JP2582665B2 - Abnormal shadow detector - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an abnormal shadow detection device that detects an abnormal shadow appearing as a substantially circular pattern in a radiation image based on image data representing the radiation image of the subject. .

(Prior art) Reading a recorded radiation image to obtain image data,
After performing appropriate image processing on the image data, reproduction and recording of the image are performed in various fields. For example, an X-ray image is recorded using an X-ray film having a low gamma value designed to be compatible with later image processing, and this X-ray image is recorded.
An X-ray image is read from a film on which a line image is recorded, converted into an electric signal (image data), subjected to image processing, and then reproduced as a visible image in a copy photograph or the like, thereby providing contrast and sharpness. It has been practiced to obtain a reproduced image having good image quality performance such as graininess and the like (see Japanese Patent Publication No. 61-5193).

In addition, the applicant (X-ray, α-ray, β-ray,
Irradiation with gamma rays, electron beams, ultraviolet rays, etc. accumulates a part of this radiation energy, and then irradiation with excitation light, such as visible light, causes a stimulable phosphor (luminous) to emit stimulated emission according to the accumulated energy. Radiation image information of a subject such as a human body is temporarily recorded in a sheet-shaped stimulable phosphor using a stimulable phosphor, and the stimulable phosphor sheet is scanned with excitation light such as laser light to radiate the light. Generates luminescent light, obtains image data by photoelectrically reading the obtained stimulating luminescent light, and outputs a radiation image of the subject as a visible image to a recording material such as a photographic photosensitive material or a CRT based on this image data. A radiation image recording / reproducing system for causing the radiation image to be reproduced has already been proposed (Japanese Patent Laid-Open No. 55-1242).
No. 9, 56-11395, 55-163472, 56-104645
No. 55-116340, etc.).

This system has the practical advantage of being able to record images over a very large radiation exposure area compared to conventional radiographic systems using silver halide photography. That is, in the case of the stimulable phosphor, it has been recognized that the amount of emitted light that is stimulated by excitation after accumulation is proportional to the radiation exposure amount over an extremely wide range. Even if fluctuates considerably, the stimulated emission light emitted from the stimulable phosphor sheet is converted into an electric signal by reading the amount of light by setting the reading gain to an appropriate value and reading it by photoelectric conversion means, and By outputting a radiation image as a visible image on a recording material such as a photographic light-sensitive material or a display device such as a CRT using, a radiation image which is not affected by a change in radiation exposure can be obtained.

In recent years, in a system using the X-ray film, the stimulable phosphor sheet, or the like, particularly a system configured for medical diagnosis of a human body, a reproduced image having good image quality suitable for observation (diagnosis) is obtained. In addition, automatic recognition of images has been performed (for example, see Japanese Patent Application Laid-Open No. Sho 62-1254).
No. 81).

Here, automatic image recognition refers to an operation of extracting a target pattern from a complex radiation image by performing various processes on image data. For example, various linear shapes such as a chest X-ray image of a human body are used. , For example, an operation of extracting a shadow corresponding to a tumor from a very complicated image mixed with a circular pattern.

Such a complicated radiographic image (for example, the chest X of the human body)
A target pattern (for example, tumor shadow) is extracted from the line image), and a visible image in which the extracted pattern is specified is reproduced and displayed, thereby assisting an observer in observation (for example, assisting a doctor in diagnosis). be able to.

(Problems to be Solved by the Invention) When automatically extracting, for example, an abnormal shadow such as a tumor shadow which appears as a substantially circular pattern on a chest X-ray image of a human body, the X-ray image is extremely complicated. Only a tumor shadow is not extracted. For example, a shadow of a blood vessel bifurcation where a blood vessel is bifurcated or a so-called blood vessel tangent shadow in which a blood vessel extends in a direction in which X-rays travel at the time of imaging are almost circular patterns. Therefore, the shadow of the blood vessel branch point, the shadow of the tangent of the blood vessel, and the like are often extracted as tumor shadows.

For this reason, it is conceivable that the blood vessel shadow is also extracted separately from the tumor shadow, and if the blood vessel shadow is extracted at the same time as the blood vessel shadow, it is determined that the blood vessel shadow is not a tumor shadow. However, this is the case that the extraction of the blood vessel shadow is performed without error, the extraction of the blood vessel shadow is usually not complete, and the pattern of the width and the length of the blood vessel shadow is almost the same.
That is, a substantially circular pattern such as a tumor shadow is often extracted as a blood vessel shadow. Therefore, as it is, a linear pattern (blood vessel shadow) is mixed in a circular pattern (tumor shadow), and a circular pattern (tumor shadow) is also extracted as a linear pattern (blood vessel shadow).
Is mixed, so how a circular pattern (tumor shadow)
It is important to distinguish only the linear pattern (blood vessel shadow) from the extracted pattern.

The present invention has been made in view of the above circumstances, and provides an abnormal shadow detection device capable of detecting an abnormal shadow appearing as a substantially circular pattern on a radiographic image with high accuracy in distinction from a linear pattern appearing on a radiographic image. It is the purpose.

(Means for Solving the Problems) The abnormal shadow detection device of the present invention is based on image data representing a radiation image of a subject.
In the abnormal shadow detection device that detects an abnormal shadow appearing as a substantially circular pattern on the radiation image, scanning the radiation image using a first film configured to extract the substantially circular pattern. The first extraction means for extracting the candidate for the abnormal shadow appearing in the radiographic image, configured to extract a linear pattern that appears on the radiographic image and includes a linear pattern having the same width and length. By scanning the radiographic image using the second filter, the second extraction means for extracting a linear pattern appearing in the radiographic image, and the candidate and the linear pattern on the radiographic image In the region where both are extracted, the area of the linear pattern in the enlarged candidate region including the candidate and extending in the vicinity of the candidate and the It is characterized in that the candidates based on the maximum width of the linear pattern is provided with a judging means for judging whether or not the abnormal shadow.

