JPH09248293A - Diagnostic device for osteoporosis - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce processing time and to improve measurement accuracy to enable to obtain the number of trabecula, by image processing to quantify the number of trabecula based on an X-ray image. SOLUTION: When an operator gives an activation instruction of a scanner 14, image information of a bone part is obtained from a set X-ray image film and stored in work areas of an RAM 3 and a VRAM 5, and the image information is displayed on a display 13. The operator designates an area to be the object within the obtained image information. The CPU 1 scans a space filter in the horizontal line direction, calculates the average of density values to each image element of the original image, and stores the average in the RAM 3. Further, a density value stored in the RAM 3 is subtracted from each of the density data on every horizontal lines in the original image and the result is digitized by a prescribed threshold value and stored in the RAM 3. The number of changing points of the digitized data on every horizontal lines is counted and a value calculated by dividing the total value by the number of horizontal lines is made to be the trabecula number.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for diagnosing bone and osteoporosis for quantifying trabecular bone structure by image analysis of an X-ray image of a bone portion, and particularly, to shorten processing time and improve measurement accuracy. The present invention relates to a technique for obtaining the number of trabeculae.

[0002]

2. Description of the Related Art Conventionally, the most common method for discovering osteoporosis is for a doctor to see the result of X-ray photography of the bone part and judge whether the sign of osteoporosis appears in the result of photography. . However, the visual judgment requires many years of experience, and different doctors may have different diagnostic contents. Bone strength (fracture risk) is known to be related to bone mineral content and trabecular running, and the bone mineral content of bone and the condition of trabecular structure of bone (eg trabecular bone A device for diagnosing osteoporosis using a quantified distribution) has been proposed.

As a method for quantifying the degree of trabecular bone, Japanese Patent Application Laid-Open No. 6-133223 proposes a bone part image analysis method. In this method, a spectrum analysis is performed on an X-ray image, a vertical and horizontal dense trabecular image (an image that appears as a thin line) is detected from the analysis result, and the degree of the denseness is determined based on the trabecular structure. It is used as an index of the degree.
In addition, the bone mineral content of the lumbar spine is measured by the DXA method for the diagnosis of osteoporosis.

[0004]

With such a proposal, the trabecular structure can be quantified, and the osteoporosis can be diagnosed more accurately than visually.

However, the above proposal still has room for improvement in the following points. JP-A-6-133
According to the technique disclosed in Japanese Patent No. 223, etc., it is necessary to perform spectrum analysis on the image to be analyzed, and the Fourier transform performed for performing the spectrum analysis requires a long calculation time, resulting in a long analysis time. I was invited to. Especially,
If the amount of X-ray image that is the analysis target image is large,
Frequently, the operator feels annoyed.

[0006] Further, in the spectrum analysis, although the spectrum distribution has a spectrum distribution determined by the width of the trabecular bone, since the distribution spreads over the entire frequency domain, the spectra of a plurality of trabecular bones overlap each other, and the quantification accuracy is good. There wasn't.

Therefore, an object of the present invention is to provide an apparatus for quantifying the information on the trabecular structure such as the number of trabecular bones without shortening the processing time and without deteriorating the measurement accuracy.

[0008]

In order to achieve the above object, according to the invention of claim 1, when an X-ray image of a bone portion is given, the X-ray image is image-processed to provide a trabecular structure. Is an osteoporosis diagnostic apparatus for quantifying the condition of
Image processing means for performing image processing for quantifying the number of trabecular bones in the state of trabecular bone structure based on the radiographic image, and output processing of information including at least the quantified result by referring to the quantified result. It is an osteoporosis diagnostic device provided with the output means to perform.

According to a second aspect of the present invention, the image processing means according to the first aspect sequentially performs the moving average processing of a predetermined number of pixels, while performing the moving average processing on all the pixels of the X-ray image. Filtering means for performing, subtraction means for performing, for each horizontal line of the X-ray radiographed image, subtraction processing for subtracting the result of moving average processing by the filter means from density data of the X-ray radiographed image, and subtraction processing results for each horizontal line Is binarized by a predetermined threshold value and the binary data, and the number of change points that change from a low level value to a high level value or vice versa is leveled. Bone prosthesis characterized by comprising a change point measuring means obtained for each line and a calculating means for quantifying the state of the trabecular structure in which the total number of change points is divided by the total number of horizontal lines to determine the number of trabecular bone structures. It is a symptom diagnosis device.

According to a third aspect of the present invention, in the filter means according to the second aspect, the predetermined number of pixels, which is the size for performing the moving average process, is approximately twice the trabecular width to be quantified. It is preferable that the osteoporosis diagnosing device is characterized in that it is composed of a muscle enhancement filter selected in size.

