JPH11155849A - Method for obtaining bone image information - Google Patents

Abstract

PROBLEM TO BE SOLVED: To effectively separate the random noises of dependent contrast from a bone section component even if an irradiated radiation dose is changed by using a threshold value dependent on an irradiated radiation dose when a radiation image to be subjected to bone processing is photographed for the threshold value processing of the skeleton processing. SOLUTION: A threshold value setting section 21 stores a function where a irradiated radiation dose at the time of photographing inputted image information and a threshold value suited for effectively eliminating random noises generated in the image at the time of receiving the irradiation are caused to correspond to each other beforehand as shown in a graph. Then, when the irradiated radiation dose at the time of photographing a radiation image is inputted from the outside, a threshold value corresponding to the stored radiated radiation dose and suited for effectively eliminating random noises is calculated. Then, the threshold value is inputted to a threshold processing section 20 to obtain a skeleton image. The obtained image is inputted to a second calculation section 30, and outputted from a morphological processing calculator 100 as bone image information which results from skeleton processing.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for acquiring bone image information, and more particularly to an improved method for obtaining bone (trabecular) image information of a human body or the like useful for diagnosis of osteoporosis or the like by skeleton processing based on a morphological operation. It is about.

[0002]

2. Description of the Related Art Bone mineral quantification, that is, quantitative measurement of the amount of calcium in bone is useful for diagnosis of fracture prevention.

[0003] In other words, the amount of bone mineral is determined by the density of the trabecular bone, which is the spongy material constituting the inside of the bone, that is, the bone density. Therefore, if the bone density is low, the density of the shadow of the bone in the bone image is low. When the bone density is high, the density of the bone shadow in the bone image is low.

[0004] Therefore, knowing a minute change in calcium in bone enables early detection of osteoporosis and has an important effect also in preventing fracture.

Therefore, conventionally, the MD method (Microdensitometry)
, SPA method (Single Photon Absorptiometry), DPA
Method (Dual Photon Absorptiometry), QDR method (Quantitat
ive Digited Radiography), QCT method (Quantitative Co
mputer Tomography), DQCT method (Dual energy Quantit
(Native Computer Tomography) has been proposed and implemented.

[0006] Each of these methods is for measuring the so-called bone density, and has been widely used for diagnosis of osteoporosis. However, even if only the numerical information of bone density is provided, the bone structure is not measured. It is difficult to grasp the state of the bones (trabecular bones) such as these alone.

Therefore, the present applicant has proposed a method of obtaining morphological information of a bone image with high accuracy by performing a skeleton process based on a morphological operation on a radiation image (Japanese Patent Application No. 8-253258). etc).

According to this method, it is possible to obtain a bone image with a high easiness of observing the morphology of the bone portion without the influence of the soft tissue and the like. It is possible to improve the diagnostic performance of osteoporosis by accurately grasping the osteoporosis.

Here, the skeleton processing based on the morphological operation will be described.

A morphological operation (hereinafter also referred to as morphological processing) is an algorithm of morphology (Morphology; also referred to as morphology or morphology) for selectively extracting only a specific image portion such as an abnormal shadow from an original image. It is a process based on, and it has been studied as an effective method to detect the microcalcification image which is a characteristic form especially in breast cancer, but the target image is not limited to the microcalcification image in such a mammogram .

In this morphological processing, processing using a structural element B corresponding to the size and shape of an image portion to be extracted is performed, and the extracted image is hardly affected by complicated background information. It does not distort.

That is, this method can detect geometrical information such as the size, shape, and density distribution of a calcified image better than a general differential process.

The outline of the morphological processing will be described below.
An example in which the present invention is applied to detection of a microcalcification image in a mammogram will be described.

(Basic calculation of morphology) Morphological processing is generally developed as a set theory in an N-dimensional space. However, for intuitive understanding, a description will be given of a two-dimensional grayscale image.

The point of coordinates (x, y) is represented by a density value f
It is regarded as a space having a height corresponding to (x, y). Here, the density value f (x, y) is a high-luminance high-signal level signal having a larger image signal value as the density is lower (the brightness is higher when displayed on a CRT).

First, for simplicity, consider a one-dimensional function f (x) corresponding to a cross section of the two-dimensional gray image. The structural element g used in the morphological operation is a symmetric function symmetric about the origin, as shown in the following equation (1).

[0017]

(Equation 1)

It is assumed that the value is 0 in the domain and the domain G is represented by the following equation (2).

[0019]

(Equation 2)

At this time, the basic form of the morphological operation is a very simple operation as shown in equations (3) to (6).

[0021]

(Equation 3)

That is, the dilation processing is performed in a range of ± m (a value determined according to the structural element B and corresponding to the mask size in FIG. 7) around the target pixel. Is a process of searching for a maximum value within the range (see FIG. 2A). On the other hand, an erosion process is a process of searching for a minimum value within a range of ± m around the pixel of interest. (See FIG. 2B). The opening process is a process of performing a dilation process after the erosion process, that is, a process of searching for a maximum value after searching for a minimum value.
ing) The process corresponds to a process of performing an erosion process after a dilation process, that is, a process of searching for a minimum value after searching for a maximum value.

In other words, in the opening process, the density curve f (x) is smoothed from the low brightness side, and a convex density variation portion (a portion having a higher brightness than the surrounding portion) which fluctuates in a spatially narrower range than the mask size 2 m. (See FIG. 3C).

On the other hand, in the closing process, the density curve f (x) is smoothed from the high brightness side, and a concave density variation portion (a portion having a lower brightness than the surrounding portion) which fluctuates in a spatially narrower range than the mask size 2 m. (See FIG. 3D).

When the structural element g is not symmetrical with respect to the origin, the dilation operation shown in the equation (5) is called Minkowski sum, and the erosion operation shown in the equation (6) is called Minkowski difference.