Here, the "linear pattern" typically includes a shadow of a blood vessel appearing in a radiation image of a human body, but may also be a shadow of a rib, for example.

(Action) Normally, filters for detecting abnormal shadows such as tumor shadows and filters for detecting linear patterns such as blood vessel shadows are devised so as to detect abnormal shadows and linear patterns as completely as possible. However, in the present invention, it is rather allowed that the filters are incomplete, and it is separately determined whether the patterns detected by both filters are abnormal shadows or linear patterns. That is, the abnormal shadow detection device of the present invention aims to mainly extract an abnormal shadow in the first filter, but allows, for example, a shadow of a blood vessel bifurcation to be extracted as a candidate for an abnormal shadow. In the second filter, a linear pattern having a width and a length similar to each other, that is, an abnormal shadow and a pattern similar thereto are also extracted as a linear pattern. In this manner, both filters are extracted. In the region, since it is determined whether or not the candidate is an abnormal shadow based on the cleaning pattern area in the expansion candidate region and the maximum width of the linear pattern, the abnormal shadow and the shadow of the blood vessel branch point are determined. An abnormal shadow such as a shadow of a blood vessel or a tangent of a blood vessel is distinguished from a confusing linear pattern, and the abnormal shadow can be detected with high accuracy.

(Example) Hereinafter, an example of the present invention will be described with reference to the drawings. Here, an example will be described in which the above-described stimulable phosphor sheet is used to detect, as an abnormal shadow, a tumor shadow that typically occurs in the lungs of a human body as a substantially spherical shape. This tumor appears on the visible image reproduced based on the image data as an almost circular pattern that is whitish (has a lower density) than the surroundings.

 FIG. 10 is a schematic diagram of an example of an X-ray imaging apparatus.

The X-ray source 11 of the X-ray imaging apparatus 10 irradiates X-rays 12 toward the chest 13a of the human body 13 and irradiates the stimulable phosphor sheet 14 with the X-rays 12a transmitted through the human body 13 so that the human body 13 is exposed. The X-ray image of the chest 13a is stored and recorded on the stimulable phosphor sheet 14.

FIG. 11 is a perspective view showing an example of an X-ray image reading device and a computer system which is an embodiment of the abnormal shadow detection device of the present invention.

After imaging is performed by the X-ray imaging apparatus 10 shown in FIG.
The stimulable phosphor sheet 14 is set at a predetermined position of the X-ray image reading device.

The stimulable phosphor sheet 14, which is set at the predetermined position and on which the X-ray image is stored, is conveyed (sub-scanning) in the direction of arrow Y by sheet conveying means 22 such as an endless belt driven by a motor 21. You. Meanwhile, laser light source 23
Is reflected and deflected by a rotating polygon mirror 26 driven by a motor 25 and rotating at high speed in the direction of the arrow, and passes through a focusing lens 27 such as an fθ lens.
The optical path is changed by 28 to be incident on the stimulable phosphor sheet 14, and the main scanning is performed in an arrow X direction substantially perpendicular to the sub-scanning direction (arrow Y direction). From the portion of the stimulable phosphor sheet 14 irradiated with the light beam 24, a stimulable luminescent light 29 of a light amount corresponding to the accumulated and recorded X-ray image information is diverged, and this stimulable luminescent light is emitted.
29 is guided by a light guide 30 and is photoelectrically detected by a photomultiplier (photomultiplier tube) 31. The light guide 30 is formed by molding a light-guiding material such as an acrylic plate, and is arranged so that the linear incident end face 30a extends along the main scanning line, and is formed in an annular shape. The light receiving surface of the photomultiplier 31 is connected to the emission end face 30b. The stimulated emission light 29 that has entered the light guide 30 from the incident end face 30a travels through the inside of the light guide 30 by repeating total reflection, exits from the emission end face 30b, is received by the photomultiplier 31, and is received by the X-ray image. Is converted into an electric signal by the photomultiplier 31.

The analog output signal S _A output from the photomultiplier 31 is logarithmically amplified by a logarithmic amplifier 32, and the A / D converter 3
Digitization is performed in step 3 to obtain digital image data _SD .

The image data _SD obtained in this manner is input to the computer system 40. This computer system
Reference numeral 40 denotes an example of the abnormal shadow detection device of the present invention. The main unit 41 includes a CPU and an internal memory, a drive unit 42 into which a floppy disk as an auxiliary memory is inserted and driven, and an operator. Keyboard 43, X for inputting necessary instructions and the like to computer system 40
It comprises a line image and a CRT display 44 for displaying necessary information.

Image data input to this computer system 40
Abnormal shadows on the X-ray image are detected based on the _SD .

In the present embodiment, each function obtained by combining the hardware constituting the computer system 40 and the software executed in the computer system 40 is regarded as an example of each means according to the present invention.

First extraction means (abnormal shadow extraction means) In the abnormal shadow extraction means in the computer system 40,
A tumor shadow appearing in the X-ray image is extracted by scanning the X-ray image using a tumor shadow extraction filter based on the image data _SD .

FIG. 2, for explaining an example of a real space filter for extracting tumor shadow, said around a predetermined pixel P ₀ on the X-ray image X
It is the figure drawn virtually on the line image.

Given pixel P ₀ is whether the pixel of the tumor Kagenai is determined. By scanning on an X-ray image using a filter as shown here, a tumor shadow appearing in the X-ray image is extracted. The filter described first below is a filter described in Japanese Patent Application No. 1-162904.