According to a fourth aspect of the present invention, the image processing means according to the first aspect performs the moving average processing on all the pixels of the X-ray image while sequentially performing the moving average processing on a predetermined number of pixels. A first filtering unit that performs the moving average process and a second filtering unit that performs the moving average process on all the pixels of the X-ray image while sequentially performing the moving average process on the number of pixels larger than the predetermined number of pixels; For each horizontal line of the X-ray photographed image, first subtraction means for performing subtraction processing for subtracting the result of the moving average processing by the first filter means from the density data of the X-ray photographed image; Second horizontal subtraction means for subtracting the result of the moving average processing by the second filter means from the density data of the X-ray image for each horizontal line, and horizontal lines formed by the first and second subtraction means. Every By paying attention to each of the two kinds of binary data and the two kinds of binary data which binarizes the arithmetic processing result with a predetermined threshold value and obtains two kinds of binary data, from the low level value to the high level value. A value obtained by dividing the sum of the change points of the first binary data by the total number of horizontal lines, and a change point measuring means for obtaining the value, or vice versa, for each horizontal line. An osteoporosis diagnosing device comprising: a calculation means for quantifying a state of a trabecular structure in which a value obtained by subtracting a value obtained by dividing a total sum of change points of value data by a total number of horizontal lines is included. Is.

According to a fifth aspect of the present invention, the image processing means according to the first aspect performs a moving average process on all the pixels of the X-ray image while sequentially performing a moving average process on different numbers of pixels. And the subtraction processing for subtracting the result of the moving average processing by the filter means from the density data of the X-ray image, for each horizontal line of the X-ray image.
Pay attention to the subtraction means performed for each filter means, the binarization means for binarizing the subtraction processing result for each filter means for each horizontal line with a predetermined threshold value, and the binary data for each filter means. By dividing the sum of the change points for each filter means and the change point measuring means for obtaining the number of change points changing from the low level value to the high level value or vice versa for each horizontal line by the total number of horizontal lines. The obtained value is taken as the number of trabecular bones detected by the filter means, and the value obtained by multiplying the number of trabecular bones detected by each filter means by a coefficient determined for each filter means is added as the trabecular bone structure number. An osteoporosis diagnostic device, comprising: an arithmetic means for quantifying the condition.

Further, the invention according to claim 6 is, in addition to the invention according to any one of claims 2, 3, 4 and 5, a memory in which a correspondence relationship between the number of trabecular bones and the content of diagnosis of osteoporosis is registered. The osteoporosis diagnosing device is characterized in that a means is provided and the output means performs a process of referring to the registered content of the storage means and outputting the diagnostic content corresponding to the obtained number of trabecular bones.

[0014]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 shows a hardware configuration diagram of the device according to the present embodiment.

The image processing apparatus 10 carries out various processes C
PU1, a ROM2 for storing data referred to by the CPU1, a system program and the like, a RAM3 functioning as a work area for various processes including image processing, and a hard disk storage device for storing at least a diagnostic program which is a diagnostic program. (HDD) 4, VRAM 5 for storing display data, and input / output interface (I / O) which is an interface with an external device.
6, 7, and 8. Each component is connected to the bus so that data can be communicated with each other.

A keyboard 11 (mouse 12) for giving various instructions and data, a display 13 for displaying and outputting the display data stored in the VRAM 5, and a scanner 1 for reading the image information of the X-ray image film.
4 is connected to the bus via I / O6, I / O7, and I / O8, respectively.

Further, the HDD 4 stores a diagnostic message table 4A in which the number of trabeculae and the progression level of osteoporosis are registered in association with each other. In FIG.
The example of the registered content of a diagnostic message table is shown. The number of trabecular bones and the Itami index indicating the progression level of osteoporosis, which is the diagnostic content, are registered in association with each other. Note that this example is a case where thin trabecular bones are measured, and the number of them is negatively correlated with the BMD, and it goes without saying that the registered content of the diagnostic message table is not limited to this example.

The diagnostic program includes an image analysis program for quantifying the trabecular bone structure, a display program for displaying various kinds of information such as guidance information and diagnostic results, and the like. , The ROM 2 stores the system program as a keyboard 11 (mouse 1
2) An instruction control program that controls the operation of each component according to an instruction given via the CPU 1 and an external device (keyboard 11, display 13, scanner 1).
Various programs such as an operation control program for controlling the operation of each component including 4) are included.

As shown in FIG. 1, this apparatus can be realized by various electronic devices, and can be realized by one information processing apparatus such as a personal computer or a workstation.

Next, the functions of the main components will be described. The CPU 1 controls the device according to the system program stored in the ROM 2 and analyzes the bone part image according to the image analysis program in the diagnostic program stored in the hard disk storage device (HDD) 4 to quantify the trabecular structure. Turn into.

