows the internal structure of the image processing apparatus of FIG. It is a figure explaining the flow of the image processing in an image processing apparatus. It is a figure which shows the gradation conversion characteristic made into the target. It is a figure which shows the flow of the conversion of the signal value at the time of a gradation conversion process.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 X-ray imaging system 10 Imaging apparatus 20 Image processing apparatus 30 Film output apparatus 40 Image server 50 Display apparatus

Claims (2)

An X-ray tube that irradiates the subject with X-rays;
An image detector that generates an X-ray image corresponding to an X-ray dose of X-rays transmitted through the subject;
An image processing apparatus that performs image processing on the generated X-ray image;
Have
The focal diameter of the X-ray tube is D (μm),
The distance from the focal point of the X-ray tube to the subject is R1 (m),
The distance from the subject to the image detector is R2 (m),
The distance from the focal point of the X-ray tube to the image detector is R3 (m),
Magnification factor M = R3 / R1,
Blur width B = D · R2 / R1 (μm),
And when
1 ≦ D ≦ 30,
3 ≦ R3 ≦ 5,
10 ≦ M ≦ 40,
The filling,
An asbestos phase contrast X-ray imaging system, wherein the image processing device performs frequency enhancement processing on the generated X-ray image in a frequency band of 500 / (10M + B) (lp / mm) or more. Irradiating a subject with X-rays from an X-ray tube;
Generating an X-ray image corresponding to an X-ray amount of X-rays transmitted through the subject by an image detector;
Applying image processing to the generated X-ray image by an image processing apparatus;
Including
The focal diameter of the X-ray tube is D (μm),
The distance from the focal point of the X-ray tube to the subject is R1 (m),
The distance from the subject to the image detector is R2 (m),
The distance from the focal point of the X-ray tube to the image detector is R3 (m),
Magnification factor M = R3 / R1,
Blur width B = D · R2 / R1 (μm),
And when
1 ≦ D ≦ 30,
3 ≦ R3 ≦ 5,
10 ≦ M ≦ 40,
The filling,
In the step of performing the image processing, the generated X-ray image is subjected to frequency enhancement processing in a frequency band of 500 / (10M + B) (lp / mm) or more, and phase contrast X-ray imaging for asbestos Method.

Images