JP2000126178A - Method of quantifying stereo surface shape and automatic identification method of malignant tumor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enable discovery of the region of a tumor such as mastadenoma at a high accuracy by calculating a shape judging parameter utilizing the stereo surface area and volume as index for measuring the degree of unevenness of the surface shape of the solid body to be compared with a threshold when judging the degree of the uneveness of the surface shape of the solid body based on a voxel data. SOLUTION: In the examination of a cancer of the breast, by an examination system 10 containing a racking means 14, a three-dimensional coordinate conversion means and a three-dimensional voxel data generation means 16, a mammary gland 301 is extracted from tan ultrasonic echo obtained using a three-dimensional position sensor 13 to identify the benignancy or malignancy of a tumor from the three-dimensional surface shape thereof. In this process, a three-dimensional ultrasonic image data 15b is generated using an ultrasonic tomographic image data and a probe coordinate data 13a and is converted to an isotropic voxel data 16a by a linear image interpolation. A parameter of a ratio between the stereo surface area S and the volume V is defined and compared with a specified threshold a judge the character of the unevenness, namely, the region of the tumor.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for evaluating a surface shape in a three-dimensional image, and in particular, quantitatively measures and evaluates geometrical irregularities (autumn) of a three-dimensionally displayed three-dimensional surface. On how to do it.

[0002] The present invention also relates to a method for quantitatively measuring and evaluating the geometric unevenness of the three-dimensional surface of a three-dimensionally displayed tumor to assist the identification of the tumor, and more particularly, to a magnetic resonance image of a living body. Magnetic Resonance Ima
ging: an MRI image) or a pathological diagnosis support for extracting a boundary between tissues represented by a three-dimensional image from a two-dimensional tomographic image such as an ultrasonic image and finding a cancer tissue (particularly, a breast cancer tissue) from normal tissues. The present invention relates to a method for automatically identifying a malignant tumor applicable to a system.

[0003]

2. Description of the Related Art In recent years, techniques for creating three-dimensional image data of a living body using an ultrasonic diagnostic apparatus or medical MRI and performing a pathological diagnosis process have been increasingly used.

[0004] In circulatory disease diagnosis for the heart and carotid arteries, three-dimensional movement of the inner and outer walls of a blood vessel with time and the three-dimensional distribution of the same site are acquired at high speed, and a spatial causal relationship between the two is obtained. It is strongly desired to be able to grasp quantitatively in order to improve diagnostic accuracy.

[0005] By the way, an X-ray mammogram is used for pathological diagnosis of a mammary gland (measurement object) tumor for detecting breast cancer which is one of the main causes of death of women in their forties in recent years.
hy and the ultrasonic echo method are used.

A characteristic of a malignant tumor is that it has a complicated contour (surface shape) with irregularities as compared with benign. In diagnostic imaging of the mammary gland, these geometric features are X
It is often used in the diagnosis of breast tumors using line mammography and ultrasonic echo images.

[0007] X-ray mammography is a transmission image of X-rays through the mammary gland, and since a relatively high spatial resolution can be obtained, several methods for extracting and evaluating a tumor by image processing have been proposed.

[0008] On the other hand, in the diagnosis of breast tumors using ultrasonic echo images, there are some advantages as compared with Mammography.

That is, 1. 1. The ultrasonic image is obtained not as a transmission image but as an ultrasonic tomographic image; 2. The inside of the mammary gland can be easily observed in real time simply by bringing the ultrasonic probe into contact with the body surface (the surface of the mammary gland); And less pain on the subject.

[0010] Therefore, the examiner scans the inside of the mammary gland while observing the ultrasonic tomographic image on the screen, so that the tumor can be found with high accuracy and the malignant tumor can be distinguished. Can be expected and become indispensable for daily diagnosis.

In the conventional ultrasonic examination of the mammary gland, an examiner (for example, a doctor) scans the inside of the mammary gland with an ultrasonic probe and observes a tomographic image, thereby finding a tumor, malignant tumor ( Cancer). The characteristics of mammary tumors displayed on ultrasound tomography are characterized by their echo levels (ie,
Image brightness). This means that it is difficult to distinguish between the two using the difference in the luminance level of the image.

Therefore, irregularities in the geometric shape of the tumor contour line in the tomographic image are used. However, it is sometimes difficult to grasp the geometrical characteristics of the malignant tumor using only the tomographic image. In such a case, if the surface shape of the tumor is displayed three-dimensionally, the geometric surface shape can be easily observed, so that it is expected that more accurate diagnosis will be possible.

Conventionally, as an image acquisition apparatus of this type, there is one as shown in Japanese Patent Application No. 4-110305 (first prior art) (first prior art, FIG. 27).

That is, the image processing apparatus of the first prior art is suitable for performing processing of two-dimensional or three-dimensional image data obtained from an MRI or an ultrasonic diagnostic apparatus, and particularly, features information useful for diagnosis is provided. An image processing apparatus capable of high-speed and efficient extraction, comprising a fuzzy element array 6A,..., 6A in which one fuzzy element is associated with each pixel of a two-dimensional image space to be processed. In this method, the value of each pixel is determined by fuzzy inference. In the case of processing three-dimensional image data, a fuzzy element in pixel units is determined by the number of tomographic planes and the number of tomographic planes defined in advance. At least the number of pixels multiplied by the number of pixels is provided, and the fuzzy elements are operated in parallel to solve the problem.

That is, the image processing apparatus of the first prior art is provided with an image processing section comprising a large number of fuzzy element arrays 6A,..., 6A associated with respective pixels in a two-dimensional image space, each of which is a fuzzy element. Fuzzy inference is performed simultaneously and in parallel on the input two-dimensional image data based on rules and membership functions defined in advance, and each fuzzy element determines a value of a pixel associated with the fuzzy element. .

Also, a plurality of tomographic plane data are input and
When performing two-dimensional image processing, the image processing apparatus of the first related art includes an image processing unit having at least a predetermined number of tomographic planes and a number of fuzzy elements obtained by multiplying the number of pixels in each tomographic plane by an image processing unit. Each fuzzy element of the unit has a predetermined rule, independent rules and membership functions for all pixels, and executes a feature extraction process such as a boundary extraction at a given point in a three-dimensional space in parallel. ing. Here, the number of fuzzy elements of the image processing unit is determined by further multiplying the number of features to be simultaneously extracted.
Further, when each tomographic plane data is sequentially input and the input speed is slower than the processing speed of the fuzzy element,
A block 6A of a fuzzy element for processing one tomographic plane,
…, One or more blocks 6 in units of 6A
A,..., 6A constitute an image processing section, and one block 6A,.

When three-dimensional image data of a living body is created by using an ultrasonic diagnostic apparatus or medical MRI and a diagnostic process is performed, generally, it is formed based on echo data captured by transmitting and receiving ultrasonic waves to and from the living body. Is done. For example, when forming a two-dimensional tomographic image, the level of the echo data captured in the two-dimensional echo data capturing area is converted into a pixel value. Also, when forming a three-dimensional ultrasonic image,
Using the echo data captured in the three-dimensional echo data capturing area, first, contour extraction of a specific tissue is performed three-dimensionally, and thereafter, for example, the tissue surface is shaded,
Thereby, a three-dimensional tissue image is formed.

When calculating the cross-sectional area of a specific tissue in a two-dimensional ultrasonic image, or when forming a three-dimensional ultrasonic image or calculating the volume of a specific tissue, the contour of the tissue (the boundary between the tissues). Need to be extracted.

An image acquisition apparatus for extracting a tissue boundary with high accuracy based on a plurality of feature amounts of a tissue image is disclosed in, for example, Japanese Patent Application No. 5-333617 (second prior art, FIG. 28). There is something.

A second prior art ultrasonic image processing apparatus is an apparatus for processing echo data for forming an ultrasonic image, and includes a plurality of echo data included in a reference area centered on echo data of interest. An average calculation unit 12B for calculating the average μ, a variance calculation unit 14B for calculating the variance σ of the plurality of echo data included in the reference area, and a tissue difference emphasis calculation based on the average μ and the variance σ And outputs a new value of the echo data of interest, and a boundary extraction unit that extracts a boundary from the image after the tissue difference enhancement, and outputs a new value of the echo data of interest. A tissue difference enhancement process is performed on the ultrasound image by performing a tissue difference enhancement operation for each, and an average value μ and a variance value σ of the echo data are obtained in a reference region centered on the echo data of interest.
The tissue difference emphasis calculation was calculated based on the average value μ and the variance value σ. As a result, the boundary extraction accuracy is improved, and a tissue difference emphasis process capable of recognizing a tissue difference is performed by comprehensively considering the average value μ and the variance value σ of the image. In this configuration, the organizational difference emphasis calculation is performed by fuzzy inference, so that a complicated table for the organizational difference emphasis calculation is not required, and a large-scale table is not required.

The ultrasonic image processing apparatus according to the second prior art includes a separability calculating section 52B for calculating the tissue separation degree from the output of the fuzzy inference section 28B and a membership function for the fuzzy inference section 28B. Means for setting parameters, a parameter setting unit 54B for optimizing a membership function based on a plurality of tissue separations obtained by sequentially changing the parameters, and a boundary extraction for the image after tissue difference enhancement. And a boundary extraction unit for performing the optimization of the membership function by a feedback loop. Further, the optimization of the membership function provided in the fuzzy inference unit 28B is achieved by a feedback loop, the degree of tissue separation is calculated while sequentially changing the parameters for determining the membership function, and the optimal parameter is calculated based on the degree of tissue separation. By performing the optimization of the membership function before the actual image processing, the tissue difference enhancement processing (a type of differentiation processing) can be performed under the most appropriate conditions for various tissues. I was

[0022]

In order to three-dimensionally display and evaluate the surface shape of a tumor (particularly, breast cancer), it is important to extract a tumor region from the acquired ultrasonic voxel image data. . Moreover, the ultrasound image of the mammary gland
Artifacts such as speckle noise and acoustic shadows,
There are also difficulties peculiar to an ultrasonic image, such as a lack of a boundary portion and a low-luminance region other than a tumor such as a muscle layer.

However, in the first prior art and the second prior art for executing image processing methods such as binarization for performing binarization using a simple threshold value and a differential operator as preprocessing of an ultrasonic image of a mammary gland, Since conventional image processing methods such as image processing binarization and a differentiation operator are used for the ultrasonic image of the above, even if various fuzzy inferences are performed in the subsequent stage, the expected extraction accuracy is fundamentally realized. It is considered difficult.

Also, using such an image processing technique,
In order to realize a tumor with high accuracy and to judge a malignant tumor, there is a problem that a complicated calculation algorithm and a large-scale computer resource are required.

By the way, the biggest feature of breast tumor diagnosis is that
The difference between good and malignant is that it involves 1cm.
The following diagnosis of small cancers is the most important. Benign tumors are "regular and smooth", such as circular or elliptical in contour. On the other hand, a malignant tumor has an irregular and irregular shape such as a crab-shaped or a star-shaped contour.