The CPU 1 further refers to the diagnostic message table 4A stored in the HDD 4 to diagnose osteoporosis. The RAM 3 has a function of temporarily storing the work area used when the CPU 1 performs various processes and the image information captured by the scanner 14.

The HDD 4 stores a diagnostic program, a diagnostic message table, and image information of the bone taken in by the scanner 14 or the like. The VRAM 5 includes image information displayed on the display 13 and the scanner 14
The image information of the image analysis target captured by is stored. The CPU 1 performs a display process of displaying the storage content of the VRAM 5 on the display 13 according to a system program or the like.

Input / output interfaces (I / O) 6, 7,
8 is a bus 11 along with a keyboard 11 (mouse 12
Information is transferred between the display 13, the scanner 14 and the components of the image processing apparatus 10.

The keyboard 11 (including the mouse 12) is
It receives operation instructions to the CPU 1 and information necessary for image analysis processing. A system program is created so that this information input can also be performed by operating the mouse 12 to position the display information on the screen of the display 13.

The display 11 displays guidance information which is information input guidance for the operator, information input through the keyboard 11, feature values of the trabecular bone structure which are the processing results of the CPU 1, for example, a CRT, a liquid crystal display. It can be realized by a display device or the like.

The scanner 14 reads the X-ray image film, converts the read density into multivalued image information, and supplies it to the VRAM 5 and the RAM 3. A configuration in which an X-ray imaging device is directly connected to the I / O 8 is also conceivable, but in this case, an output signal obtained by analog-digital conversion of an electric signal from an X-ray sensor that constitutes the X-ray imaging device is It is given to the image image processing apparatus 10.

Now, a series of processes of this embodiment will be described. FIG. 2 is a flowchart showing a processing procedure executed by the diagnostic program. Now, when the operator gives a command to start the device through the keyboard 11, the CP
The U1 loads the diagnostic program into the area in the RAM 3 and operates according to the loaded diagnostic program.

First, in step S10, the CPU 1
Displays a menu screen on the display 13. Specifically, the CPU 1 stores the information for displaying the menu screen in the area of the VRAM 5, so that the osteoporosis diagnosis menu is displayed on the display screen of the display 13.

In this embodiment, the state of trabecular bone structure is shown.
In the menu that quantifies and outputs the number of trabecular bones (interval), which is a menu that quantifies the trabecular structure, and a menu that displays and outputs the progression level of osteoporosis based on the quantified results. Two menus of a certain "menu for diagnosing osteoporosis" are prepared, and these two types of menus are displayed on the display 13.

Next, in step S20, the operator designates a desired menu via the keyboard 11. At this time, the CPU 1 stores the information of the instructed menu in the work area of the RAM 3, for example.

The calculation process of the feature amount of the trabecular bone structure in the next step S30 is a process in which the CPU 1 quantifies the number (width) of trabecular bones showing the state of the trabecular bone structure, and which menu is selected. This is a process that is executed even if is selected. Since this processing is a main part of the present invention, it will be described in detail later.

Next, in step S40, the CPU 1
Determines whether a diagnosis of osteoporosis is necessary. Specifically, the CPU 1 performs the following processing. That is, CP
The U1 refers to the menu information stored in the RAM 3, and if the instructed menu is “a menu for quantifying the trabecular bone structure”, the U1 proceeds to step S50 and determines the trabecular bone obtained in step S30. Display the structural features 13
Display output to and end the process (end).

On the other hand, if the instructed menu is "a menu for diagnosing osteoporosis", the process proceeds to step S45 to diagnose the progression level of osteoporosis based on the measured data, that is, the quantification result. , Further, step S
The process proceeds to 50 and the diagnostic result is displayed and output on the display 13 to end the process (END).

The process in step S45 will be described in detail as shown in FIG. That is, step S
Since the feature amount of the trabecular bone structure has already been obtained in 300, the CPU 1 refers to the diagnostic message table 4A, extracts the corresponding diagnostic message from the HDD 4 (step S300), and converts this into image information. (Step S310), the data is stored in the VRAM 5 and displayed on the display 13 (step S320). At this time, the CPU 1 may be configured to display the image information captured by the scanner 14 and the guide information used for osteoporosis diagnosis on the display 13.

FIG. 13 shows a state of the display screen when such a display is performed, and 101 is the scanner 1
Image information (bone image) captured from No. 4, 102
Is guide information used for osteoporosis diagnosis, and the guide information 102 corresponds to five samples (102A) of trabecular bone images corresponding to the feature amount of trabecular bone information called Itami index, corresponding to each sample. Itami index (102B) and the explanation message of the state of trabecular bone structure corresponding to each Itami index (1
02C). And 103 is a step S45.
It is the diagnostic message obtained in (Diagnostic name).