Here, in the case of a signal with a high density and a high signal level, the higher the density, the higher the density, the higher the density value f
Since the magnitude relationship is reversed with respect to the case where the image signal value of (x) is at the high luminance and high signal level, the dilation processing for the high density and high signal level signal and the erosion processing for the high luminance and high signal level (FIG. )), The erosion processing for the high-density high-signal level signal and the dilation processing for the high-brightness high-signal level (FIG. 10A) correspond to the opening processing for the high-density high-signal level signal. The closing process for a high-brightness high-signal level and the opening process for a high-brightness / high-signal level (FIG. 10C) match with the closing process for the high-brightness / high-signal level. .

In this section, a case of an image signal (brightness value) having a high brightness and high signal level will be described.

(Application to Calculated Shading Detection) For the detection of calcified shadows, a difference method of subtracting a smoothed image from an original image is considered. It is difficult to distinguish calcified shadows from elongated non-calcified shadows (such as mammary glands, blood vessels, and mammary gland supporting tissues) using a simple smoothing method. A morphological filter expressed by the following equation (7) based on the opening process has been proposed ("Extraction of microcalcification image by morphological filter using multiple structural elements" IEICE Transactions D-II Vol. .J75-D-II No.7 P1170-1176 July 1992
Tsuki, "Basics of Morphology and Its Application to Mammogram Processing" MEDICAL IMAGING TECHNOLOGY Vol.12 No.1 Jan
uary 1994).

[0029]

(Equation 4)

Here, Bi (i = 1, 2,..., N) has a linear size of m pixels and n pixels (for example, in the case of FIG. 8, 9 pixels and 4 directions, m = 9, n = 4) (these elements are collectively referred to as multiplexed structural elements in the m-pixel and n-directions). If the structuring element Bi is set to be larger than the calcified shadow to be detected, the processing by the above-described opening calculation causes a signal change portion finer than the structuring element Bi (an image portion in which the signal fluctuates in a narrow spatial range). The calcified shadow, which is a convex portion having a luminance value larger than that of the surroundings, is removed. On the other hand, a non-calcified shadow such as a shadow of an elongated mammary gland is longer than the structural element Bi, and if its inclination matches one of the four structural elements Bi, the opening processing (the equation (7)) Even if the two-term operation is performed, it remains as it is. Therefore, a smoothed image obtained by the opening process (an image from which only calcified shadows have been removed)
Is subtracted from the original image f to obtain an image including only small calcified shadows. This is the concept of equation (7).

As described above, in the case of a signal having a high density and a high signal level, the density value of the calcified shadow is lower than that of the surrounding image portion, and the density value of the calcified shadow is lower than that of the surrounding portion. Since the signal change portion has a small concave shape, a closing process is applied instead of the opening process, and Expression (8) is applied instead of Expression (7).

[0032]

(Equation 5)

The closing processing of the equation (8), which is an example of the morphological operation, will be specifically described.

That is, according to the morphological operation on the density value Sorg, which is an image signal of a high density and high signal level, for example, the density value Sor shown by the solid line in FIG.
The maximum value processing (dilation processing) is performed on the image data having the distribution of g with the linear three-pixel structural element B as shown in FIG. S _i is converted to S _i ′ employing the maximum value S _{i + 1} among three pixels (determined by the structural element B) adjacent to each other with the target pixel as the center. By performing this calculation for all the pixels, it is converted into a maximum value signal indicated by a broken line in FIG.

Next, considering the minimum value processing (erosion processing) by the structural element B for the maximum value signal obtained by this maximum value processing, the pixel of interest indicated by the broken line in FIG. The maximum value signal S _i ′ is converted into S _i ″ (= S _i ) using the minimum value S _i−1 ′ among three pixels adjacent to each other with the target pixel as a center. , The distribution of the minimum value signal Sorg ″ after the maximum value processing is indicated by the dashed line in FIG. In the image signal indicated by the dashed line, the image portion where the signal fluctuates in a spatially narrower range than the structural element B with respect to the original solid line original image data disappears,
This indicates that an image portion which is a change portion of the signal value which fluctuates in a wider range spatially than the structural element B and an image portion having no change remain in the original shape. That is, the above process (closing process) acts as a process for smoothing the image density distribution from the high density side.

The value Smor obtained by subtracting the value obtained by the closing process (the value obtained by performing the maximum value process on Sorg and then performing the minimum value process) from the original image signal Sorg is: It represents an image portion which is a change portion of a signal value which fluctuates in a narrow spatial range and is erased by the above-described closing process.

Here, the image signal is originally a position (x, y) which is a two-dimensional element and a signal value f which is a third-dimensional element.
(X, y), but in the above description, a one-dimensional image signal distribution curve which appeared on a predetermined cross section of an image developed on two dimensions has been described for ease of understanding. .

Therefore, in practice, the above description needs to be applied to a two-dimensional image, and the use of multiple structural elements is for the purpose of supporting a two-dimensional image.

Next, the skeleton processing based on the morphological operation will be described.

The skeleton processing is generally a skeleton of a figure (sk
eleton), and the skeleton can be regarded as a set of centers of disks inscribed in the figure. That is, for example, the skeleton of each figure (represented by a solid bold line) shown in FIGS.

Hereinafter, a case where the skeleton processing is performed by the above-described morphological operation will be described.
In this case, the skeleton processing is performed by the following equation (9) or (10).
Can be represented by

[0042]

(Equation 6)

[0043]

(Equation 7)

Here, the equations (9) and (10) are as described above.
This is due to the difference between representing the image as a high-density high-signal-level image signal or a high-luminance high-signal-level image signal. Equation (9) is applied to extract the skeleton of the image portion of high brightness (high brightness), while extracting the skeleton of the image portion of low brightness (high density) from the image represented by the image signal of high brightness and high signal level. In this case, equation (10) is applied, and there is no substantial difference in the operation itself.