However, the first prior art and the second
Conventional ultrasound techniques, including conventional techniques, often use such geometrical features in tomographic images, but if the tumor is still small, observing the tomographic image alone may indicate that the tumor is malignant. There are many cases (cases) where it is difficult to determine whether the disease is benign or not.

The only quantitative evaluation method (finding method) of a malignant tumor used in ordinary ultrasonic examination techniques including the first prior art and the second prior art is the aspect ratio (Depth) of the tumor area. width ratio, or D / W ratio, S / V ratio), that is, the ratio of the vertical diameter to the horizontal diameter (the ratio of the major axis to the minor axis) in the largest tomographic image of the tumor may be used.

That is, S / V of such a tumor area
When the ratio is used, the S / V ratio of cancer is larger than that of a benign tumor, but no fixed standard is found. Thus, in general, the S / V of the tumor area
Tumors with a ratio value of 0.8 or more are considered malignant, tumors with a value of 0.6 or less are benign, and tumors with a ratio between 0.6 and 0.8 are cases that need to be examined.

However, such a method for evaluating a malignant tumor is based on the premise that a maximum tomographic image of the tumor is accurately obtained. In addition, when the tumor is still small,
Even if it is a benign tumor, there is a problem that the S / V ratio of the tumor region tends to show a high value.

An object of the present invention is to solve such a conventional problem. In particular, the present invention provides a solid surface area (Su).
rface, S) and volume ratio (Volume, V) ratio parameters (S ³ / V ² ratio, S / V ratio)
Quantify the irregularities of the surface shape by defining parameters using, and quantitatively measure and evaluate the geometric irregularities of the three-dimensionally displayed three-dimensional surface based on this to identify the tumor How to support personal computer (P
C) A simple three-dimensional surface shape quantification method is provided, which enables a short-time (high-speed) calculation (image processing) with a small-scale computer resource having a calculation capability of the order, and as a result, an ultrasonic three-dimensional image It is an object of the present invention to realize a method for automatically identifying a malignant tumor, in which a region of a tumor (particularly, a mammary gland tumor) can be found with high accuracy and a judgment of a malignant tumor can be automatically extracted with good reproducibility.

Further, a parameter using a parameter S / V ratio of a ratio of a surface area S to a volume V of a tumor (benign or malignant) extracted as a three-dimensional image using a visualization technique such as an ultrasonic diagnostic method is defined. Quantification of irregularities in the surface shape of the tumor is quantified, and boundaries between tissues represented by a three-dimensional image composed of an MRI image, an ultrasonic image, and the like of a living body are extracted, and cancer tissues (particularly, breast cancer tissues) are extracted from normal tissues. (Mammary malignant tumor)), a method for automatically identifying a malignant tumor applicable to a pathological diagnosis support system for finding a personal computer (P
C) A simple three-dimensional surface shape quantification method is provided, which enables a short-time (high-speed) calculation (image processing) with a small-scale computer resource having a calculation capability of the order, and as a result, an ultrasonic three-dimensional image It is an object of the present invention to realize a method for automatically identifying a malignant tumor, in which a region of a tumor (particularly, a mammary gland tumor) can be found with high accuracy and a judgment of a malignant tumor can be automatically extracted with good reproducibility.

[0032]

According to a first aspect of the present invention, there is provided a method for quantifying a three-dimensional surface shape based on voxel data. An unevenness quantification step of calculating a shape determination parameter (γ) using the surface area (S) and volume (V) of the solid as an index for measuring the degree of unevenness of the three-dimensional surface shape; and the calculated shape determination parameter (Γ)
A shape determination step of determining that the surface of the three-dimensional object is smooth when the value is less than a predetermined threshold value, and determining that the surface of the three-dimensional object is exhibiting irregularities when the value is equal to or more than the threshold value. The shape determination parameter is a three-dimensional surface shape quantification method having a logical configuration defined by γ = (S ³ / V ² ) / κ, κ = constant.

According to the first aspect of the present invention, in order to quantify irregularities in the surface shape, a parameter using the ratio of the surface area to the volume of the extracted tumor is defined. For this,
By executing the unevenness quantification step, a shape determination parameter (γ = (S ³ / V ² ) / using the surface area and volume of the solid as an index for measuring the degree of unevenness of the solid surface shape is used.
κ) is calculated. Subsequently, by performing the shape determination step, it is determined that the three-dimensional surface is smooth when the shape determination parameter (γ) calculated in the unevenness quantification step is less than a predetermined threshold, while the shape determination parameter (γ) is greater than or equal to the threshold. In some cases, it is determined that the surface of the solid has irregularities.

The three-dimensional region obtained as a result can be displayed three-dimensionally, and the geometrical irregularities on the surface can be quantitatively measured and evaluated. The three-dimensional representation of a solid surface uses surface rendering that allows the shading of the solid surface to be similar to the shadows of light that occur when we look at an object. Thereby, it becomes possible to observe a subtle unevenness of the three-dimensional surface shape.

The calculation of the three-dimensional shape determination parameter γ can be automatically performed at high speed with a computer resource such as a compact and inexpensive PC, and is an effective means for diagnosing a three-dimensional shape.

According to a second aspect of the present invention, in the method of quantifying a three-dimensional surface shape according to the first aspect, the irregularity quantification step includes the step of determining the value of the shape determination parameter when the three-dimensional shape is a sphere. Is a three-dimensional surface shape quantification method having a logical configuration including a constant normalization step of normalizing the κ such that １ becomes 1.

In addition to displaying the three-dimensional surface shape of the three-dimensional object and evaluating its geometrical unevenness, the ratio S / V ratio between the cube of the surface area and the square of the volume of the extracted solid is used.
Using the parameter o, the irregularities of the three-dimensional surface shape are quantified. In addition, this S
/ V ratio (predetermined threshold) is used to perform grouping of the two.

Therefore, according to the second aspect of the present invention, in addition to the effect of the first aspect, in the unevenness quantification step, first, a constant normalization step is performed to make the three-dimensional shape spherical. After normalizing κ such that the value of the shape determination parameter becomes 1 at the time of, the shape determination parameter (γ = (S ³ /) using the surface area and volume of the solid as an index for measuring the degree of unevenness of the solid surface shape V ² ) / κ) is calculated. Subsequently, the above-described shape determination step is performed.

As described above, the three-dimensional display of the three-dimensional surface is performed by a shading method called surface rendering, and it is possible to easily observe a subtle uneven change in the three-dimensional shape of the surface.

According to a third aspect of the present invention, in the method for quantifying a three-dimensional surface shape according to the first aspect, the step of quantifying the irregularity of the three-dimensional surface is performed as a sum of voxels of the three-dimensional object finally extracted. This is a three-dimensional surface shape quantification method having a logical configuration including a volume calculation step of calculating the volume of the three-dimensional surface.

According to the invention described in claim 3, according to claim 1
In addition to the effects described in the above, the volume calculation step in the unevenness quantification step calculates the volume of the solid as the sum of the finally extracted voxels of the solid.

In such a volume calculation step, calculation is performed using voxels as basic units.
It can be automatically executed at high speed with computer resources such as the above, and is an effective means for initial diagnosis of breast solids and mass screening.

According to a fourth aspect of the present invention, in the three-dimensional surface shape quantification method according to the first aspect, the irregularity quantification step includes the step of quantifying the three-dimensional contour of a voxel adjacent to a voxel labeled as a three-dimensional contour. A surface area including a first step of forming a combination of three voxels, and a second step following the first step as a surface area of the three-dimensional object, the sum of the areas of triangles formed by the combination of the three adjacent voxels This is a three-dimensional surface shape quantification method having a logical configuration including a calculation step.

According to the fourth aspect of the present invention, the first aspect is provided.
In addition to the effects described in the above, in the uneven irregularity quantification step,
First, by performing a constant normalization step, κ is normalized so that the value of the shape determination parameter becomes 1 when the solid shape is a sphere, and then the first step is executed to label the solid as a solid contour. 3 for the voxel
One voxel combination is created, and the second step is executed following the first step, and the sum of the areas of the triangles formed by the combination of three adjacent voxels is defined as the surface area of the solid. Based on this, a shape determination parameter (γ = (S ³ / V ² ) / κ) using the surface area and volume of the solid as an index for measuring the degree of unevenness of the solid surface shape is calculated. Subsequently, the above-described shape determination step is performed.

The surface area calculation step including the first step and the second step can be automatically executed at high speed with computer resources such as a compact and inexpensive PC because the calculation is performed using voxels as basic units. This is an effective means for initial diagnosis of three-dimensional objects and mass screening.

The invention according to claim 5 is a method for automatically identifying a malignant tumor based on three-dimensional voxel data for determining the degree of irregularity of a tumor surface shape, wherein the index measures the degree of irregularity of a tumor surface shape. As the surface area of the tumor (S),
A tumor irregularity quantification step of calculating a tumor shape determination parameter (γ) using a volume (V); and if the calculated tumor shape determination parameter (γ) is less than a predetermined threshold, the surface of the tumor is smooth. Is determined, the shape determination step to determine that the surface of the tumor presents the properties of irregularities if the threshold value or more, the tumor determined that the surface is smooth is diagnosed as a benign tumor, A tumor diagnosis step of diagnosing, as a malignant tumor, a tumor determined that the surface has irregularities, wherein the tumor shape determination parameter is γ = (S ³ / V ² ) / κ, κ = constant This is a method for automatically identifying a malignant tumor having a logical configuration defined in (1).

The most distinctive feature of the diagnosis of breast tumors is that differentiation between benign and malignant is involved. Among them, the discrimination diagnosis of small cancers of 1 cm or less is the most important. Benign tumors are “regular and smooth” with a circular or elliptical outline.
It is.

On the other hand, the malignant tumor has an irregular and irregular shape such as a crab-shaped or star-shaped contour. In normal ultrasonography, such geometrical features in tomographic images are often used, but if the tumor is still small, observing the tomographic image alone will determine whether the tumor is malignant or benign. Sometimes it is difficult to know what is.

Therefore, the invention described in claim 5 utilizes the fact that the surface shape of a malignant tumor is more irregular than that of a benign tumor, so that the surface area is larger than that of a benign tumor having the same volume. Therefore, in order to quantify the irregularity of the surface shape, a parameter using the ratio of the surface area to the volume of the extracted tumor is defined. For this purpose, by performing the tumor irregularity quantification step, a tumor shape determination parameter (γ) using the surface area and volume of the tumor as an index for measuring the degree of irregularity of the tumor surface shape is calculated. Subsequent to the tumor irregularity quantification step, a shape determination step is performed by paying attention to the fact that the parameter S / V ratio of the surface area to volume ratio normalized in the case of a sphere shows a higher value in a malignant tumor than in a benign tumor. By doing so, it is determined that the surface of the tumor is smooth when the calculated tumor shape determination parameter is less than a predetermined threshold, while it is determined that the surface of the tumor has irregularities when the calculated parameter is not less than the threshold. I do. By performing a tumor diagnosis step following the shape determination step, a tumor determined to have a smooth surface in the shape determination step is diagnosed as a benign tumor, and a tumor determined to have a surface having irregularities is identified as a malignant tumor. Diagnose a tumor.