Figure 3 is centered on the given pixel P _0, which is a diagram showing an example of a profile of the X-ray image of the second view of the line segment L ₁ and L ₅ of extending direction (x-direction).

As shown in FIG. 2, a plurality of (eight in this case) line segments extending from a predetermined pixel P _{0 in the} X-ray image to the periphery of the X-ray image
Assuming L _i (i = 1, 2,..., 8), circles R _j (j = 1, 2, and ₃ ) having radii r ₁ , r ₂ , and r ₃ centered on a predetermined pixel P ₀ , respectively.
Assume 3). X-ray image data of a predetermined pixel P ₀ is defined as f _0, and each pixel P _ij located at each intersection of each line segment _Li and each circle R _j
(In FIG. 2, P ₁₁ , P ₁₂ , P ₁₃ , P ₅₁ , P ₅₂ , and P ₅₃ are marked with symbols.) X-ray image data is _defined as f _ij .

Here, the X-ray image data f ₀ of a given pixel P ₀ pixels P
difference delta _ij between the X-ray image data f _ij of _ij is calculated according to the following equation (1).

_{Δ ij = f ij -f 0 ...} (1) (i = 1,2, ......, 8; j = 1,2,3) then for each line segment L _i, the difference obtained in (1) The maximum value of _Δij is obtained. That is, when an example line segments L _1, L _5, for the line segment L _1, pixel P _11, P _12, P each difference corresponding to _{13 Δ 11 = f 11 -f 0} Δ 12 = f 12 - The maximum value of f ₀ Δ ₁₃ = f ₁₃ −f ₀ is obtained. In this example, Δ ₁₃ > Δ ₁₂ > ₁₁ as shown in FIG. 3, and therefore Δ ₁₃ is the maximum value.

The maximum value of the pixel P _51, P _52, P ₅₃ each difference corresponding to _{Δ 51 = f 51 -f 0 Δ} 52 = f 52 -f 0 Δ 53 = f 53 -f 0 for line L ₅ delta ₅₃ is obtained, which is the difference delta _51, delta
₅₂ is a representative value representing the delta _53.

The line segments L _i for each predetermined pixel P ₀ as a plurality of pixels P
the maximum value of the difference delta _ij with _ij, the maximum value this that the obtained representative values for the line segment.

Next, the person sets the two line segments extending in opposite directions from a given pixel P _0, i.e. the line segment L ₁ and the line segment L _5, a line segment L ₂ and the line segment L _6, a line segment L ₃ With the line segment L ₇ and each of the line segment L ₄ and the line segment L ₈ as one set, the average value of two representative values (M ₁₅ , M ₂₆ , M ₃₇ , and M ₄₈ , respectively) is obtained for each set. . The set of the segment L ₁ and the line segment L _5, the average value M ₁₅ is Is required.

By handling this way the two line segments extending in opposite directions from a given pixel P ₀ as human pair, the distribution of the image data tumor shadow 7 is in the position where a concentration gradient is not asymmetrical Can reliably detect the tumor shadow.

When the average values M ₁₅ , M ₂₆ , M ₃₇ , and M ₄₈ are obtained as described above, based on these average values M ₁₅ , M ₂₆ , M ₃₇ , and M ₄₈ , a predetermined pixel is calculated as follows. It is determined whether P ₀ is a pixel in the tumor shadow. These average values M ₁₅ , M ₂₆ , M ₃₇ ,
The determination based on the M ₄₈ is not limited to a specific determination method, the following method is employed for example.

4 is a diagram for explaining an example of a method of obtaining characteristic values C ₁ used in the determination. The horizontal axis is the average value M ₁₅ , M ₂₆ , M ₃₇ , M ₄₈ determined as described above, and the vertical axis is the average value M ₁₅ , M
Evaluation values C ₁₅ , C ₂₆ , C ₃₇ , and C ₄₈ corresponding to ₂₆ , M ₃₇ , and M ₄₈ , respectively.

If the average value M ₁₅ , M ₂₆ , M ₃₇ , M ₄₈ is smaller than a certain M _{1, the} evaluation value is zero, and if it is larger than a certain value M _{2, the} evaluation value is 1.0, M _{1 to} M ₂
In the middle of, a value between 0.0 and 1.0 is an evaluation value according to the magnitude of the value. In this way, each average value M ₁₅ , M ₂₆ , M
Evaluation values C ₁₅ , C ₂₆ , C ₃₇ , and C ₄₈ corresponding to ₃₇ and M ₄₈ are obtained, and the sum of these evaluation values C ₁₅ , C ₂₆ , C ₃₇ , and C ₄₈ is C ₁ = C ₁₅ + C ₂₆ + C ₃₇ + C ₄₈ ... (3) is a characteristic value C _1. That is, this characteristic value C ₁ is the minimum value 0.0
And any value between 4.0 and the maximum value.

This characteristic value C ₁ is compared with a predetermined threshold value Th ₁ , and C ₁ ≧ T
whether it is h1, by either a C ₁ <Th1, a given pixel P ₀ is whether the pixel of each tumor Kagenai is determined.

Upon obtaining the evaluation values _{C 15, C 26, C 37} , C 48, fourth each evaluation value by using a conversion equation as saturated at as indicated by the chain line smaller value M ₂ 'in Figure C ₁₅ , C ₂₆ , C ₃₇ , and C ₄₈ , the characteristic value C ₁ obtained according to the equation (3) becomes a characteristic value C ₁ having a large value in the case of a tumor shadow closer to a circle, and vice versa. in the Figure 4 to determine the characteristic values C ₁ with saturated no such conversion formulas to a large value M ₂ "as shown by the two-dot chain line, the characteristic value C ₁ is a large tumor shadow contrast with the surroundings a characteristic value C ₁ having a large value for. Thus appropriate conversion formula is selected depending on the purpose.