For example, on a negative film (high density and high signal level), the bone has a lower density than other image portions, and the trabecular portion has a lower density.
The portion that does not exist has a higher concentration. Therefore, since the skeleton processing is performed on the trabecular bone which is a portion having a lower density than the surroundings, Equation (9) may be applied.

Here, it is assumed that the structural element B in the equation (9) is a circle having a radius r, and the figure shown in FIG. 11 is subjected to skeleton processing. In the graphic shown in FIG. 11, the area outside the contour is a high density part, and the inside is the low density part.

First, an erosion process by the structural element B is performed on this figure in the upper part of the figure. At λ = 0 (zero erosion processing by the structural element B), the figure does not change at all.

In λ = 1 (one erosion process using the structural element B), the figure is embedded inside the structural element B by the radius r.

In the case of λ = 2 (two erosion processes using the structural element B), the portion of the figure projecting from the circle completely disappears.

By repeating the same operation, λ = N
-1 (N-1 times of erosion processing by structural element B)
And the figure is only a circle with a radius of r or less.

On the other hand, in the middle part of the figure, the image obtained by performing the erosion process (λ = 0, 1, 2,..., N−1, N) by the structural element B is further opened by the structural element B. This is a processed graphic.

The lower figure is obtained by subtracting the middle figure from the upper figure in correspondence with the number of times of processing.

At λ = 1, it can be seen that the skeleton element of the portion protruding from the circle of the original figure is extracted, and at λ = N−1, the skeleton element of the circle of the original figure is extracted.

As described above, the original figure is subjected to the erosion processing, the opening processing is further performed, and the number of processings is subtracted according to the erosion processing. Each of the obtained difference images is binarized by a predetermined threshold value. Eliminate random noise to some extent.

The result of calculating the union (sum set) of the result is the meaning of the equation (9). In equation (10), the level of density is effective for extracting a skeleton element from a figure inverted from equation (9). Dilation processing is performed on the original figure, and further closing processing is performed. Is subtracted correspondingly, and a skeleton element is extracted by obtaining a union set obtained by binarizing with a threshold value.

The contents of the morphological operation and the skeleton processing have been described above.

[0057]

By the way, in the skeleton processing based on the morphological operation described above,
False detection of random noise as trabecular bone is inevitable to some extent. For this reason, it is desired to reduce erroneously detected random noise. In the above-mentioned method proposed by the present applicant, random noise is obtained by binarizing the obtained difference image with a preset threshold value. Is being reduced.

However, according to the bone image information acquisition method described above, since the threshold value is set in advance as a unique value, the contrast changes according to the image capturing conditions, specifically, the irradiation radiation dose. In some cases, random noise components cannot be effectively removed.

The present invention has been made in view of the above circumstances, and it is possible to effectively remove random noise even when the irradiation radiation dose is varied for each image capturing condition such as the type of a subject. A first object is to provide an image information acquisition method.

As one method of acquiring radiation image information, for example, a radiation image information recording / reproducing system called CR (Computed Radiography) is well known. This radiation image information recording / reproducing system accumulates a part of this radiation energy when irradiated with radiation, and then emits stimulated emission according to the accumulated radiation energy when irradiated with excitation light such as visible light or laser light. Using a stimulable phosphor (stimulable phosphor), radiation image information of an object such as a human body is temporarily stored and recorded on a sheet-shaped stimulable phosphor sheet in which a stimulable phosphor is laminated on a support. Then, the excitation light such as a laser beam is scanned for each pixel to generate stimulated emission light sequentially from each pixel, and the obtained stimulated emission light is photoelectrically sequentially read to obtain an image signal. The stimulable phosphor sheet after reading the image signal is irradiated with erasing light to release the radiation energy remaining on the sheet, thus enabling the sheet to be used repeatedly. Over a very wide radiation exposure range as compared with the radiographic system using, it has a practical advantage of being able to record the image.

For the image information obtained by such a system, the region of interest to be diagnosed in the image only needs to have the optimum contrast for observation. Therefore, the region of interest is set to the optimum contrast. As described above, the contrast conversion processing is performed (see
4-51229).

For this reason, in the bone image information acquisition method, the image to be subjected to the skeleton processing based on the morphological operation is often subjected to the contrast conversion processing.

A second object of the present invention is to provide an image processing apparatus that can effectively perform the above-described contrast conversion processing even when the irradiation radiation amount is changed for each image capturing condition such as the type of a subject. An object of the present invention is to provide a bone image information acquisition method capable of removing random noise.

Further, in the diagnosis of trabecular bone, the temporal change of the state of the trabecular bone is an important factor. Therefore, the chronological change obtained by sequentially photographing the same subject in time series under the same photographing condition is important. It is common practice to compare a plurality of typical radiation images. When comparing a plurality of images, the plurality of images are taken under the same imaging conditions (including the irradiation radiation amount), but the radiation absorption amount of the trabecular bone changes with time between the images. Therefore, each image is subjected to the above-described contrast conversion processing under different processing conditions. As a result, the contrast of the trabecular bone itself in the image to be finally compared is substantially the same between the plurality of images, but since the imaging conditions are the same, the contrast between the images should be essentially the same. The resulting random noise has different contrast between the images due to the difference in the processing conditions after the contrast conversion, and the state of the random noise that remains after the binarization processing of the skeleton processing based on the morphological operation differs between the images. However, there may be a problem that it is not possible to accurately grasp the temporal change in the state of the trabecular bone based on the comparison of

A third object of the present invention is to provide a method for comparing and observing bone images between images on which such a contrast conversion process has been performed, to reduce the level of contrast of random noise between the plurality of images. An object of the present invention is to provide a method for acquiring bone image information that can be easily compared as the same.