As a result, a region of a mammary gland tumor having irregularities on the tumor surface, which is a feature of a malignant tumor (cancer), can be detected well, and a benign or malignant geometric feature can be easily detected from a three-dimensional display image of the tumor surface. This makes it possible to more objectively make a diagnosis of benign or malignant tumor.

That is, it is possible to three-dimensionally display a tumor region obtained by an automatic breast tumor extraction system (pathological diagnosis support system) and quantitatively measure and evaluate the geometric unevenness of the surface. become able to. The three-dimensional representation of the tumor surface uses surface rendering that allows the shading of the tumor surface to be similar to the light shading that occurs when we look at an object. As a result, it becomes possible to observe subtle changes in the unevenness of the tumor surface shape, and it becomes possible for the examiner and the patient to understand and judge based on common recognition. As a result, breast cancer diagnosis can be supported by making the extracted tumor into a three-dimensional image and quantifying the geometric unevenness of the tumor surface shape.

The calculation of the tumor shape determination parameter γ can be automatically executed at high speed with a computer resource such as a compact and inexpensive PC, and is an effective means for initial diagnosis and mass screening of breast tumors.

According to a sixth aspect of the present invention, in the method for automatically identifying a malignant tumor according to the fifth aspect, the step of quantifying the irregularity of the tumor unevenness includes the step of determining the tumor shape determining parameter when the shape of the tumor is a sphere. This is an automatic malignant tumor identification method having a logical configuration including a constant normalization step of normalizing the κ so that the value becomes 1.

To assist in the diagnosis of breast tumors, the extracted tumors are displayed three-dimensionally with their surface shapes,
Evaluate the geometric unevenness. The irregularity of the surface shape of the tumor is quantified using a parameter of the ratio S / V ratio of the cube of the surface area of the extracted tumor to the square of the volume of the tumor. In addition, the S / Vr for both benign and malignant
Atio (predetermined threshold) is used to perform grouping of the two.

Therefore, according to the invention of claim 6, in addition to the effect of claim 5, by performing a tumor irregularity irregularity quantification step, the tumor surface shape can be used as an index for measuring the degree of irregularity. A tumor shape determination parameter (γ = (S ³ / V ² ) / κ) using the surface area and volume of the target is calculated.
If the surface of the tumor is smooth, γ will be small, and if the surface has irregularities, γ will have a large value. Therefore, by performing the shape determination step following the tumor unevenness quantification step, it is determined that the tumor surface is smooth when the calculated tumor shape determination parameter is smaller than a predetermined threshold (S / V ratio). On the other hand, the threshold (S / Vra
tio) When it is equal to or more than the above, it is determined that the surface of the tumor exhibits unevenness. By performing a tumor diagnosis step following the shape determination step, a tumor determined to have a smooth surface in the shape determination step is diagnosed as a benign tumor, and a tumor determined to have a surface having irregularities is identified as a malignant tumor. Diagnose a tumor.

As described above, the three-dimensional display of the tumor surface is performed by the shading method called surface rendering, and it is possible to easily observe a subtle unevenness of the three-dimensional shape of the surface.

According to a seventh aspect of the present invention, in the method for automatically identifying a malignant tumor according to the fifth aspect, the step of quantifying the irregularity of the tumor irregularities is performed by calculating a sum of voxels constituting the finally extracted tumor. An automatic malignant tumor identification method having a logical configuration including a tumor volume calculation step of calculating the volume of the tumor.

According to the invention of claim 7, according to claim 5,
In addition to the effects described in (1), by providing a tumor volume calculation step in the tumor irregularity quantification step, the tumor volume is calculated as the sum of voxels constituting the finally extracted tumor. According to the volume of the tumor, a tumor shape determination parameter (γ = (S ³ / V ² ) / κ) using the surface area and volume of the tumor is calculated as an index for measuring the degree of unevenness of the tumor surface shape. Following the tumor irregularity quantification step, the above-mentioned shape determination step is executed, and the smoothness of the tumor surface and the nature of the irregularity are determined according to the calculated tumor shape determination parameters. A tumor diagnosis step is executed following the shape determination step, and a benign / malignant diagnosis of the tumor is made in the shape determination step according to the surface determination result.

In the tumor volume calculation step, since the calculation is performed using voxels as basic units, it can be automatically executed at high speed with computer resources such as a compact and inexpensive PC, and is effective as an initial diagnosis of breast tumor and a mass screening. Means.

According to an eighth aspect of the present invention, in the method for automatically identifying a malignant tumor according to the fifth aspect, the step of quantifying the irregularity of the tumor unevenness includes three steps adjacent to the voxel labeled as the outline of the tumor. A surface area comprising: a first step of forming a combination of three voxels; and, following the first step, a second step in which the sum of the areas of triangles formed by the combination of the three adjacent voxels is the surface area of the tumor This is an automatic malignant tumor identification method having a logical configuration including a tumor surface area calculation step of calculating.

According to the invention described in claim 8, according to claim 5,
In addition to the effects described in (1), by providing a tumor volume calculation step in the tumor irregularity quantification step, the tumor volume is calculated as the sum of voxels constituting the finally extracted tumor. In addition, a tumor surface area calculation step is provided in the tumor irregularity quantification step, and the first step is executed to create a combination of three adjacent voxels with respect to the voxel labeled as the outline of the tumor. Further, by executing the second step following the first step, the total sum of the areas of the triangles formed by the combination of three adjacent voxels is calculated as the surface area of the tumor. In addition,
According to the volume of the tumor, a tumor shape determination parameter using the surface area and volume of the tumor is calculated as an index for measuring the degree of unevenness of the tumor surface shape. Following the tumor irregularity quantification step, the above-mentioned shape determination step is executed, and the smoothness of the tumor surface and the nature of the irregularity are determined according to the calculated tumor shape determination parameters. A tumor diagnosis step is executed following the shape determination step, and a benign / malignant diagnosis of the tumor is made in the shape determination step according to the surface determination result.

In the tumor surface area calculation step including the first step and the second step, since the calculation is performed using voxels as basic units, it can be automatically executed at high speed with a computer resource such as a compact and inexpensive PC. This is an effective means for early diagnosis and mass screening.

[0063]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS First, based on the drawings, a three-dimensional image acquisition process which is a pre-process of a breast cancer screening system 10 for executing a tumor extraction process of the present invention, and a device for executing the process (a personal computer (PC)) An embodiment will be described. FIG. 1 is a block diagram showing a basic configuration of a PC.

[Three-dimensional image acquisition processing] First, an ultrasonic echo (= probe) acquired using the probe 12 (ultrasonic probe 121) to which the three-dimensional position sensor 13 (AC magnetic field position sensor 131) is attached. A mammary gland 301 (measured object 30) is extracted from the data 12a), and a three-dimensional image in the breast cancer screening system 10 using an ultrasonic image for performing breast cancer diagnosis for identifying benign or malignant of a tumor from its three-dimensional surface shape. PC and 3 which are most suitable for the system which generates an ultrasonic image
The dimensional image acquisition processing will be described.

The breast cancer screening system 10 performs a discriminating diagnosis of breast cancer by evaluating geometrical irregularities in the shape of a three-dimensional region (three-dimensional image) of a breast tumor received by a PC. In the three-dimensional processing required for the breast cancer screening system 10 that extracts such a breast tumor, voxel data 16a regarding the tumor site is indispensable.

The PC is a breast cancer screening system 10 described later.
, A program code describing a three-dimensional image acquisition process which is a pre-process of the above, a program code describing a tumor extraction process (an automatic algorithm for extracting a breast tumor) which is a post-process of the breast cancer screening system 10, and a post-process of the breast cancer screening system 10 Is a hardware for physically executing a program code describing the malignant tumor automatic identification process (automatic identification algorithm for breast tumor).

The breast cancer screening system 10 centered on such a PC has a three-dimensional area (three-dimensional image) of a breast tumor.
Is automatically extracted by a program (including a program code of a three-dimensional image acquisition process described later), and the tumor surface is three-dimensionally displayed (displayed as a three-dimensional image) to support the discrimination diagnosis of breast cancer (the breast cancer screening system 10). It has a three-dimensional ultrasonic image data collection function.

The breast cancer screening system 10 shown in FIG. 1 comprises an ultrasonic probe 121, an AC magnetic field position sensor 131, a PC
Hardware configuration.

The ultrasonic probe 121 is connected to the object 3 to be measured.
While scanning the surface of the object 0, the object to be measured 30 is probed to generate probe data 12a relating to the internal structure and the surface structure.

In this embodiment, while scanning the surface of the mammary gland 301 using ultrasonic waves, ultrasonic tomographic image data relating to the internal structure and the surface structure acquired by performing a probe based on the ultrasonic echo from the mammary gland 301. The ultrasonic probe 121 for generating the ultrasonic probe 121 a
Used as

The AC magnetic field position sensor 131 measures the spatial position and posture of the ultrasonic probe 121 during scanning while being attached to the ultrasonic probe 121, and obtains coordinate data of the ultrasonic probe 121. It is generated in synchronization with the probe operation of the measured object 30.

The AC magnetic field position sensor 131 has three-dimensional coordinates (x ₀ , y ₀ , z ₀ ) relating to its own spatial position and attitude.
And position data (x ₀ , y ₀ , z ₀ , ψ, θ,...) Representing the inclination direction of the vehicle itself as angles (azimuth angle ψ, elevation angle θ, rollover angle Ф).
位置) is a position sensor that measures in real time and outputs it as probe coordinate data 13a.

The AC magnetic field position sensor 131 measures the spatial position and posture of the ultrasonic probe 121 while scanning the surface of the mammary gland 301 while being attached to the ultrasonic probe 121, and The coordinate data 121 (x ₀ , y ₀ , z ₀ ) is generated in synchronization with the probe operation of the mammary gland 301.

Thus, the breast cancer screening system 10 (P
C), when acquiring the ultrasonic tomographic image data 121a, collects the probe coordinate data 13a in synchronization with the acquisition of the ultrasonic tomographic image data 121a, and acquires the ultrasonic tomographic image data 121a and the ultrasonic tomographic image data 121a. The three-dimensional ultrasonic image data 15b is generated using the probe coordinate data 13a synchronized with the
b is converted to isotropic voxel data 16a by performing linear image interpolation. If two or more different values correspond to the same voxel, the average value of these values is used as voxel voxel data 16a. be able to.