In this way, by displaying the diagnostic message and the guide information and the like used for the diagnostic on the same screen, convenience of the diagnostic is achieved. As described above, by performing the series of processing shown in FIG. 2, the number of trabecular bones showing the state of the trabecular bone structure is quantified and osteoporosis is diagnosed.

Now, the process of step S30 of FIG. 2, which is a main part of the present invention, will be described with reference to the drawings.
FIG. 3 is a flowchart for explaining the process of step S30 of FIG. 2 in more detail.

In step S100, when the operator sets the X-ray image film on the scanner 14 and gives an instruction to activate the scanner 14 via the keyboard 11, the scanner 14 determines the bone part from the set X-ray image film. The image information of the bone part is acquired and stored in the work area of the RAM 3, and stored in the VRAM 5 to display the image information of the bone on the display 13.

It is to be noted that the CPU 1 is configured so as to filter the acquired image information in order to remove noise components such as unnecessary signals mixed in during image reading and dust on the X-ray film. It is preferable.

Further, in step S100, a guide screen for inputting parameters required for the subsequent analysis processing is displayed. The parameters input here will be described later.

Next, step S110 executed by the CPU 1,
The process of obtaining the feature amount of the trabecular bone structure performed in 120 and 130 will be described. First, in step S110, the operator designates an area to be analyzed in the acquired image information. For example, vertical 1200 pixels, horizontal 1200
When the image information of pixels is displayed on the display 13, the rectangular area formed of a predetermined number of pixels is designated so that the maximum height is 1200 pixels and horizontal 1200 pixels.
Such designation is performed by inputting the number of vertical and horizontal pixels and the position coordinate of the upper left point of the rectangular area with the keyboard 11, or operating the mouse 12 so that the rectangular area is formed in the display image. Then, by clicking the positions of the upper left point and the lower right point, it is possible to specify the quadrangle area having the upper left point and the lower right point as vertices as the analysis target area.
You just need to create a diagnostic program.

In step S120, the operator inputs parameters required for subsequent processing. Examples of such parameters include the size of a spatial filter (hereinafter, simply referred to as “filter” as appropriate) used for extracting trabecular bone, a threshold value for binarizing image data, and image data within a predetermined size. An example is a small figure removal area having a predetermined size for removing the set from the original image. In addition,
The image processing performed using such parameters will be described later. The spatial filter here is a line enhancement filter that obtains a moving average of the densities of a predetermined number of pixels, and this line enhancement filter is hereinafter simply referred to as a spatial filter.

In step S130, step S110
Image processing is performed on the two-dimensional image data of the analytic symmetric region designated by 1. to obtain the trabecular feature amount. This will be described in more detail with reference to FIGS. After the process of step S130 is completed, the process returns and returns to step S40.

The following process is a process performed by the CPU 1 according to the image analysis program. Now, assume that the size of the original image (referred to as “image 0”) to be subjected to image analysis is 100 pixels vertically and 100 pixels horizontally, for example. Therefore, the number of trabecular bones in the image data of 100 pixels in the vertical direction and 100 pixels in the horizontal direction is obtained.

The size of the median filter is 3 ×
With 3 pixels (9 pixels in total), the median filter is
It has a function of obtaining an average value for every 9 pixels and setting the obtained average value as the density value of the central pixel of the median filter. So if you scan this filter in the horizontal line direction,
The moving average of the concentration on the line is obtained.

Now, in step S131, the image 0
In order to perform an averaging process for obtaining an average value for all pixels of
RA is sequentially calculated to obtain an average value for each pixel and RA
Store in M3. The image that has been subjected to the moving average processing by the spatial filter is referred to as image 1.

FIG. 7A shows the image 0 which is the original image. The illustrated example shows a state in which one trabecular bone exists in the vertical direction. Further, FIG. 7B shows a density distribution (solid line) in a certain horizontal line LL ′ and a spatial filter (filter size AB) in the horizontal line LL ′ direction.
The moving average processing result (dotted line) is shown. That is,
The dotted line in FIG. 7B shows the density distribution of a certain horizontal line LL ′ in the image 1.

Next, in step S132, the image 0
RAM3 from the density data of image 0 for each horizontal line of
The subtraction process for reducing the result of the moving average process performed by the spatial filter stored in is stored. The CPU 1 stores the result of this subtraction processing in the RAM 3 for each horizontal line.
The image thus subtracted is referred to as image 2.

FIG. 7C shows the state of the density distribution of a certain horizontal line LL 'in the image 2. Next, in step S133, the subtraction processing result for each horizontal line stored in the RAM 3 is binarized by a predetermined threshold value. For this purpose, the subtraction processing result is compared with the threshold value T as shown in FIG. Specifically, the CPU 1 generates binary image data by setting, in the image 2, pixels having a density equal to or higher than the threshold value to “1” and other pixels to “0”. In addition, CPU
1 stores such binarized data for each horizontal line. In the example of FIG. 7C, a process of adding a predetermined value to the subtraction result is performed so that the subtraction result does not become negative.