[0066]

According to a first bone image information acquiring method of the present invention, a threshold value of a threshold value process in a skeleton process is set according to an irradiation radiation dose.

That is, in the first bone image information obtaining method of the present invention, a skeleton process based on a morphological operation is performed on image information representing a radiation image, thereby enhancing or extracting a bone image in the radiation image. In the bone image information acquisition method, in performing the threshold processing in the skeleton processing, a threshold value depending on an irradiation radiation dose when a radiation image to be subjected to the skeleton processing is captured is used as the threshold value of the threshold processing. It is assumed that.

The skeleton processing based on the morphological operation refers to the processing represented by the above-mentioned equation (9) or (10).

The bone image specifically refers to an image of a trabecular bone, a bone marrow cavity, a cortical bone, or the like.

The threshold processing is a processing for removing the random noise by performing a binarization processing using a threshold suitable for separating the above-described random noise component and bone image component.

The threshold value depending on the irradiation radiation amount means a threshold value which increases as the irradiation radiation amount decreases and decreases as the irradiation radiation amount increases.

As a method of making the threshold value dependent on the irradiation radiation dose, for example, the following method may be used. Note that the term “irradiation dose” includes those that do not directly depend on the irradiation dose but eventually depend on the irradiation dose, such as radiation tube voltage. The same applies to the following description and other inventions.

That is, a look-up table in which the irradiation radiation dose is associated with a threshold suitable for effectively removing random noise when receiving the irradiation radiation dose is created in advance, and a look-up table is prepared from outside. An operator or the like inputs the radiation dose at the time of the radiation image capturing, and a threshold suitable for effectively removing random noise corresponding to the input radiation dose is referred to by referring to the lookup table. Just ask.

Here, the lookup table is created by
The irradiation radiation dose is changed stepwise to capture a so-called “solid image” without any subject, and the contrast of random noise generated in each solid image (specifically, for example, the average density value and noise of the solid image) Difference from the density value of the pixel corresponding to), and with reference to the contrast of this noise, for each irradiation radiation dose, a contrast suitable for effectively removing noise is set as a threshold value, respectively. Good.

The input of the radiation dose is not limited to the direct input of the radiation dose by the operator or the like as described above, and imaging information or the like capable of specifying the radiation dose in each radiation image information is added as additional information. The additional information may be input when the image information is read.

According to the second bone image information acquiring method of the present invention, the threshold value of the threshold processing in the skeleton processing is set based on the image of the reference object taken together with the subject in the radiation image.

That is, according to the second bone image information obtaining method of the present invention, a skeleton process based on a morphological operation is performed on image information representing a radiation image to thereby enhance or extract a bone image in the radiation image. In the image information acquisition method, in the radiation image, the radiation absorption difference between adjacent sections is set to be substantially the same as the radiation absorption of the trabecular bone along with the subject, and the radiation absorption differs stepwise between the sections. A radiation absorption amount composed of a pattern includes an image of a known reference object, a density variation width in the same section in the image of the reference object is obtained, and a threshold value processing in the skeleton processing is performed. Larger than the density fluctuation width within the same section and smaller than the density fluctuation width between adjacent sections in the image of It is characterized in that constant.

Here, the skeleton processing, the bone image, and the threshold processing based on the morphological operation are the same as those in the above-described first bone image information acquiring method of the present invention, and therefore the description is omitted.

The reference object is composed of six sections d1, d2, and d6 each having a stepwise different amount of radiation absorption as shown in FIG.
A step wedge having a structure in which d3, d4, d5, and d6 are arranged can be applied.

In the image of the reference object, there is basically no density fluctuation within the same section. However, in reality, the above-mentioned density fluctuation occurs due to the influence of the random noise described above. Contrast.

The third bone image information acquiring method according to the present invention eliminates the effect of the contrast conversion processing when the object of the skeleton processing based on the morphological operation is image information subjected to the contrast conversion processing. The skeleton processing is performed after the contrast reverse conversion processing is performed.

That is, according to the third bone image information acquiring method of the present invention, the skeleton based on the morphological operation is applied to the image information which has been subjected to the contrast conversion processing for converting the contrast of the region of interest in the radiation image into a predetermined contrast. By performing the process, in the bone image information acquisition method of enhancing or extracting the bone image in the radiation image,
Prior to the skeleton processing, the image information on which the contrast conversion processing has been performed is subjected to a contrast inverse conversion processing for eliminating the effect of the contrast conversion processing, and the image information after the contrast inverse conversion processing is performed. At the time of threshold processing in the skeleton processing, a plurality of thresholds suitable for separating a plurality of trabecular components and noise components respectively set in advance according to a plurality of imaging conditions of the radiation image, A threshold value corresponding to a radiographic image capturing condition to be performed is selected.

Here, the skeleton processing, the bone image and the threshold processing based on the morphological operation are the same as those in the first bone image information acquiring method of the present invention, and the description is omitted.

The region of interest refers to the region to be observed or the region to be observed and its vicinity, and is all or a part of the subject.

The contrast conversion processing is disclosed in
No. 9, etc., means a signal processing performed on the entire original image or an image signal (image information) corresponding to the region of interest so that the region of interest is expressed as a visible image having an optimum contrast. A process of converting into a preset contrast for each type of image (a type of a region of interest, a type of a subject, a type of a photographing position, and the like) (also referred to as a normalization process) corresponds to the image information. This is generally performed in the process of converting to an image.

[0086] The contrast reverse conversion process for eliminating the effect of the contrast conversion process is based on the premise that the contrast conversion process is a reversible process with respect to at least the region of interest. This means that the conversion processing is performed so as to obtain the contrast before the processing.

As a method for selecting a threshold, for example, the following method can be applied.