That is, instead of the conventional manual scanning or mechanical scanning operation that requires tracking the position and orientation of the ultrasonic probe 121 in real time, the ultrasonic tomographic image (ultrasonic tomographic image data 121a) and the ultrasonic Position data of the acoustic probe 121 (probe coordinate data 13a = position data (x ₀ , y ₀ , z ₀ , ψ, θ, Ф))
Is used as the correction data for the position of the ultrasonic probe 121, the ultrasonic tomographic image (ultrasonic tomographic image data 121a) acquired by the AC magnetic field position sensor 131 using the AC magnetic field attached to the ultrasonic probe 121.
Is converted to three-dimensional voxel data 16a with little artifact and high reproducibility as in the case of the conventional manual scanning.
Can be converted to As a result, when mechanically scanning the mammary gland 301 with the ultrasonic probe 121, the breast gland 301 is a very flexible and delicate tissue that may compress the tissue and fail to obtain a highly reproducible tumor shape. The ultrasonic tomographic image data 121a with high accuracy can be generated for the measured object 30 as well.

When two or more different values correspond to the same voxel, the average value of these values is used as the voxel voxel data 16a, so that the acquired ultrasonic tomographic image (probe data 12a) can be converted into three. After the dimensional coordinate transformation, linearly interpolated isotropic voxel image data (voxel data 16a) can be generated.

When collecting the probe data 12a, the breast cancer screening system 10 (PC) collects the probe coordinate data 13a in synchronization with the collection of the probe data 12a.

The PC generates three-dimensional image data 15a using the probe data 12a and the probe coordinate data 13a synchronized with the probe data 12a.

The PC performs linear image interpolation on the three-dimensional image data 15a to obtain isotropic voxel data 16a.
When two or more different values correspond to the same voxel, the average value of these values is used as the voxel data 16a of the voxel.

The PC sends the probe magnetic field position sensor 13a as probe coordinate data 13a synchronized with the probe data 12a.
1, a transformation matrix T [a _ij ], (i, j) based on the azimuth angle ψ, the elevation angle θ, and the rollover angle 相 対,
= 1, 2, 3) and the coordinates (x,
y, 0) multiplied by a transformation matrix T [a _ij ] to obtain 3
The coordinates (x ′, y ′, z ′) of the dimensional image data 15a are generated.

That is, an alternating current magnetic field position sensor 131 is attached to the ultrasonic probe 121, and the alternating current magnetic field position sensor 131 has its own (that is, the alternating magnetic field position sensor 131) spatial position data (position and position data). The data 13a of the three-dimensional coordinates related to the posture = position data (x ₀ , y ₀ , z ₀ , ψ, θ, Ф) is measured in real time and output as probe coordinate data 13 a. Therefore P
C is the probe data 12a (the ultrasonic tomographic image data 121
Probe coordinate data 13a can be collected in synchronization with the collection of a). In response to this, the PC transmits the probe data 12a (ultrasonic tomographic image data 121a) and the probe data 12a.
In order to generate three-dimensional image data 15a using probe coordinate data 13a synchronized with (ultrasonic tomographic image data 121a), coordinates (x, y, 0) of probe data 12a (ultrasonic tomographic image data 121a) are used. _Is multiplied by a transformation matrix T [a _ij ] to generate coordinates (x ′, y ′, z ′) of the three-dimensional image data 15a. In other words, instead of the conventional manual scanning or mechanical scanning operation that requires tracking the position (x ₀ , y ₀ , z ₀ ) and posture (ψ, θ, ψ) of the ultrasonic probe 121 in real time,
Probe coordinates data 1 of the ultrasonic probe 121 at the same time as the coordinates (x, y, 0) of the ultrasonic tomographic image data 121a
The transformation matrix T [a _ij ] as 3a is converted to the ultrasonic probe 1
The coordinates (x, y, 0) of the ultrasonic tomographic image data 121 a acquired by the AC magnetic field position sensor 131 based on the AC magnetic field attached to the ultrasonic probe 121 are used as the correction data of the position 21 in the conventional method. Similar to the manual scanning, the coordinates (x ', y', 3 ') of the three-dimensional voxel data 16a with less artifact and high reproducibility.
z ′).

As a result, when the mammary gland 301 is mechanically scanned by the ultrasonic probe 121, the mammary gland 301 is a very flexible and delicate tissue that may compress the tissue and fail to obtain a highly reproducible tumor shape. High-precision ultrasonic tomographic image data 121 for the object 30 to be measured
The coordinates (x, y, 0) of a can be generated.

When two or more different values correspond to the same voxel, the average value of these values is used to determine the coordinates (x ′, y ′, x ′) of the voxel voxel data 16a.
z ′), the obtained ultrasonic tomographic image (probe data 12a) is subjected to three-dimensional coordinate conversion, and then linearly interpolated isotropic voxel image data coordinates (x ′,
y ′, z ′) can be generated.

Here, each matrix element a _ij in the above-mentioned transformation matrix T [a _ij ] is represented by a ₁₁ = cos (ψ) · cos (θ) and a ₁₂ = cos (ψ) · sin (θ) · sin (Ф) −sin (ψ) · cos (Ф), a ₁₃ = cos (ψ) · sin (θ) · cos (Ф) + sin (ψ) · sin (Ф), a ₂₁ = sin (ψ) · cos (θ), a ₂₂ = cos (ψ) · cos (Ф) + sin (ψ) · sin (θ) · sin (Ф), a ₂₃ = sin (ψ) · sin (θ) · cos (Ф) − cos (ψ) · sin (Ф), a ₃₁ = −sin (θ), a ₃₂ = cos (θ) · sin (Ф), a ₃₃ = cos (θ) · cos (Ф) ... Equation (1-1) ) To (1-9).

That is, instead of the conventional manual scanning or mechanical scanning operation that needs to track the position and orientation of the ultrasonic probe 121 in real time, a conversion matrix T [a _ij represented by a simple linear function is used. ] As the correction data of the position of the ultrasonic probe 121, the coordinates (x, y, and x) of the ultrasonic tomographic image data 121a acquired by the AC magnetic field position sensor 131 based on the AC magnetic field attached to the ultrasonic probe 121. 0) can be realized by using small-scale hardware such as a personal computer (PC) having not so high computational power, as in the case of the conventional manual scanning, with less artifacts and higher reproducibility of a three-dimensional voxel. Data 16
The coordinates can be converted to the coordinates (x ′, y ′, z ′) of a. As a result, when mechanically scanning the mammary gland 301 with the ultrasonic probe 121, the breast gland 301 is a very flexible and delicate tissue that may compress the tissue and fail to obtain a highly reproducible tumor shape. The coordinates (x, y, 0) of the high-accuracy ultrasonic tomographic image data 121a can be generated at high speed and at low cost even with small-scale hardware for the measured object 30.

More specifically, the breast cancer screening system 1 shown in FIG.
The basic hardware configuration of “0” will be described.

In this embodiment, the ultrasonic probe 121
After the A / D conversion of the video signal from the ultrasonic diagnostic apparatus 20 while tracking the ultrasonic probe 121 by attaching the AC magnetic field position sensor 131 to the
us (PeripheralComponent In
terconnect Bus), and at the same time, the three-dimensional position (x ₀ ,
y ₀ , z ₀ ) and orientation data (azimuth angle ψ, elevation angle θ, rollover angle Ф) are output via RS232C, and the memory (Frame Grabber) or PC of the breast cancer screening system 10 is output.
We built a system to transfer data directly to the memory.

Here, a small size (in this embodiment, external dimensions = approximately 2.8 × 2.2 × 1.5 cm ³ ) and light weight (17.0
By attaching the Kg) AC magnetic field position sensor 131 to the ultrasonic probe 121, a skilled examiner can scan the breast surface by manual scanning in the same manner as in daily diagnosis. In addition, an ultrasonic image (tip data 12
a = ultrasonic tomographic image data 121a) is transferred to a memory (not shown) that can be directly accessed by a PC via a PCI bus.
21a) can be accelerated.

Hereinafter, a specific configuration of the breast cancer screening system 10 will be described.

As the ultrasonic diagnostic apparatus 20, for example, SSD-2000 (trade name) manufactured by Aloka Co., Ltd.
Can be diverted. The ultrasonic probe 121 is
A concave electronic scanning method with a frequency of 7.5 MHz is used.
The reason for using the concave shape is to acquire an image without changing the shape of the mammary gland as much as possible. This ultrasonic probe 121
, An ultrasonic image is usually acquired at a setting of depth range = 5 cm, and the size of each pixel is 0.
It becomes 014 × 0.014 cm ² . Further, an AC magnetic field position sensor 131 is attached to the tip of the probe, so that the position and direction of the ultrasonic probe 121 can be tracked in real time.

The ultrasonic probe 121 has an AC magnetic field position sensor 131 (for example, manufactured by Polhemus).
Fasttrak tracking system (trade name)) is attached, and the position and orientation (azimuth angle ψ, elevation angle θ, rollover angle Ф (unit is [degree (or rad))) of the ultrasonic probe 121 are detected and tracked. The user can arbitrarily acquire an ultrasonic tomographic image (probe data 12a) of the affected part.

A probe tracking system (Po)
Fastraktracking s manufactured by lhemus
system (product name)). This system is composed of an AC magnetic field position sensor 131, a transmitter, and a system electronic control unit. The AC magnetic field position sensor 131 is attached to the ultrasonic probe 121, and the AC generated from the transmitter arranged immediately below the bed. The magnetic field is detected, and the position and orientation data of the AC magnetic field position sensor 131 with respect to the transmitter arrangement are calculated by the system electronic control unit. The obtained position data is RS232C
And transferred to the computer in real time.

That is, while the AC magnetic field position sensor 131 measures the position (x ₀ , y ₀ , z ₀ ) and angle (azimuth angle ψ, elevation angle θ, rollover angle で) of the ultrasonic probe 121, ultrasonic diagnosis is performed. Ultrasonic tomographic image (probe data) data observed by the device 20 is directly loaded into a memory (not shown) of a PC (Pentium (trade name) 166 MHz, RAM 192 MB, manufactured by DEC).

In the present embodiment, the image capture is performed by a PC
I Bus Frame Grabber (Data
The processing is performed at a transfer rate of 15 frames / sec using DT3155 (trade name, manufactured by Translation Corporation). The PCI Frame Grabber is mounted on a PCI bus of a PC as an image board, and the image board can take in an external video signal into a memory (DRAM) of the PC in real time via the PCI bus after A / D conversion. The amount of images that can be taken into the PC depends on the capacity of the memory of the PC, and the larger the memory, the more ultrasonic tomographic image data 121a can be taken into the PC.

As the PC, specifically, DEC P
C (CPU = Pentium166MHz, DRAM1
92MB, OS (Operating System)
= Windows 95).

Thus, in the breast cancer screening system 10, the examiner can arbitrarily track the position (x ₀ , y ₀ , z ₀ ) and posture (azimuth angle ψ, elevation angle θ, rollover angle の) of the ultrasonic probe 121. A tomographic image of the affected area can be acquired.