FIG. 7D shows an example of binarized data for a certain horizontal line LL '. Next, in step S134, paying attention to the binary data, the number of change points changing from a low level value to a high level value is obtained for each horizontal line. The CPU 1 stores the number of change points for each horizontal line in the RAM 3. In the example of the binarized data for a certain horizontal line LL ′ shown in FIG. 7D, the number of change points changing from a low level value to a high level value is one (illustrated by an upward arrow).

Generally, as shown in FIG. 9, for each horizontal line, the number of change points at which the low level value (“0”) changes to the high level value (“1”) is calculated, and the horizontal level is changed. It may be the number of trabecular bones in the line. It should be noted that the number of change points that change from a high level value (“1”) to a low level value (“0”) for each horizontal line may be obtained and used as the number of trabecular bones on the horizontal line.

Next, in step S135, the CPU
1 is the sum of the number of change points stored in the RAM 3, and the value obtained by dividing the calculated sum total value by the total number of horizontal lines (100 from 100 pixels in the vertical direction) is taken as the number of trabecular bone RA.
Store it in M3.

By performing the above processing, the number of trabecular bones existing in the analysis target image can be obtained. In addition,
The CPU 1 can quantify the distance between the trabeculae (corresponding to the width information of the trabeculae) by obtaining the distance between the change points with reference to the binarized data stored in the RAM 3. Is. For this purpose, the CPU 1 may refer to the binarized data and create an image analysis program so as to obtain the interval between the change points. Of course, the distance between the trabeculae is the characteristic amount of the trabecular bones, so that it should be displayed on the display 13, or the characteristic amount and the diagnosis content should be associated with each other and registered in the diagnosis message table so that the diagnosis can be performed. You can leave it.

Here, the influence of the filter size of the spatial filter (median filter) on the accuracy of this processing will be described. Generally, due to the nature of the spatial filter, it is extremely difficult to extract trabecular bone having a thickness equal to or larger than the filter size, and trabecular bone sufficiently smaller than the filter size is relatively difficult to extract. It is easy to extract the trabecular bones with. This will be described with reference to FIG.

FIG. 8A shows an original image which is an analysis target image. This illustrated example is an example in which thicker trabecular bones are present on both sides of a thin trabecular bone that is present near the center in the vertical direction.

A certain horizontal line L in this original image
The concentration distribution at L'is shown in FIG. Next, FIG. 8C shows the density distribution on a certain horizontal line LL ′ of the image obtained by performing the filter processing described in step S131 on this original image.

The right side of FIG. 8 (c) shows the result of processing by the first filter, which is smaller in size.
Further, the left diagram of FIG. 8C shows the processing result by the second filter having a larger filter size and a larger size than the first filter. In each case, the solid line is the actual concentration distribution on a certain horizontal line LL ′, and the dotted line is
The processing result by the filter is shown.

Next, the right diagrams of FIGS. 8D and 8E are shown in FIG.
In the density distribution as shown on the right side of (c), the subtraction process of step S132 and the binarization process of step S133 are performed. Similarly, next, as shown in FIG.
The left diagram of (e) is the subtraction process of step S132 in the density distribution as shown in the left diagram of FIG.
This shows how the binarization processing of 133 is performed.

Here, the filter size of the smaller filter is slightly larger than that of the thin trabecular bone, and the filter size of the larger filter is about twice that of the thick trabecular bone. As shown in the right series of FIG. 8 (right diagrams of FIGS. 8C, 8D, and 8E), when the trabecular bone width is larger than the filter size, the concentration distribution The peak value becomes smaller and is removed by the binarization process. Specifically, the thick trabecular bone is removed from the image information.

On the other hand, the left series of FIG. 8 (FIG. 8 (c),
As shown in (d) and (e) left diagrams), trabecular bone with a filter size of half or less has a slightly smaller peak value of the concentration distribution of the trabecular bone after the filtering process is executed, but Since the thickness is small and the peak value of the concentration distribution is small,
Frequently, it is removed by the binarization process.

Therefore, it becomes easy to extract a trabecular bone having a thickness of about half the filter size, and in measuring the number of trabecular bones having a desired thickness, the filter size is taken as the trabecular bone to be measured. It is preferable that the width is approximately half the width.

In order to measure the number of thin trabecular bones, a method of setting a small filter size in accordance with this can be considered. However, when the thin trabecular bones do not exist, bones thicker than a desired thinness can be considered. Since the number of beams will be measured,
The embodiment described below provides a process for measuring only the number of trabecular bones having a desired thinness.