Note that the determination based on the average values M ₁₅ , M ₂₆ , M ₃₇ , and M ₄₈ is not limited to the method using the characteristic value C ₁ , and for example, thresholds M ₁₅ , M ₂₆ , M ₃₇ , and M ₄₈ are used. M ₁₅ , M ₂₆ , M
_37, M ₄₈ Kyo該the predetermined pixel P ₀ if the threshold Th2 or more may be determined that the pixel of tumor Kagenai, and M
_15, may be determined as M _26, M _37, M the predetermined pixel P ₀ when three or more is the threshold Th2 or more of the ₄₈ is a pixel of tumor Kagenai, further M ₁₅ , M ₂₆ , M ₃₇ , M ₄₈ sum M = M ₁₅ +
M ₂₆ + M ₃₇ + M ₄₈ may be obtained, and the sum M may be compared with a threshold Th 3 to determine that the predetermined pixel P ₀ is a pixel in the tumor shadow when M ≧ Th 3.

The algorithm of the filter for extracting the tumor shadow is not limited to the above algorithm. Hereinafter, another example will be described (see Japanese Patent Application No. 1-162905).

The gradient ▽ f _ij of the pixel data f _ij of each pixel P _ij (i = 1, 2,..., 8; j = 1, 2, 3) in FIG. 2 is obtained.

FIG. 5 is a diagram showing the gradient and the calculation method described below.

After the gradient ▽ f _ij is determined, the length of the vector of these gradients ▽ f _ij is adjusted to 1.0. That is, _{assuming that} the magnitude of the gradient ▽ f _ij is | ▽ f _ij |, the normalized gradient ▽ f _ij / | ▽ f _ij | is obtained.

Then, the normalized gradient _{▽ f ij / | ▽ f ij} | , the direction component of the line segment L _i is calculated. That is, a unit vector from each pixel P _ij toward a predetermined pixel P ₀ is And when (Where * represents the inner product).

Thereafter, the components a inward (direction of the predetermined pixels P ₀₎ positive for, when the outward and negative, each line segment L _i (i = 1, 2,
…, 8) each maximum value Is required.

Furthermore, each of these maximum values Value obtained by adding Is required. By dividing the added value by the number of line segment L _i (8 in this embodiment) becomes an average value. Therefore, this added value is obtained by simply multiplying the average value by a constant, and can be identified with the average value.

This sum As the feature amount C ₂ and threshold the characteristic quantity C ₂ of a predetermined Th4
And whether C ₂ ≧ Th4 or C ₂ <Th4 determines whether each of the predetermined pixels P ₀ is a pixel in the tumor shadow.

This filter has the magnitude of the gradient ▽ f _ij | ▽ f _ij
| Normalized, that by focusing only on the (degree of direction difference between the line segment L _i), the feature amount C ₂ which shape regardless of the contrast with the surrounding has a large value by a circular
Is obtained, whereby the tumor shadow is extracted with great accuracy.

Further, in the above embodiment, as shown in FIG.
Each image data f corresponding to each pixel P _ij on the line segments L _{1 to} L _{8
Although ij} is used, it is needless to say that the number of line segments does not need to be eight, and may be, for example, sixteen. In addition, the calculation was performed for three distances r ₁ , r ₂ , and r ₃ with respect to the distance from the predetermined pixel P ₀ , but the calculation is not limited to the three distances and the size of the tumor shadow to be extracted When the distance is almost constant, the distance may be one (in this case, the calculation for obtaining the representative value is unnecessary). To extract the tumor shadows of various sizes with higher precision,
It may perform operations for almost continuous number of distances from r ₁ to a distance r _3.

Next, a filter having a different algorithm will be described (see Japanese Patent Application No. 1-162909).

Figure 6, in order to illustrate this algorithm, a diagram depicting virtually on the image around a predetermined pixel P ₀ on the X-ray image.

As shown in FIG. 6, a plurality of (eight in this case) line segments extending from a predetermined pixel P _{0 in the} X-ray image to the periphery of the X-ray image
Assume L _i (i = 1, 2,..., 8), and further assume a circle R having a radius r and centered on a predetermined pixel P ₀ . Also, each pixel P _i (i = 1, 2,...) At the intersection of the circle R with the center area Q ₀ including the predetermined pixel P ₀ and each of the line segments L _i (i = 1, 2,..., 8). consider each peripheral area Q _i, including 8). Incidentally, the radius r, the center region Q _0, Suto of each peripheral area of the region Q _i, and assuming that the peripheral region, the size of the tumor shadow of interest, determination accuracy, to properly consider the operation speed, etc. Determined. Also, if in the present embodiment, it is assumed each pixel P _i away from the predetermined pixels P ₀ equal distance r, which for example, the extraction target tumor shadow having major axis in the X direction of FIG. 6, etc. for selecting a pixel from the pixel P _1, pixel P ₀ as P ₅ in the far may have different distances from a given pixel P ₀ for each pixel P _i.

Average values Q ₀ , Q _i (i = 1, 2,..., 8) of a large number of image data corresponding to a large number of pixels in the central region Q ₀ and the peripheral regions Q _i assumed as described above are obtained. Can be For the sake of simplicity, the same symbol is used for a symbol indicating each area Q ₀ , Q _i (i = 1, 2,..., 8) and a symbol indicating the average value of image data in each area. ing.

Here, the average value of these differences Δ _i (i = 1, 2,..., 8) And the distribution Is required.

Next, these average values The ratio C ₃ of the variance σ ² is Obtained as this ratio C ₃ is compared with a predetermined threshold value Th1, in the case of C ₃ ≧ Th5 relatively average value Is large and the variance σ ² is small, the predetermined pixel P ₀ is determined to be a pixel in the tumor shadow, and if C ₃ <Th5, the predetermined pixel P
₀ is determined to be a pixel outside the tumor shadow.