The above-mentioned Fasttrack tracking
The system is used to measure the position and orientation of an object at a remote place using an alternating magnetic field. Fa
The track tracking system first generates a magnetic field from three stationary coils installed in a transmitter (Transmitter), receives these magnetic fields by a sensor including three receiving coils for remote sensing, and uses a predetermined calculation algorithm. The three-dimensional position of the sensor relative to the transmitter (x ₀ ,
y ₀ , z ₀ ) and posture (azimuth angle ψ, elevation angle θ, rollover angle Ф) are calculated.

Therefore, the ultrasonic probe 121
The digitized AC magnetic field position sensor 131 is an ultrasonic probe 1
3D coordinates of 21 current positions (scanning positions)
(X ₀, Y₀, Z₀) And its orientation to an angle (azimuth ψ, elevation
Angle data (x, y, z,
ψ, θ, φ) are measured in real time and RS232C is used.
Output to a PC.

Here, the measured coordinates are determined by the direction in which the transmitter (not shown) of the AC magnetic field is arranged. In the reference coordinate system of the transmitter in the ultrasonic probe 121, the vertical direction is the Z direction, and the direction from the affected part toward the examiner is the X direction.

It is desirable that the transmitter of the ultrasonic probe 121 is disposed almost immediately below the affected part. See FIG. 2 for the three-dimensional spatial coordinate system of the Fasttrack tracking system.

While measuring the position and the angle of the ultrasonic probe 121 with the AC magnetic field position sensor 131, the tomographic image data observed by the ultrasonic diagnostic apparatus 20 is converted to PC (D
(Pentium 166MHz, 192MB manufactured by EC)
Directly into memory. Image capture is PCI B
us Frame Grabber (Data Tra
DT3155 manufactured by nslation Co., Ltd.) and the actual transfer rate 15 frames / sec together with the position (x ₀ , y ₀ , z ₀ ) and orientation data (azimuth angle ψ, elevation angle θ, rollover angle Ф) of the ultrasonic probe 121. Was realized.

Next, the process of creating the voxel data 16a will be described.

FIG. 3 is a flowchart for explaining one embodiment of the three-dimensional image acquisition processing executed by the breast cancer screening system of FIG.

The three-dimensional image acquiring process (the process of producing the voxel data 16a) of the present invention shown in FIG. 3 includes a probe step (step S1) and a three-dimensional position detecting step (step S1).
2) It comprises a three-dimensional coordinate conversion step (step S3) and a three-dimensional voxel data generation step (step S4).

In the present PC, since the extraction of the mammary gland tumor region is performed by three-dimensional processing, image data of the voxel data 16a which is equivalent to the tumor site is required. Therefore, the acquired tomographic image group is first subjected to three-dimensional coordinate conversion, and then converted to voxel data 16a by linear interpolation.

Now, the position (X) of the AC magnetic field position sensor 131 facing the transmitter (ultrasonic probe 121).
Axis direction, Y-axis direction, and Z-axis direction), the direction (azimuth (Azimuth)) of the AC magnetic field position sensor 131 using the AC magnetic field relative to the transmitter (the probe).
ψ, elevation angle (Elevation) θ, rollover angle (Rol)
l) Ф) (Step S1, Step S2). The acquired ultrasonic tomographic image (the ultrasonic tomographic image data 121a
The coordinates (x, y, 0) of (probe data 12a) are represented by matrix P, the coordinates (x ', y', z ') after three-dimensional conversion are represented by matrix Q, the current coordinate position of the ultrasonic probe 121, and The coordinates (x ₀ , y ₀ , z) of the probe coordinate data 13a indicating the direction
_When (x ₀ , y ₀ , z ₀ ) in ( ₀ , ψ, θ, Ф) is defined as a matrix R, a matrix operation (matrix product operation) using a transformation matrix T [a _ij ] is performed to obtain the matrix R. (X, y, 0) from the coordinates (x, y, 0) of the obtained ultrasonic tomographic image (ultrasonic tomographic image data 121a (probe data 12a)).
y ′, z ′) are obtained (step S3).

That is, it is given by Q ^t = TP ^t + R ^t . Here, each of the matrices Q ^t , P ^t , and R ^t is
It means a transposed matrix for each of the matrices Q, P, and R.

Here, the transformation matrix T [a _ij ] is expressed as follows: a ₁₁ = cos (ψ) · cos (θ), a ₁₂ = cos (ψ) · sin (θ) · sin (Ф) −sin (ψ) · Cos (Ф), a ₁₃ = cos (ψ) · sin (θ) · cos (Ф) + sin (ψ) · sin (Ф), a ₂₁ = sin (ψ) · cos (θ), a ₂₂ = cos (Ψ) · cos (Ф) + sin (ψ) · sin (θ) · sin (Ф), a ₂₃ = sin (ψ) · sin (θ) · cos (Ф) −cos (ψ) · sin (Ф) , A ₃₁ = −sin (θ), a ₃₂ = cos (θ) · sin (Ф), a ₃₃ = cos (θ) · cos (Ф) ... Equations (1-10) to (1-18) Given. With respect to the ultrasonic image acquired in this manner, the position (x ₀ , y ₀ , z
₀ ) and the direction data (azimuth angle ψ, elevation angle θ, rollover angle Ф) to obtain voxel data 16a (step S4).

By the way, depending on the scanning method of the ultrasonic probe 121, the acquired voxel data 16a has voxel data 1 for which the luminance value is not given.
6a exists or has the same voxel data 16a
May correspond to two or more different luminance values.

Therefore, in the present embodiment, in particular, when there is voxel data 16a to which no luminance value is given, luminance interpolation (Image Brightness) of the image is performed.
s Interpolation) is used to determine the luminance value of the voxel data 16a.

When two or more different luminance values correspond to the same voxel data 16a, the average luminance value of these luminance values is used as a representative luminance value of the voxel data 16a.

Next, an embodiment of image interpolation will be described.

In the present embodiment, the linear interpolation method (one form of the linear interpolation method LI) is extended to three-dimensional spatial processing, and voxel
The data 16a is interpolated.

That is, in this PC, image data of isotropic voxel data 16a is required.
The acquired ultrasonic tomographic image (probe data) is converted into voxel data 16a by linear interpolation after three-dimensional coordinate conversion.

Specifically, first, the voxel to be interpolated
With respect to the data 16a, linear interpolation processing is performed by obtaining six voxel data 16a having luminance values that are the nearest neighbors in three directions of the X-axis direction, the Y-axis direction, and the Z-axis direction. This linear interpolation processing has advantages that the processing speed is high and the gray scale of the tumor image can be kept smooth.

As described above, the breast cancer screening system 1
According to the three-dimensional image acquisition processing which is the pre-processing of 0, since the three-dimensional position sensor is attached to the ultrasonic probe 121, the ultrasonic probe 121 during scanning is
Is measured, the probe coordinate data 13a synchronized with the probe operation of the measured object 30 can be generated. Further, while scanning the surface of the measured object 30 using the ultrasonic probe 121, the measured object 30
When the probe is generated to generate the probe data 12a, the PC
The probe coordinate data 13a is collected in synchronization with the collection of the probe data 12a. Subsequently, the PC reads the probe data 12
The three-dimensional image data 15a is generated using the probe coordinate data 13a synchronized with the probe data 12a. Subsequently, the PC performs linear image interpolation on the three-dimensional image data 15a to convert it into isotropic voxel data 16a, while averaging the two if two or more different values correspond to the same voxel. The value is used as voxel data 16a of the voxel. In other words, a series of scans obtained by arbitrarily scanning the ultrasonic probe 121 using position data (coordinate data of the ultrasonic probe 121) measured by the AC magnetic field position sensor 131 attached to the ultrasonic probe 121. Can be converted into the voxel data 16a.

[0117] Such a PC can be used to measure an object 3 to be measured such as a mammary gland 301 having a very flexible and delicate tissue structure.
It is suitable for an application such as creating the voxel data 16a by probing 0. That is, a conventional skilled examiner manually scans the surface of the mammary gland 301 and finely adjusts the position and orientation of the ultrasonic probe 121 to collect the probe data 12a. Instead of the mechanical scanning (scanning) operation of mechanically scanning the ultrasonic probe 121,
The ultrasonic probe 12 is being scanned by an AC magnetic field position sensor 131 attached to the ultrasonic probe 121.
1 to generate probe coordinate data 13a synchronized with the probe operation of the mammary gland 301 (measured object 30), and the PC collects (scans) the probe data 12a. Synchronized probe coordinate data 13a
Is collected, the position data of the ultrasonic probe 121 (the ultrasonic probe 121
Coordinate data) and the acquired probe data 12a
Can be converted to three-dimensional voxel data 16a with less artifacts and high reproducibility as in conventional manual scanning and mechanical scanning by using the data as position data.

[Tumor Extraction Process] Next, an embodiment of a tumor extraction process (automatic breast tumor extraction algorithm) which is a post-process of the breast cancer screening system 10 will be described with reference to the drawings.

In the following description, 3 using fuzzy inference is described.
An algorithm for automatically extracting a tumor (particularly, a breast tumor) from the dimensional voxel data 16a will be described. First, a process of automatically creating a membership function using a three-dimensional LoG (Laplaceof Gaussian) filter in order to stably and automatically extract a good breast tumor region will be described. Next, using fuzzy inference and relaxation method,
Classification of each voxel into three classes, "tumor", "normal tissue" and "boundary" between them, and determination of a three-dimensional region of the tumor will be described.

For the diagnosis of breast cancer, image diagnosis such as X-ray mammography and ultrasonic examination is often used in addition to visual and palpation and fine needle aspiration cytology. A feature of malignant tumors (cancer) is that they have a complex contour with irregularities as compared to benign. Such geometric features are often used in diagnostic imaging. X-ray Mammogram
hy is an X-ray transmission image to the mammary gland 301, and a relatively high resolution can be obtained. Therefore, several methods for extracting and evaluating a tumor by image processing have been proposed.

On the other hand, the diagnosis of breast tumor by ultrasonic examination is that the inside of the breast gland 301 can be easily observed in real time simply by bringing the ultrasonic probe 121 into contact with the surface of the breast, and that there is little pain to the subject. , And that there is no need to worry about exposure and that it is safe
In such cases, tumors can be extracted, and so are essential for daily diagnosis. In the ultrasonic examination of the mammary gland 301, an examiner scans the inside of the mammary gland 301 and observes a tomographic image to find a tumor and distinguish a malignant tumor (cancer). The feature of the breast tumor displayed on the ultrasound tomographic image is that the echo level, that is, the brightness of the image is lower in both benign and malignant tissues than in normal tissue. This means that it is difficult to distinguish between the two using the difference in the luminance level of the image. Therefore, the irregular shape of the tumor contour in the tomographic image is used, but it may be difficult to grasp the geometrical characteristics of the malignant tumor (cancer) only with the tomographic image. In such a case, if the surface shape of the tumor is displayed three-dimensionally, the geometric surface shape can be easily observed, so that it is expected that more accurate diagnosis will be possible.