FIG. 5 shows the flow of processing, and FIG. 10 shows an explanatory view of processing. First, it is assumed that the first spatial filter (median filter) and the second spatial filter (median filter) are used.

Then, the filter size Fs2 of the second spatial filter is set to the filter size F of the first spatial filter.
It is made larger than s1 (Fs1 <Fs2). First, in step S200, the number of trabecular bones is measured using the first spatial filter. The basic processing is step S131.
13 to 135, here, between steps S133 and 134, a small figure removing process, which will be described later, is performed to correct the binarized data, and noise components are removed. .

First, in step S131, in order to perform the averaging process for obtaining the average value for all the pixels of the original image, the average value for each pixel is obtained while sequentially moving the first spatial filter and stored in the RAM3. Keep it.

FIG. 10A shows an original image. As shown in the figure, a thin trabecular bone exists in the center and a relatively thick trabecular bone exists around it. Further, FIG. 10B schematically shows the processing result by the first spatial filter.

Next, in step S132, the RAM 3 is extracted from the density data of the original image for each horizontal line of the original image.
Subtraction processing for reducing the result of the averaging processing stored in the first spatial filter is performed. The CPU 1 stores the result of this subtraction processing in the RAM 3 for each horizontal line.

Next, in step S133, the RAM
The subtraction processing result for each horizontal line stored in 3 is binarized with a predetermined threshold value. Next, small figure removal processing is performed by steps not shown.

This processing is performed for binary image data
When the aggregate area of the “1” data is smaller than the small figure removal area (Sa1) which is the size of the area given in advance, all the “1” data are set as the “0” data and the binary image data is created. It is a process of fixing. Specifically, the CPU 1
First, referring to the data stored in the RAM 3, the number of pixels in which the data “1” is connected is checked for each horizontal line, and then the connection state of the data “1” between the horizontal lines is checked. When it is determined that there is an area in which the number of pixels to which the data "1" is connected exceeds a predetermined number, the small figure removal processing is performed by using the area as a target area of the small figure removal processing. This process is effective for removing noise components. Of course, the CPU 1 stores the binary data subjected to the small figure removal processing in the RAM 3.

Next, in step S134, paying attention to the binary data subjected to the small figure removal processing, the number of change points at which the low level value changes to the high level value is obtained for each horizontal line. The CPU 1 stores the number of change points for each horizontal line in the RAM 3. Next, step S13
In 5, the CPU 1 obtains the total sum of the number of change points stored in the RAM 3, and further stores the value obtained by dividing the obtained total sum value by the total number of horizontal lines in the RAM 3 as the number of trabecular bones. By performing the above processing, the number of trabecular bones is obtained by the first spatial filter (measurement result B).

Next, in step S210, the number of trabecular bones is measured using the second spatial filter (measurement result A). The contents of processing using the second spatial filter are basically the same as those in step S200, and therefore, redundant description will be omitted. Of course, the size of the spatial filter (Fs
2), the small figure removal area (Sa1 <Sa2) is changed. Note that FIG. 9B schematically shows the processing result by the second spatial filter.

Next, in step S220, from the measurement result B, which is the number of trabecular bones obtained using the first spatial filter, to the measurement result A, which is the number of trabecular bones obtained using the second spatial filter. Is subtracted to obtain the number of thin trabecular bones. This state is schematically shown in FIG.

As described above, the number of trabecular bones having a desired thickness can be accurately measured by using two spatial filters having different sizes. Generally, it is possible to perform processing using three or more types of spatial filters as shown below.

Now, using a spatial filter having a filter size of Fsi and a small figure removal area SAi, step S130
The number of trabecular bones measured by performing the process of step S135 (including the small figure removing process) is Ni. However, i is a parameter that determines the type of spatial filter. Therefore, assuming that n spatial filters are used, i
= 1, 2, ..., N.

Now, n spatial filters ((Fs1, S
A1), (Fs2, SA2), ..., (Fsi, SA
i), ..., (Fsn, SAn)), the number of trabecular bones measured using “N1, N2, ..., Ni ,.
The result obtained by multiplying the coefficients ai determined for each filter and obtaining the sum is set as A, and this is set as the number of trabecular bones obtained from the analysis target image.

A = Σ (ai × Ni) (where Σ is i =
In the description of the above-mentioned embodiment, when one spatial filter is used, “n = 1, a1 = 1”, and when two spatial filters are used, , "N = 2, a1 = 1, a2 =-
1 ".

Now, an example of how to determine the coefficient ai will be described. In this embodiment, the number of trabecular bones and BMD
It is known that (Bone Mineral Density) has a positive or negative correlation depending on the filter size and the size of the small figure removal area. If there is a positive correlation, "a
i = 1 ”, on the other hand, if there is a negative correlation,“ ai =
-1 ”, the number of trabecular bones obtained from the analysis target image A
Should be obtained.