A further different real space filter will be described with reference to FIG.

As shown in FIG. 2, a plurality of (eight in this case) line segments extending from a predetermined pixel P _{0 in the} X-ray image to the periphery of the X-ray image
Assuming L _i (i = 1, 2,..., 8), three circles R _j (j = 1, ₂ ) with respective radii r ₁ , r ₂ , r ₃ centered on a predetermined pixel P ₀
2,3) is assumed. The central area including the predetermined pixel P ₀ is defined as Q _0, and each pixel P _ij located at each intersection of each line segment _Li and each circle R _j
(The symbols for P ₁₁ , P ₁₂ , P ₁₃ , P ₅₁ , P ₅₂ , and P ₅₃ are shown in FIG. 2.) Q _ij (i = 1, 2,..., 8; j
= 1,2,3) (However, in FIG. 2, Q ₀ and
Only Q ₁₁ , Q ₁₂ , Q ₁₃ , Q ₅₁ , Q ₅₂ and Q ₅₃ are shown. ).

For each of the regions Q ₀ and Q _ij (i = 1, 2,..., 8; j = 1, 2, 3), a large number of pixels corresponding to a large number of pixels in each of the regions Q ₀ , Q _ij Average values Q ₀ , Q _ij (i = 1, 2,..., 8; j = 1,
2,3) is required. For simplicity, each area Q
_The same symbol is used for the signal indicating ₀ , Q _ij (i = 1, 2,..., 8; j = 1, 2, 3) and the signal indicating the average value of the image data in each area.

Next, each difference Δ _ij (i = 1, 2,..., 8; j = 1, 2, 3) between the average value Q _{0 of the} central region and the average value Q _ij of each peripheral region is Δ _ij = Q _ij -Q ₀ ... determined as (5), further each line segment L _i, the maximum value delta _i of the difference delta _ij is calculated.

Next, a second characteristic value V representing the variation of the maximum value Δ _{i (i} = 1~8) first characteristic value representative of the U and the maximum value Δ _{i (i} = 1~8) is obtained. For this purpose, first, the characteristic values U _{1 to} U ₄ and V _{1 to} V ₄ are obtained according to the following arithmetic expressions.

_{U 1 = (Δ 1 + Δ} 2 + Δ 5 + Δ 6) / 4 ... (6) U 2 = (Δ 2 + Δ 3 + Δ 6 + Δ 7) / 4 ... (7) U 3 = (Δ 3 + Δ 4 + Δ 7 + Δ 8 ) / 4 ... (8) U 4 = (Δ 4 + Δ 5 + Δ 8 + Δ 1) / 4 ... (9) V 1 = U 1 / U 3 ... (10) V 2 = U 2 / U 4 ... (11) V ₃ = U ₃ / U ₁ (12) V ₄ = U ₄ / U ₂ (13) Here, for example, a case where the characteristic value U ₁ is obtained according to the equation (6) will be described. delta ₁ and delta ₂ or delta adding about ₅ and delta ₆₎ means smoothing, region of opposite sides sandwiching the pixel P ₀ (delta _1,
The addition of + Δ ₂ and Δ ₅ + Δ ₆ ) is performed so that even if there is a tumor shadow in an area having a concentration gradient, the tumor shadow can be detected.

Further, for example, (10) will be described for obtaining the characteristic value V ₁ in accordance with formula, is a characteristic value determined for directions perpendicular to each other and the characteristic value U ₁ and the characteristic value U _3, thus the tumor shadow as shown in FIG. 3 If 57 is circular, V ₁ ≒ 1.0, and if it deviates from the circle, that is, if pixel P ₀ is within a linear shadow, V ₁ will deviate from 1.0.

As a first characteristic value U representing the maximum value Δ _{i (i} = 1~8) of the difference, the maximum value U = MAX of _{U 1 ~U 4 (U 1,} U 2, U 3, U 4) ... (14) is employed, as the second characteristic value V representing the variation of the maximum value of the difference delta _{i (i} = 1 to 8), the maximum value V = MAX of V ₁ ~V ₄ (V _1, V ₂ , V ₃ , V ₄ ) (15) are adopted. Thus, the first and second characteristic values
U, the V is obtained, as a characteristic value C ₄ for a given pixel P ₀ to determine whether the pixel of tumor Kagenai, the ratio of these first and second characteristic values There is employed, the characteristic value C ₄ is compared with a predetermined threshold value Th6, or a C ₄ ≧ Th6, by either a C ₄ <Th6, the pixel
It is determined whether each of P ₀ is a pixel in the tumor shadow.

Incidentally, in the above-described film example, using the mean value Q _ij of the image data corresponding to each peripheral region Q _ij including a pixel P _ij of the second eight as shown in FIG line L ₁ ~L ₈ However, the number of the line segments does not need to be eight, and may be, for example, sixteen. The same applies to the embodiment described with reference to FIG. Further, in the above-described embodiment which was carried out with reference to FIG. ₂ , the calculation was performed for the three distances r ₁ , r ₂ , and r _3. However, the calculation is not limited to the three distances, but may be of various sizes. in order to extract the tumor shadow more accurately the distance a may be performed continuously varied computed from r ₁ to r _3.

Scanning the X-ray image using any of the above real space filters or a combination thereof or other known filters detects tumor shadows that typically appear as circular patterns on the X-ray image However, since the determination means described later determines whether or not the tumor shadow detected in this way is a new tumor shadow, the tumor shadow detected in this manner is used as a candidate tumor shadow. Shall be called.