In order to three-dimensionally display and evaluate the surface shape of a tumor, it is important to extract a tumor region from a plurality of acquired ultrasonic tomographic image groups. However,
In the ultrasound image of the mammary gland 301, there are artifacts such as speckle noise and acoustic shading, lack of boundaries, and low-luminance areas other than tumors such as muscle layers, so that conventional images such as binarization and differentiation operators are used. It cannot be realized by the processing method.

Under such technical background, all the tumor extraction processing in the tumor extraction processing of the present embodiment is performed in a three-dimensional space, and the luminance of the image obtained from the reference voxel area near the voxel of interest. Since the statistic such as the average value is more reliable than that from the two-dimensional reference area, better extraction of the tumor can be expected.

In the tumor extraction processing according to the present embodiment, which is executed following the three-dimensional image acquisition processing which is the preprocessing of the breast cancer screening system, a concave electronic scan having a frequency of 7.5 MHz is performed in order to confirm the effectiveness of the breast cancer screening system. Breast tumors were extracted using an ultrasonic diagnostic apparatus of the system. Malignant tumor (cancer)
The results applied to the cases of and benign tumors (fibroadenoma) were consistent with the contours traced by the doctor, suggesting the effectiveness of the tumor extraction processing of the present embodiment. Further, when the extracted tumor was three-dimensionally displayed by surface rendering, a difference in surface shape could be observed.

Since the area of the mammary gland tumor in the ultrasonic image has a lower luminance than that of the normal tissue, the tumor is extracted based on the level of the luminance value. However, since an ultrasonic image has remarkable artifacts such as acoustic shadows and relatively low-luminance regions such as muscle layers, it is difficult to extract a tumor by simple binarization.

In the tumor extraction processing of this embodiment, the “tumor” and “normal tissue” are first applied to the three-dimensional voxel data 16a, ie, all voxel data, using the concept of fuzzy reasoning. , And their "boundaries", to determine the "likeness" of the three classes, and from there, gradually determine the area of the tumor, while gradually resolving the contradictions in the three-dimensional space.

This algorithm is roughly divided into two stages (a first stage and a second stage) as shown in FIG.
First, the first stage has a logical configuration centered on a feature value calculation and a membership function automatic generation process.
A membership function {μ _t , μ _n , μ _b } (Membership Func) used for fuzzy inference from the output of an oG (Laplace of Gaussian) filter
) is automatically created. The second stage following the first stage has a logical configuration centered on the fuzzy inference process and the defuzzy process, and the obtained membership function ｛μ
_t, μ _n, by using the μ _b} fuzzy rules (Fuzzy
Rule) and relaxation method (Relaxation M)
Defuzzif (Defuzzif)
y) The tumor area is determined by the processing.

These processes are described in a program code executable by a CPU such as a PC constituting the breast cancer screening system 10. In the present embodiment, three-dimensional voxel data 16a is used as an ultrasonic image.

Next, the definition of the feature value used in this embodiment will be described.

In order to simplify the process of extracting the tumor of the mammary gland 301, the ultrasonic images of the mammary gland tumor are classified into three types of regions. That is, a “tumor” (tum) with low brightness
or), “normal tissue” (normal ti) with high brightness
ssue), and the "boundary" of both
y). Breast tumors, both benign and malignant, have lower brightness than normal tissues and are surrounded by surrounding normal tissues.

In daily diagnosis, first, a "tumor" region is extracted by using such features, and a shape such as a malignant unevenness / irregularity is determined from a benign smooth / regular outline. It distinguishes benign and malignant by using features.

However, the ultrasonic image of the mammary gland 301 has a statistical value such as an average luminance of the image under the skin of the subject, in addition to the problem of the lack of speckle / noise boundaries and the deterioration of image quality inherent to the ultrasonic image. Since it changes depending on the thickness of fat, the state of surrounding tissues, and the like, it is difficult to automatically extract a tumor region on conventional image processing such as binarization or a differentiation operator. The doctor's judgment of the tumor boundary takes into account the uniformity of the echo distribution inside the tumor and the intensity of the boundary echo.

With respect to the ultrasonic image of the tumor, the “tumor” has lower luminance than the “normal tissue”, and the distribution of the luminance is small. The luminance center of gravity at the “boundary” is largely biased toward the “normal tissue” after the “boundary”, and the “normal tissue”
Of the reference area substantially coincides with the geometric center of the reference area. In the present embodiment, these expressions are quantified by the following three statistics {u, d, v}. Reference voxel volumes of the same size (gx, gy, gz)
Ask about.

Hereinafter, each processing of the tumor extraction processing of the present embodiment will be described in detail.

The feature value calculation in the first stage is performed for each of the three-dimensional voxels with the three-dimensional voxel data 16.
Probability distribution of [0,1] section of membership function {μ _t , μ _n , μ _b } using distribution of predetermined statistics related to tumor region and surrounding normal tissue and boundary between them on _f . Is a process expressed as

Here, the predetermined statistic used in the feature value calculation is a luminance average value u for each voxel, a luminance centroid (g _x , g _y , g _z ) for each voxel, and a distance between geometric centers. d, four parameters of luminance variance v for each voxel.

The average luminance value u is defined by the following equation: Average luminance value u = {f (i, j, k)} / N ³ .

Where i, j, k = 0, 1, 2,..., N
-1 (N: natural number), f (i, j, k) represents a luminance value in the voxel (i, j, k), N represents the size of the reference voxel volume (gx, gy, gz), Σ means sum operation.

That is, it is considered that the average luminance value u is small in the “tumor” region, large in the “normal tissue” region, and takes an intermediate value in the “boundary” region. The predetermined statistic used in the quantity calculation is calculated by calculating the average luminance value u for each voxel by the formula (2-
Defined in 1). As a result, it is difficult to automatically extract a tumor region on conventional image processing such as binarization and a differential operator, so that "tumor" (tumo) having low luminance is difficult.
r, the following suffix is t), “normal tissue” (nor
mal tissue, the following suffix is n), and the "boundary" (boundary, hereinafter the suffix b) of both can be identified.

The luminance centroid (g _x ,
g _y , g _z ) is g _x = {f (i, j, k) · (i + 1)} / {f (i, j, k)}, g _y = {f (i, j, k) ・ (j + 1)｝｝ / {f (i, j, k)}, g _z = {f (i, j, k) ・ (k + 1)} / ｛ΣΣΣf (i, j, k)｝... are given by equations (2-2) to (2-4).

[0141] The distance d between the center of gravity and geometric center of the luminance for each _{voxel, d = {(g x -c} x) 2 + (g y -c y) 2 + (g z -c z) 2} ^1/2 is given by equation (2-5).

Where i, j, k = 0, 1, 2,..., N
And f (i, j, k) is the voxel (i, j,
k), and N (natural number) means the size of the reference voxel volume.

That is, the value of the distance d between the center of gravity of the luminance and the geometric center of the geometric center and the center of the geometric center is equal to or smaller than the center of the luminance and the geometric center of the reference volume in the region of “normal tissue”. Are almost equal to each other, it is considered to be small. In the area of the “boundary”, the luminance is considered to be large due to the one-sided luminance at the boundary surface. If the target pixel is the “boundary”, the distance is large. On the contrary, if the distance is “normal tissue”, the distance becomes a small value.
-4) and Equation (2-5), and the predetermined statistic used in the feature calculation in the first stage is at least:
It is determined based on the center of gravity of the luminance for each voxel and the distance d of the geometric center.

The luminance variance v for each voxel is as follows: v = {f (i, j, k) -u} ² } / N ³ (2-6) where i, j, k = 0,1 , 2, ..., N-1;
f (i, j, k) represents the luminance value at the voxel (i, j, k), and N (natural number) is given by the size of the reference voxel volume (gx, gy, gz).

Here, considering that the luminance distribution is small in the “tumor” region and large in the “normal tissue” and “boundary”, the predetermined distribution used in the feature value calculation in the first stage is considered. The statistics include the luminance variance v for each voxel
Expression (2-6) is applied by paying attention to the fact that

The luminance average u is considered to be small in the “tumor” region and large in the “normal tissue” region. In the "boundary" area, it is considered to take an intermediate value.

On the other hand, the value of the distance d between the luminance center and the geometric center of the luminance center and the geometric center is “normal tissue”.
In the area of, the center of gravity of the luminance and the geometric center of the reference volume are almost the same, so it is thought to be small. However, in the area of “boundary”, the luminance is biased to one side with the boundary surface as a large value. Conceivable. In addition, "tumor"
In the area of, a low-luminance part and a slightly high-noise part may be mixed, so the value of d has a part that overlaps with the d value of “normal tissue”.
Tends to be smaller than the value.

On the other hand, the luminance distribution v is considered to be small in the region of “tumor” and large in the region of “normal tissue” and “boundary”. FIG. 5 shows a two-dimensional conceptual diagram of the center of gravity of luminance and the distance d between the center of gravity of the geometric center and the geometric center. That is, if the pixel of interest is “boundary”, d is large, and if it is “normal tissue”, d is small.

In the first stage of the tumor extraction processing of the present embodiment,
Is the membership function ｛μ_t, Μ_n, Μ _bAutomatic creation of｝
Do. Membership function ｛μ_t, Μ_n, Μ_bGot｝
By creating for each three-dimensional voxel data 16a
Tumor region can be extracted stably regardless of the system environment.
Is expected. Membership function ｛μ_t, Μ_n,
μ_b｝ Is a three-dimensional LoG (L
output of the place of Gaussian filter
Force, that is, greater than zero. Zero crossing point and
Hist of each feature distribution in three regions smaller than 0
Calculated from grams. Figure 6 shows the membership function
｛Μ_t, Μ_n, Μ_bThe procedure (flow) for creating 作成 is shown below.

As shown in FIG. 6, the membership function automatic generation step in the first stage is performed by automatically generating a membership function {μ _t , μ _n , μ _b } using a three-dimensional Gaussian Laplace filter (three-dimensional LoG filter). The creation process,
Subsequent to the feature calculation, the membership functions ｛μ _t , μ _n , μ
_{b This} is a process of automatically extracting a tumor region by summarizing the distribution of statistics based on a fuzzy inference process with｝.

A voxel used to create the membership function {μ _t , μ _n , μ _b } is a three-dimensional Gaussian Laplace filter (three-dimensional LoG filter) g represented by the following equation.
(R) Determined from the output of (Equation (2-9) described later).

The LoG filter is a kind of second-order differential filter or band-pass filter, and is often used in image processing to extract a target boundary. When this filter is used, the second derivative becomes zero at the position of the boundary, or the output of the filter shows a positive / negative change, that is, a zero crossing. Connecting the zero crossings of these outputs forms a boundary. Therefore, a three-dimensional Gaussian Laplace filter (three-dimensional L
When the zero crossing point of the output of the oG filter is connected, it becomes a boundary surface of the target region to be extracted.

In the tumor extraction processing of this embodiment, in addition to the extraction of the “boundary” by zero crossing, the positive value of the output is “tumor” with low luminance and the negative value is “normal tissue” with high luminance. Utilization is used to classify voxels into three classes.