In FIG. 11, the filter size "11 pixels × 1
The spatial filter A having 1 pixel "and the small figure removal area of" 0 pixel ", and the spatial filter B having the larger filter size" 31 pixels x 31 pixels "and the small figure removal area of" 25 pixels ". The relationship between the number of trabecular bones and BMD, obtained by using each of them, is shown.

The measurement results for the spatial filter A and the spatial filter B are shown by black circles and white triangles, respectively. According to the data analysis, the BMD indicated by the data plotted by the black circles
The correlation with B has a positive correlation of about 0.05, while the correlation with BMD indicated by the data plotted by the white triangles has a negative correlation of about -0.4. Therefore, the spatial filter A, The coefficients for the spatial filter B are “ai =
If 1 ”and“ ai = −1 ”are set and the number A of trabecular bones is obtained, this coefficient may be used.

In this way, the value obtained by multiplying the number of trabecular bones detected using each spatial filter by the coefficient determined for each spatial filter and adding the result is used as the number of trabecular bones to quantify the state of the trabecular bone structure. It is possible to improve the accuracy of quantification.

In the embodiment described above, the storage means is the HDD 4, the output means is a step including at least step S50, and the image processing means is step S13.
Corresponds to 0. Further, the filter means, the subtraction means, the binarization means, the change point measurement means, and the calculation means respectively have step S
It corresponds to 131, S132, S133, S134, and S135.

As described above, according to the present invention,
It is possible to provide a device that quantifies information on trabecular structure such as the number of trabecular bones without shortening the processing time and without deteriorating the measurement accuracy.

The filter means for improving the image quality is not limited to the median filter, and other types of spatial filters such as a smoothing filter having a large weight in the center can be used.

In the above description, the line for measuring the number is set to the horizontal direction, and the description has been limited to the measurement of the number of trabecular bones in the vertical direction.
It is also possible to measure the number of trabecular bones in a direction other than the vertical direction by inclining the line for measuring the number.

[0085]

As described above, according to the first aspect of the present invention, the state of the trabecular bone structure is quantified by obtaining the number of trabecular bones by the image processing means. Since image processing using a spatial filter is performed without performing complicated processing, it is possible to realize a device that does not require a long processing time.

According to the second aspect of the invention, for each horizontal line of the X-ray photographed image, the result of the averaging process is subtracted from the density data of the X-ray photographed image, and the subtraction result is binarized.
Furthermore, the number of trabeculae that changes from a low level value to a high level value is obtained for each horizontal line, and the total number of change points is divided by the total number of horizontal lines to obtain the number of trabecular bones automatically. be able to. At this time, as in the invention described in claim 3, since the size of the filter means is set to be approximately half the width of the trabecular bone to be quantified, the bone having a specific width Beams can be extracted efficiently.

According to the invention described in claim 4, from the number of trabecular bones obtained by using the first filter means using the two filter means having different sizes, the second filter having a larger size than the first filter means is used. By reducing the number of trabecular bones obtained by using the filter means of No. 2, it becomes possible to obtain the number of trabecular bones having a fineness that cannot be extracted by the second filter means.

According to the fifth aspect of the present invention, a plurality of filter means having different sizes are used, and a value obtained by multiplying the number of trabecular bones detected by each filter means by a coefficient determined for each filter is added as the number of trabecular bones. By quantifying the trabecular structure, the accuracy of quantification can be improved.

According to the sixth aspect of the invention, since the correspondence relationship between the number of trabecular bones and the content of diagnosis of osteoporosis is registered, if the number of trabecular bones is obtained, the corresponding diagnostic content is output. It is possible to realize a diagnostic device that does.

[Brief description of drawings]

FIG. 1 is a configuration diagram showing a system configuration of an embodiment according to the present invention.

FIG. 2 is a flowchart illustrating a processing procedure executed by a CPU 1;

FIG. 3 is a flowchart showing a processing procedure executed by a CPU 1.

FIG. 4 is a flowchart showing a processing procedure executed by the CPU 1.

FIG. 5 is a flowchart illustrating a processing procedure executed by a CPU 1;

FIG. 6 is an explanatory diagram of the principle of trabecular bone extraction processing.

FIG. 7 is an explanatory diagram of the influence of the filter size on the extraction result.

FIG. 8 is an explanatory diagram of the principle of measuring the number of trabecular bones.

FIG. 9 is an explanatory diagram of the principle of measuring fine trabecular bone by combining two types of filters having different filter sizes.

FIG. 10 is an explanatory diagram showing a relationship between a measurement result and BMD.

FIG. 11 is an explanatory diagram showing the relationship between the measurement results of fine trabecular bone and BMD.