That is, a look-up table is created in advance that associates imaging conditions with thresholds suitable for effectively removing random noise when radiation is applied under the imaging conditions, and a look-up table is prepared from outside. An operator or the like inputs imaging conditions at the time of the radiographic image imaging, and selects a threshold suitable for effectively removing random noise corresponding to the input imaging conditions by referring to the lookup table. do it.

Here, the lookup table is created by
For each of a plurality of images photographed under various photographing conditions, the occurrence state of random noise is investigated, and contrast (specifically, for example, the average density value of a solid image and the density value of a pixel corresponding to noise are compared). Difference), respectively,
By referring to the noise contrast, a contrast suitable for effectively removing noise may be set as a threshold value for each irradiation radiation dose.

The input of the imaging conditions is not limited to the input by the operator or the like as described above, and imaging information or the like for specifying the imaging conditions (particularly the irradiation radiation dose) is added to each radiation image information as additional information. When reading this image information, additional information may be input.

Note that the imaging conditions need only include information that can specify at least the amount of irradiation radiation, and may include not only the amount of irradiation radiation but also the tube voltage of a radiation source.

According to the fourth method for acquiring bone image information of the present invention, a plurality of pieces of image information to be compared and contrasted which are photographed in time series and subjected to the above-described contrast conversion processing are subjected to contrast reverse conversion processing. In this way, the appearance states of random noise generated in each piece of image information are unified to the same level.

That is, the fourth method for acquiring bone image information according to the present invention provides a plurality of radiation images which are obtained by sequentially photographing the same subject in time series under the same photographing conditions and which are compared and contrasted over time. Contrast conversion processing for converting the contrast of the same region of interest in each radiographic image into a predetermined contrast is applied to each image, and each obtained image information is subjected to skeleton processing based on a morphological operation. According to the bone image information acquisition method of enhancing or extracting the bone image in each of the radiation images, prior to the skeleton processing, the effect of the contrast conversion process on each image information subjected to the contrast conversion process Each of the contrast reverse conversion processes to be eliminated is performed, and the image information after the contrast reverse conversion process is performed is applied to each image information. The threshold value of the threshold process in the skeleton process of, is characterized in that to match with each other.

Here, the skeleton processing based on the morphological operation, the bone image, the region of interest, and the contrast reverse conversion processing are the same as those in the above-described third bone image information acquiring method of the present invention, and therefore description thereof is omitted.

A plurality of chronological radiation images obtained by sequentially photographing the same subject in time series under the same radiographing condition means that, for example, the same part of the same patient is photographed under the same radiographing condition. This means a radiation image or the like used for observing a temporal change of the portion, such as a plurality of radiation images taken every month.

The bone image information obtained by each of the first to fourth bone image information obtaining methods of the present invention is not used only for reproducing a visible image, but is used as an index value for a bone structure state. May be used for the process of calculating As an index value indicating such a bone structure state, for example, an index value obtained by a star volume (Star volume), a node strut analysis (Node-strut analysis), or the like can be applied.

[0097]

According to the first bone image information acquisition method of the present invention, the threshold value of the threshold processing for separating the bone component and the noise component on the image signal in the skeleton processing is set according to the irradiation radiation dose. By setting, even when the irradiation radiation amount fluctuates for each image capturing condition such as the type of the subject, it is possible to effectively separate the random noise of the contrast depending on the irradiation radiation amount from the bone component, S /
N can be improved.

According to the second bone image information acquiring method of the present invention, the threshold value of the threshold processing in the skeleton processing is set to be larger than the density fluctuation width in the same section in the image of the reference object taken together with the subject in the radiation image. By setting it smaller than the density fluctuation width between adjacent sections, even if the irradiation radiation amount fluctuates for each image capturing condition such as the type of subject, random noise of contrast depending on this irradiation radiation amount is effective. It can be separated from the bone component, and the S / N can be improved.

According to the third bone image information acquiring method of the present invention, when the object of the skeleton processing based on the morphological operation is the image information on which the contrast conversion processing has been performed, the contrast conversion processing is performed prior to the skeleton processing. The image information before the contrast conversion processing, that is, the known imaging, is obtained even if the irradiation radiation amount is changed for each image shooting condition such as the type of the subject by performing the contrast inverse conversion processing that eliminates the effect of The image information is returned to the original image captured according to the conditions, and a skeleton process is performed by selecting an appropriate threshold value according to the known imaging conditions. Can be effectively separated from the bone component, and the S / N can be improved.

According to the fourth method for acquiring bone image information of the present invention, prior to the skeleton processing, a plurality of pieces of image information of a target to be compared and contrasted, which are photographed in time series and subjected to a contrast conversion processing, include: By performing each contrast reverse conversion process, each image information is returned to the image information before the contrast conversion process, that is, the original image information captured under the same shooting condition. The appearance state of random noise generated in each information can be unified to the same level, so that there is no difference in random noise between comparative images, and the temporal change of trabecular state based on comparison of images can be accurately grasped be able to.

[0101]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A specific embodiment of the bone image information acquiring method of the present invention will be described below with reference to the drawings.

FIG. 1 shows a morphology operation device 100 serving as a bone image information acquisition device for implementing the first bone image information acquisition method of the present invention, and a bone image information obtained by the morphology operation device 100. A bone structure index value calculating means 200 for calculating an index value (for example, an index value obtained by node strut analysis, etc.) representing the state of the structure by a numerical value, and a morphological operation device
FIG. 3 is a block diagram showing a configuration of a bone measurement device including image display means 300 for displaying a visible image represented by the image information based on the bone image information obtained by 100.