Grades μ _t , μ _n , and μ g belonging to three classes of “tumor”, “normal tissue”, and “boundary” of each voxel.
μ _b is the membership function ｛μ _t , μ calculated in the first stage
_n , μ _b } and a fuzzy rule such as the following equation (2-7).

Therefore, the fuzzy inference step is a fuzzy inference
Perform grade calculation and determine membership function
Of the average luminance value u generated in the processing
Function ｛μ_t, Μ_n, Μ_b｝ | U, luminance centroid and geometric
Membership function ｛μ for center distance d_t, Μ_n, Μ
_bMembership function for｝ | d and luminance variance v
｛Μ _t, Μ_n, Μ_b｝ ｜ Consists of v and fuzzy rules
Based on the fuzzy inference process performed,
Grade t belonging to class R1 of "tumor", "normal tissue"
Grade n belonging to class R2 of class R of "boundary"
3 grades {t, n, b} of grade b belonging to
This is the process of executing the classification by using.

Here, the fuzzy rule is if the
It is described in rules expressed in the form of a conditional statement. The fuzzy rule (Equation (2-7)) is shown below.

R1: if (uismall) and (dis medium) and (vismall) then the voxelis “tumor”, R2: if (uislarge) and (dis medium) is large) then the voxel is “normal tissue”, R3: if (uis medium) and (dis large) and (vis medium) then the voxel is a formula, “both formula” Membership function used in fuzzy inference process
(｛Μ_t, Μ_n, Μ_b｝ | U, ｛μ_t, Μ_n, Μ_b｝ | D, ｛μ
_t, Μ_n, Μ_b｝ | V) is used to determine the membership function.
Function for the average luminance value u generated
｛Μ_t, Μ_n, Μ _b｝ | U, the center of gravity and the geometric center
Membership function ｛μ for distance d_t, Μ_n, Μ_b｝ ｜
d, membership function 輝 度 μ for luminance variance v_t, Μ_n,
μ_b｝ | V is included at least.

In the fuzzy inference step, a first logical process and a second logical process are executed.

The first logical processing is a membership function ({μ _t , μ _n , μ _b } | u, a distance between the center of gravity of the luminance and the geometric center, with respect to the luminance average value u generated in the membership function determination processing. membership function ｛μ _t , μ _n , μ
_The grade μ belonging to three classes for each of the feature quantities {u, d, v} in each voxel based on the membership function {μ _t , μ _n , μ _b ｝ | v) for _b ｝ | d and the luminance variance v _t | u, μ _t | d, μ _t | v, μ _n | u, μ
_n | d, μ _n | v, μ _b | u, μ _b | d, and μ _b | v.

The second logical processing is performed by using the grade μ _t | u, μ _t |
d, μ _t | v, μ _n | u, μ _n | d, μ _n | v, μ _b | u,
μ _b | d, μ _b | based on each value of v, the analog value that defines the analog value mu _t, "normal tissue" ness which is input to the mechanism of the fuzzy inference defining a "tumor" likelihood of each voxel a process of obtaining the analog value mu _b defining a mu _n or "boundary" ness.

The fuzzy inference mechanism is expressed by equation (2-8).

[0162] _{R1: μ t = min (μ} t | u, μ t | d, μ t | v), R2: μ n = min (μ n | u, μ n | d, μ n | v), R3 : Μ _b = min (μ _b | u, μ _b | d, μ _b | v) Expression (2-8) where min (a ₁ , a ₂ , a ₃ ) is a ₁ , a ₂ , a ₃ It is expressed by an operation that selects the minimum value from.

As a result, if shown in the equation (2-7)
The “and” logical operation in the “if to then” conditional statement in the rule expressed in the “then else” conditional statement format means that the probability that a certain fact can occur is the minimum probability of each condition. Conditional expressions can be expressed by fuzzy logical expressions.

The membership function automatic generation step is performed when the statistical amount including the luminance distribution on the mammary gland 301 ultrasonic image changes due to the thickness of the subcutaneous fat of the subject and the state of the surrounding mammary gland 301 tissue. Membership functions used for fuzzy inference (， μ _t , μ _n , μ _b ｝ | u, ｛μ _t , μ _n , μ _b ｝
| D, ｛μ _t , μ _n , μ _b ｝ | v) are converted into a three-dimensional Gaussian Laplace filter (three-dimensional LoG filter)
This is a process of automatically creating a job based on the output of the process.

That is, it is a kind of second-order differential filter or band-pass filter. In image processing, a simple three-dimensional Gaussian Laplace filter (three-dimensional LoG filter) often used in target boundary extraction is used. As a result, the second derivative becomes zero at the position of the boundary of the voxel data, or a change in the output of the filter is positive or negative, that is, zero crossing appears.
Connecting the zero crossings of these outputs forms a boundary. Therefore, when the zero crossing point of the output of the three-dimensional Gaussian Laplace filter (three-dimensional LoG filter) is connected, it becomes the boundary surface of the target area to be extracted. Therefore, in addition to the extraction of the "boundary" by zero crossing, the voxel is classified into three classes by utilizing the fact that the positive value of the output is "tumor" with low luminance and the negative value is "normal tissue" with high luminance. Be able to classify. By executing such a simple class classification, a membership function (｛μ _t , μ) capable of performing calculation (image processing) in a short time (high speed) with a small computer resource having a calculation capability of about PC. _n , μ _b ｝ | u, ｛μ _t , μ _n , μ _b ｝ | d,
｛Μ _t , μ _n , μ _b ｝ | v) can be automatically created.

In the membership function automatic generation step, the membership function (｛μ _t , μ _n , μ _b ｝ | u, ｛μ _t , μ _n ,
The voxels used to create μ _b ｝ | d, ｛μ _t , μ _n , μ _b ｝ | v) are represented by a three-dimensional Gaussian-Laplace filter (three-dimensional LoG filter) g (r). Laplace filter (three-dimensional LoG filter) A three-dimensional Gaussian Laplace filter (three-dimensional LG filter) obtained from the output
oG filter) processing and the zero-crossing point of the output of the three-dimensional Gaussian-Laplace filter (three-dimensional LoG filter) to connect the "boundary" of the target area to be extracted
, A three-dimensional Gaussian Laplace filter (three-dimensional LoG filter) classifying voxels whose output values are positive into “tumors” with low luminance, and a three-dimensional Gaussian Laplace filter (3 Dimension Lo
G filter) A voxel whose output shows a negative value has high luminance.
A normal tissue extraction process classified as “normal tissue” and a voxel classified as “tumor” in the tumor extraction process are subjected to dilation / contraction processing to classify as a “tumor” that is isolated within normal tissue Expansion / contraction processing that removes the voxels classified as “borders” and voxels classified as “borders”, and separates voxels classified as “tumors” from other low-brightness closed areas when they are connected. A feature amount operation is performed only on the voxels that have been classified into the “boundary” before and after the contraction processing, and the luminance average value u, the distance d between the center of gravity and the geometric center and the luminance are calculated. A voxel selection / feature amount calculation process for calculating three feature amounts {u, d, v} of variance v, and three feature amounts {u, d, v} obtained in the voxel selection / feature amount calculation process On the other hand, members corresponding to each probability density function Membership function _{({μ t, μ n,} μ b}
_{| U, {μ t, μ} n, μ b} | d, {μ t, μ n, μ b} | v)
And a membership function determination process for determining

Here, the three-dimensional Gaussian Laplace filter (three-dimensional LoG filter) g (r) is given by the equation (2-9)
It is represented by

G (r) = (R ² −3σ ² ) / {(2π ³ ) ^1/2 · σ ⁷ } · exp {−r ² / 2σ ² } Equation (2-9) where r is the origin The distance from σ is defined as the standard deviation of Gaussian.

A three-dimensional Gaussian-Laplace filter (3
If the (dimensional LoG filter) processing is performed by convolution in the real space, there is a practical problem that the processing time becomes enormous. Therefore, in the tumor extraction processing of the present embodiment, the processing is speeded up by using the FET.

By providing such a three-dimensional Gaussian-Laplace filter (three-dimensional LoG filter) process, classifying “tumor” voxels and “normal tissues”
Three feature quantities {u, d, v} can be calculated for all voxels. Also, by calculating three feature values {u, d, v} only for voxels classified as “boundary” before and after the dilation / shrinkage processing, the feature value of “boundary” is changed to another class. Even if the number of voxels is small compared to, erroneous "boundary" voxels can be excluded as much as possible.

The voxel selection / feature calculation processing is performed by
Before and after the contraction processing, the feature amount calculation is performed only on the voxels that have been classified into the “boundary”, and the brightness average value u (first feature amount), the center of gravity of the brightness and the geometric The three feature quantities {u, d, v} of the center distance d (second feature quantity) and the luminance variance v (third feature quantity) are calculated.

The membership function determination processing is performed for the three feature quantities ｛u, d,
v}, membership functions ({μ _t , μ _n , μ _b } | u, ｛μ _t , μ _n , μ
_b ｝ | d, ｛μ _t , μ _n , μ _b ｝ | v).

The feature value average luminance, the distance between the center of gravity of the luminance and the geometric center, and the luminance variance defined in the tumor extraction processing of the present embodiment are three of “tumor”, “normal tissue” and “boundary”. It is necessary to find membership functions {μ _t , μ _n , μ _b } indicating “likeliness” belonging to two classes. Therefore, as described above, the three-dimensional Gaussian Laplace filter (three-dimensional LoG filter) is applied to the voxel data, and the data is classified into three regions based on the positive / negative and zero crossing points of the output.

However, since the voxel region obtained here contains voxels classified incorrectly, the “tumor”
For voxels classified as
(Tion & Erosion) treatment, and "tumor" voxels and "boundaries" that are isolated in normal tissue
The voxel is removed, and when the "tumor" is connected to another low-luminance closed area, the two are separated. For this purpose, in the present embodiment, the membership function {μ _t , μ _n , μ _b } is approximated by the Rician function P _A (x).

The three feature amounts are calculated for all the “tumor” voxels and the “normal tissue” voxels classified by such processing. Since the number of voxels in the feature value of the “boundary” is smaller than that of the other classes, it is necessary to exclude erroneous “boundary” voxels as much as possible. Therefore, three feature values are calculated only for voxels classified as “boundary” before and after the expansion / contraction processing.

One idea is to use the probability density coefficient of the feature calculated from the classified voxels as the membership function {μ _t , μ _n , μ _b }. Classification by the Laplace filter (three-dimensional LoG filter) is insufficient, and it is necessary to consider that voxels classified incorrectly are included.

It is known that the probability density coefficient of the brightness of an ultrasonic image is represented by a Rician function P _A (x) represented by the following equation.

Therefore, in the present embodiment, the probability density coefficients for the three characteristic quantities {u, d, v} in the membership function determination processing are defined by the Rician function P _A (x). Here, the Rician function P _A (x) is given by Expression (2-10).