FIG. 12 is an explanatory diagram of a display example of a display.

FIG. 13 is an explanatory diagram showing an example of registered contents of a diagnostic message table.

[Explanation of symbols]

 1 CPU 2 ROM 3 RAM 4 HDD 5 VRAM 6, 7, 8 I / O 10 Image Processing Device 11 Keyboard 12 Mouse 13 Display 14 Scanner

Front page continuation (72) Inventor Manabu Tanigawa 1 Asahi Kasei Kogyo Co., Ltd. 2 at Samejima, Fuji City, Shizuoka Prefecture (72) Inventor Yoshio Fukunaga 193-1 Futako, Kurashiki City, Okayama Prefecture

Claims (6)

[Claims] 1. An osteoporosis diagnostic apparatus for quantifying the state of trabecular bone structure by image-processing an X-ray image of a bone portion, the X-ray image being obtained. Based on, image processing means for performing image processing to quantify the number of trabecular bones in the state of trabecular structure,
An osteoporosis diagnostic apparatus comprising: an output unit that refers to the quantification result and performs an output process of information including at least the quantification result. 2. The image processing means according to claim 1, wherein the image processing means performs moving average processing on all the pixels of the X-ray image while sequentially performing moving average processing on a predetermined number of pixels, and the X means. For each horizontal line of the radiographic image, subtraction means for performing subtraction processing for subtracting the result of the moving average processing by the filter means from the density data of the X-ray radiographic image, and the subtraction processing result for each horizontal line are predetermined thresholds. A binary point conversion means for binarizing by a value and the binary point data, and a change point measurement for each horizontal line to obtain the number of change points changing from a low level value to a high level value or vice versa. Means and
An osteoporosis diagnosing device comprising: a calculation means for quantifying a state of a trabecular structure in which a value obtained by dividing a total sum of change points by a total number of horizontal lines is used. 3. The filter means according to claim 2, wherein the predetermined number of pixels, which is a size for performing moving average processing, is selected to be approximately twice the trabecular width to be quantified. An osteoporosis diagnostic device comprising a muscle enhancement filter. 4. The first filter means according to claim 1, wherein the image processing means performs moving average processing on all pixels of the X-ray image while sequentially performing moving average processing on a predetermined number of pixels. Second filter means for performing moving average processing on all pixels of the X-ray image while sequentially performing moving average processing on a number of pixels larger than the predetermined number of pixels,
First subtraction means for performing subtraction processing for subtracting the result of the moving average processing by the first filter means from the density data of the X-ray radiographic image for each horizontal line of the X-ray radiographic image;
Second subtraction means for performing subtraction processing for each horizontal line of the X-ray radiographed image from the density data of the X-ray radiographed image to subtract the result of the moving average processing by the second filter means,
The subtraction processing result for each horizontal line by the first and second subtraction means is binarized by a predetermined threshold value, and two kinds of 2
Paying attention to each of the two kinds of binary data and the binary conversion means for obtaining the value data, the values from the low level to the high level,
Alternatively, on the contrary, from the value obtained by dividing the sum of the changing points of the first binary data by the total number of horizontal lines and the changing point measuring means for obtaining the number of changing points for each horizontal line, the second binary data is obtained. An osteoporosis diagnosing device comprising: a calculation means for quantifying the state of the trabecular structure in which the number of trabecular bones is obtained by subtracting a value obtained by dividing the sum of the change points of the total number of horizontal lines. 5. The X-ray processing system according to claim 1, wherein the image processing unit sequentially performs moving average processing for different numbers of pixels.
A plurality of types of filter means for performing moving average processing on all pixels of the radiographic image, and subtracting the result of the moving average processing by the filter means from the density data of the radiographic image for each horizontal line of the radiographic image. Subtraction means for performing the subtraction processing for each filter means, binarization means for binarizing the subtraction processing result for each filter means for each horizontal line with a predetermined threshold value, and two for each filter means
Paying attention to the value data, the change point measuring means for obtaining the number of change points changing from the low level value to the high level value or vice versa for each horizontal line, and the sum of the change points for each filter means are calculated. The value divided by the total number of horizontal lines is the number of trabecular bones detected by the filter means, and the value obtained by multiplying the number of trabecular bones detected by each filter means by a coefficient determined for each filter means and adding it And an arithmetic means for quantifying the condition of the trabecular bone structure, and an osteoporosis diagnostic device. 6. The storage unit according to claim 2, further comprising a storage unit for registering a correspondence relationship between the number of trabecular bones and a diagnosis content of osteoporosis, and the output unit includes the storage unit. Refer to the registration contents of the means,
An osteoporosis diagnosing device characterized by performing a process of outputting a diagnostic content corresponding to the obtained number of trabecular bones.