As shown in FIG. 2, the morphology operation device 100 shown in the figure performs dilation processing on the input image information S using a structural element having a size and shape suitable for extracting trabecular bone. (A (0)), and then further subjected to erosion processing (B (0)), and the result B
(0) is subtracted from the original image information (image information before dilation processing) S (S-B (0)), and the same processing is performed on the image information A (0) after dilation processing as original image information S And the image information S representing each skeleton
.., S (n) for calculating (0), S (1),.
And each processed image information S obtained by the arithmetic unit 10
(0), S (1),..., S (n) are compared with the threshold value set by the threshold value setting unit 21 to output a value exceeding the threshold value, and an output from the threshold value processing unit 20 And a second calculation unit 30 that calculates a predetermined union with respect to.

As shown in FIG. 1, the threshold setting unit 21 includes an irradiation radiation amount when the input image information is captured and a random noise generated in the image when the irradiation of the irradiation radiation amount is performed. A function suitable for effectively removing a threshold value and a function associated with the threshold value in advance as shown in the illustrated graph are stored in advance. According to the stored function, a threshold value corresponding to the input radiation dose and suitable for effectively removing random noise is obtained.

Then, the obtained threshold value effective for the removal of random noise is input to the threshold value processing unit 20, and the threshold value processing unit 20 uses the input threshold value to convert each of the difference image information S (n). Each is binarized, and the skeleton image information S
kl (n) = {Skl (0), Skl (1), Skl (2), S
kl (3), Skl (4), Skl (5), Skl (6), Skl
(7) Find｝.

The skeleton image information Skl (n) obtained by the threshold processing section 20 is input to the second arithmetic section 30,
The calculation unit 30 receives the input skeleton image information Skl
Using (i, j) satisfying i ≦ n ≦ j in (n), the logical sum SUM (i, j) = Skl (i) + Skl (i + 1) +
... Sum set image information represented by + Skl (j) is obtained. In the example shown in FIG. 2, SUM (0, 1), SUM (2,
5), SUM (3,6) and SUM (4,7).

The four sum-set image information SUM (0,1), SUM (2,5), SUM (3,3)
6), SUM (4,7) and eight skeleton image information S
kl (0), Skl (1), Skl (2), Skl (3), Skl
(4), Skl (5), Skl (6), and Skl (7) are output from the morphology operation device 100 as bone image information resulting from the skeleton processing.

Here, the threshold processing (binarization processing) by the threshold processing section 20 is performed on each processed image information S (0), S (1),..., S ( In n), a process for separating the bone component and the random noise component is performed. This threshold value in the conventional skeleton process is a constant value, while the contrast of the random noise component increases as the irradiation radiation dose decreases. To do
Depending on the magnitude of the irradiation radiation, the bone component and the random noise component may not be properly separated.

However, according to the morphological operation device 100 for implementing the bone image information acquiring method of the present invention, the threshold value increases as the irradiation radiation dose decreases, and decreases as the irradiation radiation dose increases. Therefore, the above-described problem does not occur because the random noise is set to a value suitable for effectively separating random noise.

In this embodiment, the radiation dose is input to the threshold setting unit 21. However, a signal having a close correlation with the radiation dose, such as a radiation tube voltage, may be input. However, in this case, it is necessary to set a function in which the input index is directly associated with the threshold value, or to set the correlation between the input index and the irradiation radiation dose as a function.

The input of the radiation dose and the like to the threshold value setting unit 21 is not limited to the direct input by the operator or the like as described above, but may be imaging information or the like which can specify the radiation dose in the radiation image information. May be added as additional information, and the additional information may be input when the image information is read.

The radiation dose and the threshold may be associated with each other using a lookup table or the like instead of using the above-described function.

Thus, the morphology operation device 1
The image information representing the bone component from which the random noise component has been effectively removed obtained by 00 is input to the bone structure index value calculation means 200 and the image display means 300, respectively.
The bone structure index value calculation means 200 to which the image information representing the bone component is input calculates the index value indicating the structural state of the bone component represented by the input image information by the following processing, and the image display means 300 displays a visible image representing a bone component based on the input image information.

The calculation of the index value by the bone structure index value calculating means 200 is performed, for example, by using the star volume Vt as an index value.
The calculation of is described with reference to FIG.

In FIG. 3, the hatched portion indicates the trabecular bone, and the other portions indicate the medullary cavity. Star volume, the size of the trabecular bone of the bone marrow cavity by devising the sampling method of specimen mm ³
Or a stereological index no gap expressed as the value of [mu] m ³ or three-dimensional, Vtj = (π / 3) × Σl 1 4 / Σl 1 where, l ₁ in any direction about the point j, Indicates the length of continuous trabecular bone. In addition, l ₁ is obtained radially in all directions around the point j, and Σ represents obtaining the sum in all directions (for example, 40 directions every 9 degrees). Then, Vtj is calculated for each sampling point, and the average value is set as Vt.

The index value Vt obtained by the bone structure index value calculating means 200 is large when the trabecular bone continuity is high, and becomes small when the trabecular bone is lost or perforated much, so that it is quantitative in the diagnosis of osteoporosis. In addition to the above information, it is possible to perform highly accurate diagnosis of osteoporosis in combination with the visual effect of the visible image displayed on the image display means 300.

FIG. 4 includes a morphology operation device 100 ', which is a bone image information acquisition device for implementing the second bone image information acquisition method of the present invention, a bone structure index value calculation means 200, and an image display means 300. FIG. 2 is a block diagram showing a configuration of a bone measurement device.

The morphological operation device 100 'shown in FIG.
The configuration and operation are the same as those of the morphology operation device 100 of the embodiment shown in FIG. 1 except that the threshold setting unit is different. The same applies to the bone structure index value calculation means 200 and the image display means 300. Therefore, description of components other than the threshold value setting unit 21 'in the present embodiment will be omitted.

The radiation image to be subjected to the skeleton processing by the morphological operation device 100 'of the present embodiment is one in which a step wedge SW as shown in FIG. 5 is photographed together with the radiation image of the subject.