Second Extraction Means (Vessel Shadow Extraction Means) Immediately before and after the detection of the tumor shadow candidate, the vascular shadow extraction means in the computer system 40 executes the following vascular shadow extraction filter based on the image data _SD. By scanning on the X-ray image using the image, the blood vessel shadow appearing in the X-ray image is extracted.

An example of a blood vessel shadow extraction filter will be described with reference to FIG. 6 virtually drawn on an X-ray image.

Given pixel P ₀ is whether the pixel of the vessel Kagenai is recognized. By scanning on an X-ray image using a filter as shown here, a blood vessel shadow appearing in the X-ray image is extracted.

Figure 7 is centered on the given pixel P _0, shows an example of a profile of the X-ray image of the 6 view line L ₁ and L ₅ of extending direction (x-direction), FIG. 8 FIG. 7 is a diagram schematically depicting a blood vessel shadow and a filter shown in FIG. Here, the predetermined pixel P ₀ is
It is assumed that the shadow is substantially at the center of the blood vessel shadow 1.

As shown in FIG. 6, a plurality of (eight in this case) line segments extending from a predetermined pixel P _{0 in the} X-ray image to the periphery of the X-ray image
_{L i (i = 1,2, ...} , 8) assumed, and further assuming a circle R of radius r centered at the given pixel P _0, also the central region Q ₀ which includes the given pixel P _0, Each peripheral area Q _i including each pixel P _i (i = 1, 2,..., 8) on the intersection of each of the line segments L _i (i = 1, 2,. Think. Incidentally, the radius r, the area of the central region Q ₀ and the peripheral area Q _i, and the number i such peripheral areas Q _i vascular shadow thickness to be subjected to recognition, noise mixed in the X-ray image It is appropriately determined in consideration of the component size, recognition accuracy, calculation speed, and the like.

Here, the average values Q ₀ , Q _i (i = 1, 2,..., 8) of a large number of image data corresponding to a large number of pixels in the central area Q ₀ and each peripheral area Q _i assumed as described above. Is required. Here, for the sake of simplicity, the same symbol is used for the symbol indicating each of Q ₀ , Q _i (i = 1, 2,..., 8) and the symbol indicating the average value of the image data in each area. I have.

When the average values Q ₀ , Q _i of the image data of the central area Q ₀ and the peripheral areas Q _i are obtained in this manner, each difference Δ _i is calculated based on the following equation: Δ _i = Q _i −Q ₀ (17) sought Te, then these maximum values .DELTA.max = MAX of the differential _{Δ i (Δ i) ... (} 18) is obtained, the output value the maximum value .DELTA.max corresponds to a predetermined pixel P _0. This implements a blood vessel shadow recognition filter. The maximum value Δmax is set as the image data value of the predetermined pixel P ₀ , and the same operation is performed for each pixel of the X-ray image to obtain an X-ray image in which the blood vessel shadow is emphasized.し き い 値 max
The blood vessel shadow is extracted by extracting only the pixel that becomes Th7.

However, in the above embodiment even when the profile 3 shown in broken lines even when the profile 2 the profile of the image shown in Figure 7 is indicated by the solid line, the same output values when the .DELTA.max = delta _5. That is, the predetermined pixel P ₀ is the blood vessel shadow 1
And the case near the boundary line 3a, the result is similarly recognized. Even if the boundary line 3a is extracted as a blood vessel shadow, there is no problem because it is excluded from the determination of the tumor shadow by the determination means described later, but the boundary line 3a is extracted as a blood vessel shadow as another example of the blood vessel shadow extraction filter below. 5 shows a blood vessel shadow extraction filter that does not have any problem.

For as in the above embodiments the central region Q ₀ and the peripheral region Q _i the average value Q _0, Q _i of the image data, further the above (1
7) After obtaining each difference delta _i based on the formula, the line segment L _{i (i} = 1,2 as shown in Figure 6, ..., 2 that extend in opposite directions from a given pixel P ₀ of 8) Two peripheral regions on the line segment of
That is, Q ₁ and Q ₅ , Q ₂ and Q ₆ , Q ₃ and Q ₇ , Q ₄ and Q ₈ are each a pair, and the difference Δ _{i, i + 4} (i, 1,2,3,4) for each pair but the following equation _{Δ i, i + 4 = Δ} i + Δ i + 4 - | Δ i -Δ i + 4 | ... determined according to (19), the maximum value .DELTA.max = MAX of the difference _{Δ i, i + 4 (Δ} i _{, i + 4} ) (20) is required. Here, considering the case of i = 1, (19)
The equation is as follows: Δ _1,5 = Δ ₁ + Δ ₅ | Δ ₁ −Δ ₅ | (21) Here, in the case of the profile 2 indicated by the solid line in FIG. 7, since Q ₁ ≒ Q ₅ , Δ ₁ ≒ Δ ₅ and thus | Δ ₁ −Δ ₅ | ≒ 0. _1,5 ≒ 2Q ₁ −2Q ₀ … (22) On the other hand, if the broken line of profile 3 shown in Figure 7, because it differs significantly from the Q _1, Q _5, for example, Q ₅ ≒ Q ₀
Then, Δ ₅ ≒ 0, and thus Δ _1,5 ≒ 0 (23). That is, by performing the calculations of the equations (19) and (20), the boundary line is prevented from being recognized as a blood vessel shadow. However, a substantially circular pattern such as a tumor shadow having substantially the same diameter as the width of the blood vessel shadow shown in FIG. 8 has the same output value. Patterns are recognized similarly. An example of this point will be described below. Referring to FIGS. 7 and 8, when the predetermined pixel P ₀ is substantially at the center of the blood vessel shadow 1, Q ₀ = Q ₂ = Q ₃ = Q ₇ = Q ₈ (24) Q ₁ = Q ₄ = Q ₅ = Q ₆ (25) As shown in the above equation (22), Δ _1,5 ≒ 2Q ₁ −2Q ₀ (22), and Δ ₂₆ = Δ ₃₇ = Δ ₄₈ = 0 (26), and Δmax = Δ ₁₅ = 2Q ₁ −2Q ₀ (27).