The step wedge SW is composed of six sections d1, d2, d3, d4, d5, d6 whose radiation absorption amounts differ stepwise.
Are arranged, and the difference in radiation absorption between adjacent sections is set in advance to be substantially the same as the radiation absorption of the trabecular bone.

The threshold setting unit 21 'extracts the step wedge SW portion from the entire input radiographic image information,
Alternatively, the image information from which the step wedge SW portion has already been extracted is input to the threshold value setting unit 21.

The threshold setting unit 21 'determines the density fluctuation width (Δd) in the same section (for example, d6) and the density between adjacent sections (for example, d6 and d5) for the image information of the step wedge SW. The variation width (ΔD) is obtained.

Here, the density fluctuation width Δd in the same section
Is caused by the contrast of random noise. On the other hand, since the density variation width ΔD between adjacent sections is set to substantially match the contrast of the trabecular bone, the threshold value setting unit 21 ′ sets Δd <Th <ΔD The threshold value Th in the range of
To enter.

Thus, even when the irradiation radiation amount changes for each image capturing condition such as the type of the subject, the threshold value processing unit
20 can effectively remove the random noise of the contrast depending on the irradiation radiation dose, and can improve the S / N of the image information of the bone component.

The operation of the components below the threshold value processing section 20 is the same as that of the morphological operation device 100 of the embodiment shown in FIG.

FIG. 6 shows a morphological operation device 100 ″ as a bone image information acquiring device for implementing the third bone image information acquiring method of the present invention, a bone structure index value calculating device 200, an image display device 300, and an input device. Contrast conversion processing means for converting the contrast of a region of interest in a radiation image represented by the image information into a predetermined contrast for the obtained image information
FIG. 4 is a block diagram showing a configuration of a bone measurement device provided with a reference numeral 400.

The morphological operation device 100 ″ shown in FIG.
Contrast conversion processing means for original image information representing an original image so that a region of interest has an optimum contrast for diagnosis.
The image information that has been subjected to the contrast conversion processing by 400 is subjected to the skeleton processing.

This contrast conversion processing is also performed by changing the reading conditions when reading an image from the standard reading conditions, and by performing signal processing on image information read under predetermined reading conditions. Is also performed.

The morphological operation device of the present embodiment 10
0 ″ is a contrast reverse conversion processing unit 40 that performs contrast reverse conversion processing on input image information so as to eliminate the effect of the contrast conversion processing by the contrast conversion processing means 400 on the image information, Calculation unit 10 and threshold processing unit similar to those of the above-described embodiments.
20; a second arithmetic unit 30; and a threshold setting unit 21 ″ for setting a threshold value for threshold processing by the threshold processing unit 20.

Here, the first operation unit 1 which is the same as the constituent part in each of the above embodiments, and has the same operation and effect.
Descriptions of 0, the threshold processing unit 20 and the second calculation unit 30 are omitted.

First, the morphological operation device of the present embodiment
To 100 ″, image information that has been subjected to contrast conversion processing by the contrast conversion processing means 400 so that the contrast of the bone component, which is the region of interest, is optimal for diagnosis is input. The image information is accompanied by information indicating the content of the contrast conversion processing by the contrast conversion processing means 400, and the contrast reverse conversion processing unit 40 reads the content of the added contrast conversion processing and determines the effect of the conversion processing. An inverse conversion processing condition to be eliminated is set, and an inverse conversion process according to the inverse conversion processing condition is performed on the input image information (image information on which the contrast conversion processing has been performed).

As a result, the contrast reverse conversion processing means 40
Is basically the same as the image information before the contrast conversion processing is performed.

In principle, in the contrast conversion processing, the objective is achieved if at least only the region of interest is converted to the optimum contrast, so that the contrast conversion processing for the image region other than the region of interest is irreversible conversion ( For example, there may be a process of setting the maximum density value irrespective of the input value (irradiation field blackening process, etc.). It is impossible to return.

It should be noted that such irreversible transformation does not have a problem since it does not target a region of interest, as represented by blackening processing outside the irradiation field.

The image information that has been returned to the image information before the contrast conversion processing is performed is input to the first arithmetic unit 10 to obtain difference image information, as in each of the above embodiments. The threshold value is input to the threshold value processing unit 20, and threshold processing is performed by comparison with the threshold value set by the threshold value setting unit 21 ″.

Here, as shown in FIG. 6, the threshold setting unit 21 ″ stores shooting conditions when the input image information is shot and random noise generated in the image when shooting is performed under the shooting conditions. A look-up table in which a threshold suitable for effective removal is associated in advance is stored, and when an imaging condition at the time of radiographic image capturing by an operator or the like is input from outside, the threshold setting unit 21 ″ Refers to the stored lookup table and sets a threshold suitable for effectively removing random noise corresponding to the input photographing condition.

As a result, the threshold value processing unit 20 can effectively remove random noise from image information on which contrast conversion processing has been performed, even when image shooting conditions such as the type of subject are changed. , The S / N of the image information of the bone component can be improved.

The operation of the components below the threshold value processing section 20 is the same as that of the morphological operation device 100 of the embodiment shown in FIG.

Here, the same subject is sequentially photographed in time series under the same photographing conditions by using the morphology operation device 100 ″ as the bone image information acquiring device for implementing the bone image information acquiring method of the present embodiment. Obtained by this, for a plurality of ray images that are compared and contrasted over time, each separately subjected to a contrast conversion process of converting the contrast of the same region of interest in each radiation image into a predetermined contrast, each obtained A method for enhancing or extracting a bone image in each radiation image by performing skeleton processing based on a morphological operation on image information will be described.

As described above, each image information representing a plurality of radiation images to be compared and compared is sequentially converted into a morphological operation device.
Enter 100 ″ to acquire bone image information.

At this time, since the plurality of radiographic images are images taken in time series, the bone components, which are the regions of interest, have the same contrast even if they are taken under the same imaging conditions. Not necessarily.

On the other hand, since the contrast suitable for the diagnosis of the bone component is constant, each image is subjected to contrast conversion processing by the contrast conversion processing means 400 under different processing conditions.

Here, the contrast of the random noise of each image is the same for the images having the same shooting conditions, and the contrast of the random noise of each image subjected to the contrast conversion processing under the same processing conditions is also the same.
The plurality of images to be compared and compared with each other in which the contrasts of the random noises were substantially matched by being photographed under the same photographing conditions are subjected to contrast conversion processing under different processing conditions by the contrast conversion processing means 400. The contrast of the random noise is different, and when performing comparative observation between these images, there is a possibility that the diagnostic performance may be reduced due to the difference in the influence of the random noise on each image.

However, in the processing by the morphological operation device 100 ″ of the present embodiment, the contrast conversion processing means
Since the effect of the contrast conversion processing on each image by the 400 is eliminated by the contrast inverse conversion unit 40, the plurality of images output from the contrast inverse conversion unit 40 have the same random noise contrast.

On the other hand, since the photographing conditions input to the threshold setting unit 21 ″ are the same between the images to be compared and compared,
The threshold value set by the threshold value setting unit 21 ″ matches between these images.

Therefore, the noise level removed by the threshold processing unit 20 is the same between these images, and the noise level remaining in the image is also the same.

Therefore, the morphological operation device 100 ″ of the present embodiment is used to sequentially compare and contrast the objects obtained over time by photographing the same object under the same photographing conditions in the same photographing condition. A contrast conversion process for converting the contrast of the same region of interest in each radiation image into a predetermined contrast is performed on each of a plurality of radiation images, and a skeleton based on a morphological operation is applied to each of the obtained image information. By performing the processing, when enhancing or extracting the bone image in each of the radiation images, the effect of random noise on each of the images to be compared and matched coincides, and a decrease in diagnostic performance can be prevented.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a bone measurement device including a bone image information acquisition device that implements a first bone image information acquisition method of the present invention.

FIG. 2 is a diagram showing a skeleton process performed by a morphology operation device as the bone image information acquisition device shown in FIG. 1;

FIG. 3 is a diagram for explaining a star volume as an index value.

FIG. 4 is a block diagram showing a bone measurement device including a bone image information acquisition device that implements a second bone image information acquisition method of the present invention.

FIG. 5 is a view showing a reference object (step wedge).

FIG. 6 is a block diagram showing a bone measurement device including a bone image information acquisition device that implements the third and fourth bone image information acquisition methods of the present invention.

FIG. 7 is a diagram showing a basic operation of morphology processing;
(A) dilation processing, (B) erosion processing, (C) opening (openin)
g) processing, (D) closing processing

FIG. 8 is a diagram showing an example of a structural element Bi used for morphological processing;

FIG. 9 specifically illustrates a closing process.

FIG. 10 is a diagram showing various figures and their skeletons;

FIG. 11 is a diagram specifically illustrating skeleton processing;

[Explanation of symbols]

 10 First calculation unit 20 Threshold processing unit 21 Threshold setting unit 30 Second calculation unit 100 Morphology calculation device (image information acquisition device) 200 Bone structure index value calculation means 300 Image display device

Claims (4)

[Claims] 1. A bone image information acquisition method for emphasizing or extracting a bone image in a radiation image by performing a skeleton process based on a morphological operation on image information representing a radiation image, wherein a threshold value process in the skeleton process is performed. A bone image information acquiring method, wherein a threshold value dependent on an irradiation radiation amount when a radiation image to be subjected to the skeleton processing is captured is used as the threshold value of the threshold processing. 2. A bone image information acquisition method for emphasizing or extracting a bone image in the radiation image by performing skeleton processing based on a morphological operation on image information representing the radiation image, wherein the radiation image includes: With the subject, the radiation absorption difference between adjacent sections is set to be approximately the same as the radiation absorption of the trabecular bone. Including the image, determine the density variation width in the same section in the image of the reference object, at the time of the threshold processing in the skeleton processing, the threshold value of the threshold processing is larger than the density variation width in the same section in the reference object image, A bone image information acquiring method, wherein the bone image information acquiring method is set to be smaller than a density fluctuation width between adjacent sections. 3. A skeleton process based on a morphological operation is performed on image information on which contrast conversion processing for converting a contrast of a region of interest in a radiographic image into a predetermined contrast is performed, so that bones in the radiographic image are processed. In the bone image information acquisition method of enhancing or extracting an image, prior to the skeleton process, the image information subjected to the contrast conversion process is subjected to a contrast reverse conversion process for eliminating the effect of the contrast conversion process; At the time of the threshold processing in the skeleton processing on the image information after the contrast reverse conversion processing is performed, a plurality of trabecular components and noise components that are respectively set in advance according to a plurality of imaging conditions of the radiation image are separated. Of the thresholds suitable for Bone image information acquiring method characterized by selecting a threshold value corresponding to the photographing condition of the image. 4. With respect to a plurality of radiation images which are obtained by sequentially photographing the same subject in time series under the same photographing conditions and which are compared and contrasted with time, the same interest in each radiation image is obtained. A contrast conversion process for converting the contrast of the region into a predetermined contrast is separately performed, and the obtained image information is subjected to a skeleton process based on a morphological operation, thereby enhancing the bone images in each of the radiation images. Alternatively, in the bone image information obtaining method to be extracted, prior to the skeleton processing, each of the image information subjected to the contrast conversion processing is subjected to a contrast reverse conversion processing for eliminating the effect of the contrast conversion processing, Threshold processing in the skeleton processing for each image information after the inverse conversion processing is performed Bone image information acquisition method, wherein a threshold value, to match each other.