On the other hand, if the given pixel P ₀ is in the approximate center of the tumor shadow 4 shown in FIG. _{8, Q 1 = Q 2 = Q} 3 = Q 4 = Q 5 = Q 6 = Q 7 = Q 8 ... (28) Δ ₁₅ = Δ ₂₆ = Δ ₃₇ = Δ ₄₈ = 2Q ₁ -2Q ₀ (29), and the maximum value Δmax is Δmax = 2Q ₁ -2Q ₀ (30) as in the above equation (27). That is, this filter does not distinguish between the blood vessel shadow 1 and the tumor shadow 4 shown in FIG. The same applies to the above-described embodiment. That is, the blood vessel shadow extraction filter also extracts a tumor shadow having a diameter approximately equal to the thickness of the blood vessel shadow as a blood vessel shadow.

In the above blood vessel shadow extraction filter, was used as the average value Q ₀ of the image data of the central area Q ₀ and the mean value Q _i of the image data of each peripheral area Q _i, each peripheral region and these central regions Q ₀ Q _i
Is set according to the magnitude of the noise superimposed on the X-ray image, and therefore, when the noise is small or when noise removal processing is separately performed, the average value
Instead of using Q ₀ and Q _i , the above-described predetermined pixel P ₀ and the image data itself corresponding to each pixel P _i may be used.

In the above two examples of the blood vessel shadow extraction filter, the predetermined distance r (see FIG. 6) is considered to be fixed. The blood vessel shadow may be thin or thick, and when only the blood vessel shadow of a specific width is to be recognized, the predetermined distance r may be fixed. Is performed for various predetermined distances r to obtain a plurality of values, and by adopting the maximum value among the plurality of values, a vascular shadow having various widths from a thin vascular shadow to a thick vascular shadow is obtained. May be recognized.

Determination means In the determination means in the computer system 40, in the region where both the candidate and the blood vessel shadow are extracted on the X-ray image where the candidate for the tumor shadow and the blood vessel shadow are extracted as described above, Thus, it is determined whether the extracted pattern is a tumor shadow or a blood vessel shadow.

FIG. 9 is a diagram showing a blood vessel shadow and a diagram virtually drawn on the blood vessel shadow to explain a method of obtaining the thickness (width) of the blood vessel shadow, which is necessary for performing this determination. is there. As it follows width of vascular Pictures 1 at the position of a given pixel P ₀ is calculated.

First, a large number of two line segments extending in opposite directions (in the example shown in FIG. 9, L ₁ and L ₅ , L ₂ and L ₆ , L ₃ and L ₇ , L ₄ and L ₈ ) It sets a number of line segments extending toward the end of a predetermined pixel P ₀ in X-ray images so as to form a (this four sets in the example)
Each intersection C ₁ , C of L ₁ , L ₂ ,..., L ₈ and the boundary line 1 a, 1 b of the blood vessel shadow 1
₂ ,…, C ₈ are required.

When the intersections C ₁ , C ₂ ,…, C ₈ are obtained in this way,
Then each intersection and a predetermined pixel _{P C 1, C 2, ...} , the distance d ₁ between C _8,
d ₂ ,..., d ₈ are obtained, and then these distances d ₁ , d ₂ ,
…, D ₈ are added. That is, d ₁ + d ₅ , d ₂ + d ₆ , d ₃ + d ₇ ,
d ₄ + d ₈ is required. When the distances d ₁ + d ₅ , d ₂ + d ₆ , d ₃ + d ₇ , d ₄ + d ₈ between the intersections of each group are obtained in this way, the minimum value of the distances between the intersections, that is, in this example, d ₁ + d ₅
The minimum value d ₁ + d ₅ is determined as the width (thickness) of the blood vessel shadow at the position of the predetermined pixel P ₀ .

FIGS. 1A, 1B, and 1C schematically show a pattern of an X-ray image in an enlarged candidate region that includes a region extracted as both a tumor shadow candidate and a blood vessel shadow and spreads near the region. 1A, 1B, and 1C show a tumor shadow, a shadow of a branch point of a blood vessel, and a shadow of a tangent of a blood vessel, respectively.

When the tumor shadow extraction filter is used, the shadow of the branch point of the blood vessel shown in FIG. 1B and the shadow of the tangent of the blood vessel shown in FIG. 1C may be extracted as a tumor shadow because they form a substantially circular pattern.

On the other hand, when a blood vessel shadow extraction filter is used, a tumor shadow as shown in FIG. 1A, which has a diameter substantially equal to the thickness of the blood vessel shadow as described above, is also extracted as a blood vessel shadow.

Therefore, the enlarged candidate region 4 is considered to include the regions extracted as both the tumor shadow candidate 7 and the blood vessel shadow 1, and the area S of the blood vessel shadow 1 in the enlarged candidate region is set. As described above, the enlarged candidate area 4 refers to an area including the tumor shadow candidate 7 and the surrounding area as described above. For example, the binarization processing is performed by comparing the output values of the various filters with a certain threshold value. When a tumor shadow candidate is extracted by performing the process, the threshold may be lowered, and the area may be extended to the periphery of the tumor shadow candidate, and the expanded area may be used as an enlarged candidate area. , A circular area extending to the range of the tumor shadow candidate may be set, and the set circular area or the like may be set as the enlarged candidate area 4. Further, the thickness (width) of the blood vessel shadow 1 is obtained in accordance with the above-described algorithm for obtaining the thickness of the blood vessel shadow at each pixel position while moving a predetermined pixel P ₀ in the enlarged candidate region 4 and the extracted blood vessel shadow 1. And the maximum width d is obtained. Note that the tumor shadow as shown in FIG. 1A is also extracted as a blood vessel shadow. Therefore, the diameter of the tumor shadow is also calculated as the blood vessel shadow thickness (width) according to the above-described algorithm for calculating the blood vessel shadow width (thickness). Can be

Thus, the area S of the blood vessel shadow 1 in the enlargement candidate area 4
Is determined, whether or not the candidate for the tumor shadow is truly a tumor shadow is determined based on the area S and the maximum width d. Here, the arithmetic expression for determination is not limited to a specific one, but for example, , And this value is compared with a predetermined threshold value Th8. If E ₁ ≦ Th8 (32), it is determined to be a tumor shadow, and if E ₁ > Th8 (33), the shadow of the vascular tangent or It is determined that the shadow is a shadow of a branch point of a blood vessel and is not a tumor shadow.

As an arithmetic expression for this determination, for example, E ₂ = S / d (34) may be used.

However, it is not necessary to apply this determination process to a tumor shadow candidate or the like having a large area that is apparently not a blood vessel shadow.

In the computer system 40 shown in FIG. 11, whether or not the once extracted tumor shadow is truly a tumor shadow is determined as described above.

After the determination as described above, for example, when a visible image is reproduced and displayed on the CRT display 44, it is possible to assist the observer by specifying the region determined as a true tumor shadow.

The above embodiment is an example of extracting a tumor shadow that typically appears as a circle in a human chest X-ray image obtained using a stimulable phosphor, but the present invention is limited to a chest X-ray image. It is not limited to a system using a stimulable phosphor, and the abnormal shadow is not limited to a tumor shadow but may be, for example, a calcified shadow, and a linear pattern is not limited to a blood vessel shadow. Instead, it may be a rib shadow or the like, and has a configuration that can be widely used when detecting an abnormal shadow on a radiation image based on image data representing the radiation image of the subject.

(Effects of the Invention) As described in detail above, the abnormal shadow detection device of the present invention extracts a response between a candidate for an abnormal shadow and a linear pattern, and, in a region where both of the above are extracted on a radiographic image, Considering the expansion candidate area, it is determined whether or not the pattern in the expansion candidate area is a true abnormal shadow based on the area and the maximum width of the extracted linear pattern in the expansion candidate area. Even if the abnormal shadow extraction filter and the linear pattern extraction filter are imperfect such that they enter into each other and are extracted, the abnormal shadow can be distinguished from the linear pattern. Can be detected with high accuracy.

[Brief description of the drawings]

FIGS. 1A to 1C are diagrams schematically showing an X-ray image pattern of an enlarged candidate region including a region extracted as both a tumor shadow candidate and a blood vessel shadow and extending in the vicinity of the region. figure for explaining an example of a real space filter for extracting tumor shadow, virtually drawn figure on the image around a predetermined pixel P ₀ on the X-ray image, Fig. 3, the above predetermined the pixel P ₀ centering a diagram showing an example of a profile of the X-ray image of the second view of the line segment L ₁ and L ₅ of extending direction (x-direction), FIG. 4, the predetermined pixels P ₀ FIG. 5 is a diagram for explaining an example of a method of obtaining a characteristic value used for determining whether or not a pixel is a pixel in a tumor shadow. FIG. 5 is a diagram showing a vector such as a gradient ▽ f _ij of image data f _ij . FIG. 6 is a diagram virtually drawn on an X-ray image for explaining an example of a blood vessel shadow extraction filter and an example of a tumor shadow extraction filter; 7 illustration, centered on the given pixel P _0, of the Figure 6 line L ₁
And illustrates an example of a profile of the X-ray image on the direction (x-direction) of extension of the L _5, FIG. 8 is a vascular shadows and drawing and was schematically shown superimposed filter shown in FIG. 6, the FIG. 9 is a diagram showing a blood vessel shadow and a diagram drawn virtually on the blood vessel shadow to explain a method of obtaining the thickness (width) of the blood vessel shadow. FIG. FIG. 11 is a schematic diagram showing an example of an apparatus, and FIG. 11 is a diagram showing an example of an X-ray image reading apparatus and a computer system including an example of the abnormal shadow detection apparatus of the present invention. 1 ... Vessel shadow 2,3 ... X-ray image profile 3a ... Boundary line, 4 ... Tumor shadow 10 ... X-ray imaging device, 11 ... X-ray source 14 ... Storable phosphor sheet 20 ... ... X-ray image reader, 23 ... Laser light source 29 ... Stimulated luminescence 31 ... Photomultiplier 40 ... Computer system 44 ... CRT display

Claims (1)

(57) [Claims] 1. An abnormal shadow detecting apparatus for detecting an abnormal shadow appearing as a substantially circular pattern on a radiation image based on image data representing a radiation image of the subject, wherein the substantially circular pattern is extracted. By scanning on the radiographic image using the first film that has been performed, first extraction means for extracting candidates for the abnormal shadow appearing in the radiographic image, appearing on the radiographic image, the width and length are By scanning on the radiation image using a second filter configured to extract a linear pattern including a similar linear pattern, a second pattern for extracting a linear pattern appearing in the radiation image Extracting means, and in a region where both the candidate and the linear pattern are extracted on the radiographic image, the region includes the candidate and spreads in the vicinity of the candidate. Determining that the candidate is the abnormal shadow based on the area of the linear pattern in the enlargement candidate area and the maximum width of the linear pattern in the enlargement candidate area; Characteristic abnormal shadow detection device.