P _A (x) = x / σ ² · exp {(− x ² + σ ² ) / σ ² } · I ₀ (xs / σ ² ) Equation (2-10) where I ₀ (x ) Is a 0th-order modified Bessel function of the first kind.

The Rician function P _A (x) is a Rayleigh distribution function when s = 0 (FIG. 7)
As the s / σ increases, it approaches the Gaussian (Gaussian) distribution function (see FIG. 7). In the case where the scatterer, which is the reflection source of the ultrasonic wave, is smaller than the wavelength and the scatterer is randomly distributed, a speckle pattern appears in a so-called image, and the luminance fluctuation is represented by Raylei expressed by the following equation.
gh distribution function. On the other hand, it is shown that when a reflection source larger or smaller than the wavelength is mixed, the peak position of Rician gradually moves away from the origin, and approaches the Gaussian function from the Rayleigh distribution function.

On the other hand, the probability density function of the average luminance value u in the feature quantities {u, d, v} is approximated to the probability density function of an ultrasonic wave. Therefore, the Rayleigh distribution is used in the “tumor” region having relatively low luminance. It is a function, and Gauss with large reflective sources is mixed in "boundary" and "normal tissue".
This can be expected to be a sian function (Gaussian distribution function).

Therefore, the Rician function P is defined as the probability density coefficient for each of the three feature quantities {u, d, v}.
By using _A (x), the membership function ({μ _t , μ _n , μ _b } | u, {μ _t , μ _n , μ _b }
| D, ｛μ _t , μ _n , μ _b ｝ | v)
By approximating by the ailelight distribution function, other "normal tissues" and "boundary" can be approximated by the Gaussian function.

FIG. 8 shows “tumor”, “normal tissue”, and “boundary”.
And the approximated membership functions {μ _t , μ _n , μ _b } for the voxels belonging to. From the figure, it can be seen that the two agree well.

The distance d between the center of gravity and the geometric center in the feature amount {u, d, v} tends to be small in the “normal tissue” and “tumor” regions and large in the “boundary” region. is there. As can be seen from the histogram of the distance d between the center of gravity of the luminance and the geometric center shown in FIG. 9, in each case, Rayleigh is higher than Gaussian.
The distribution is better approximated. Therefore, the membership function ｛μ _{t of} the feature d (the distance between the center of gravity and the geometric center),
μ _n , μ _b } are approximated by a Rayleigh distribution (FIG. 11).
reference).

The luminance variance v in the feature amounts {u, d, v}
Is shown in FIG. The approximation is better in the Rayleigh distribution as shown in FIG. Therefore, the membership function {μ _t , μ _n , μ _b } of the luminance variance v
Is approximated by a Rayleigh distribution.

The three-dimensional Gaussian-Laplace filter (3
When the histogram is calculated from the voxels classified according to the output of the (dimensional LoG filter), the Rayle assumed in each case is affected by the voxels classified incorrectly.
The distribution becomes different from that of the light distribution or Gaussian. As an example, the histogram is approximated to the histogram obtained for the “boundary” of the feature amount d (the distance between the luminance center and the geometric center).

As described above, Rici is used as the probability density coefficient.
By using the an function P _A (x), the membership function (｛μ _t ,
μ _n , μ _b ｝ | u, ｛μ _t , μ _n , μ _b ｝ | d, ｛μ _t , μ _n ,
μ _b ｝ | v) can be approximated by a Rayleigh distribution function.

With the same intent, the luminance variance v in the feature quantity {u, d, v} is calculated as Ric as a probability density coefficient.
By using the ian function P _A (x), the luminance variance v
Membership functions (｛μ _t , μ _n , μ _b ｝ | u, ｛μ _t ,
μ _n , μ _b ｝ | d, ｛μ _t , μ _n , μ _b ｝ | v) are Rayle
It can be approximated by an igh distribution function.

Therefore, in the present embodiment, the membership function ({μ _t , μ _n , μ _b } | u,
｛Μ _t , μ _n , μ _b ｝ | d, ｛μ _t , μ _n , μ _b ｝ | v)
The "tumor" is approximated by a probability density function expressed by a Rayleigh distribution function.

Thus, the Rician function P _A (x)
Is a Rayleigh distribution function when s = 0,
As s / σ increases, it approaches the Gaussian function.
In the case where the scatterer which is the reflection source of the ultrasonic wave is smaller than the wavelength and the scatterer is randomly distributed, a speckle pattern appears in a so-called image, and the luminance variation is represented by a Rayleigh distribution function represented by the following equation. Becomes On the other hand, it is shown that when a reflection source larger or smaller than the wavelength is mixed, the peak position of Rician gradually moves away from the origin, and approaches the Gaussian function from the Rayleigh distribution function.

On the other hand, the probability density function for the average luminance value u in the feature amounts {u, d, v} is approximated to the probability density function of the ultrasonic wave, so that the Rayleigh distribution is used in the “tumor” region with relatively low luminance. It is a function, and Gauss with large reflective sources is mixed in "boundary" and "normal tissue".
This can be expected to be the sian function. For this reason, the membership function determination processing is performed for each of the three feature quantities {u, d, v} obtained in the voxel selection / feature quantity calculation processing by a membership function (｛ μ _t , μ _n , μ _b ｝ | u, ｛μ _t , μ _n , μ _b ｝ | d, ｛μ
_t , μ _n , μ _b ｝ | v), the membership function (｛μ _t , μ _n , μ _b ｝ | u, ｛μ _t ,
μ _n , μ _b ｝ | d, ｛μ _t , μ _n , μ _b ｝ | v)
Is approximated by a probability density function expressed by a Rayleigh distribution function, and "normal tissue" and "boundary" are approximated by a probability density function expressed by a Gaussian distribution.

That is, by using the Rician function P _A (x) as the probability density coefficient for the first feature value, the membership function (｛μ _t , μ _n , μ _n ) of the average luminance value u is obtained.
_{b} | u, {μ t} , μ n, μ b} | d, {μ t, μ n, μ b} |
For v), “tumor” is approximated by Rayleigh distribution function, and “boundary” and other “normal tissues” are Gaussian.
It can be approximated by the ian function. In this way, Ray
By approximating the membership function {μ _t , μ _n , μ _b } with the Leigh distribution or Gaussian, it is considered that the effect of voxels that are incorrectly classified can be reduced.

Also, the distance between the center of gravity of the luminance and the geometric center
d to the membership function (｛μ _t, Μ_n, Μ_b｝ | U,
｛Μ_t, Μ_n, Μ_b｝ | D, ｛μ_t, Μ_n, Μ_b｝ | V) Ra
Approximation by probability density function expressed by yleigh distribution function
I'm going to.

Thus, the Rician function P _A (x)
Is a Rayleigh distribution function when s = 0,
As s / σ increases, it approaches the Gaussian function.
In the case where the scatterer which is the reflection source of the ultrasonic wave is smaller than the wavelength and the scatterer is randomly distributed, a speckle pattern appears in a so-called image, and the luminance variation is represented by a Rayleigh distribution function represented by the following equation. Becomes

On the other hand, it is shown that when a reflection source larger or smaller than the wavelength is mixed, the peak position of Rician gradually moves away from the origin and approaches the Gaussian function from the Rayleigh distribution function. . On the other hand, the probability density function of the average luminance value u in the feature amounts {u, d, v} is approximated to the probability density function of the ultrasonic wave.
The Gaussian distribution function becomes an ailelight distribution function.
It can be expected that it will be an function. For this reason, the membership function determination processing is performed for each of the three feature quantities {u, d, v} obtained in the voxel selection / feature quantity calculation processing by a membership function (｛ μ _t ,
μ _n , μ _b ｝ | u, ｛μ _t , μ _n , μ _b ｝ | d, ｛μ _t , μ _n ,
When calculating μ _b ｝ | v), the membership function (｛μ _t , μ _n ,
μ _b ｝ | u, ｛μ _t , μ _n , μ _b ｝ | d, ｛μ _t , μ _n , μ _b ｝
| V) is approximated by a probability density function represented by a Rayleigh distribution function.

That is, by using the Rician function P _A (x) as the probability density coefficient for the second feature, the distance d between the center of gravity of the brightness and the geometric center in the feature {u, d, v} can be calculated. , "normal tissue" and "tumor" small in area, tends to increase the "boundary" area, Rician function P _A as the probability density factor
By using (x), the membership function ({μ _t , μ _n , μ _b ｝ |) of the distance d between the center of gravity and the geometric center is
_{u, {μ t, μ n} , μ b} | d, {μ t, μ n, μ b} | v) is as can be approximated by the Rayleigh distribution function.

Also, membership functions (輝 度 μ _t , μ _n , μ _b ｝ | u, ｛μ _t , μ _n , μ _b ｝ | d, ｛μ
_t , μ _n , μ _b ｝ | v) are approximated by a probability density function expressed by a Rayleigh distribution function.

Thus, the Rician function P _A (x)
Is a Rayleigh distribution function when s = 0,
As s / σ increases, it approaches the Gaussian function.
In the case where the scatterer which is the reflection source of the ultrasonic wave is smaller than the wavelength and the scatterer is randomly distributed, a speckle pattern appears in a so-called image, and the luminance variation is represented by a Rayleigh distribution function represented by the following equation. Becomes On the other hand, it is shown that when a reflection source larger or smaller than the wavelength is mixed, the peak position of Rician gradually moves away from the origin, and approaches the Gaussian function from the Rayleigh distribution function.

On the other hand, the probability density function for the average luminance value u in the feature amounts {u, d, v} is approximated to the probability density function of the ultrasonic wave. Therefore, the Rayleigh distribution is used in the “tumor” region having relatively low luminance. It is a function, and Gauss with large reflective sources is mixed in "boundary" and "normal tissue".
This can be expected to be the sian function. For this reason, the membership function determination processing is performed for each of the three feature quantities {u, d, v} obtained in the voxel selection / feature quantity calculation processing by a membership function (｛ μ _t , μ _n , μ _b ｝ | u, ｛μ _t , μ _n , μ _b ｝ | d, ｛μ
_t , μ _n , μ _b ｝ | v), the membership function (｛μ _t , μ _n , μ _b ｝ | u, ｛μ _t ,
μ _n , μ _b ｝ | d, ｛μ _t , μ _n , μ _b ｝ | v)
It is approximated by the probability density function represented by the eigh distribution function.

That is, by using the Rician function P _A (x) as the probability density coefficient for the third feature, the luminance variance v of the feature {u, d, v} is calculated as the Rician function P _A as the probability density coefficient. By using (x), the membership function (輝 度 μ _t , μ _n , μ _b ｝ | u, ｛μ _t , μ _n , μ _b ｝ | d, ｛μ
_t , μ _n , μ _b ｝ | v) can be approximated by a Rayleigh distribution function.

Next, each processing of the second stage of the tumor extraction processing of the present embodiment will be described in detail.

In the second stage, the tumor area is extracted by fuzzy inference using the membership functions {μ _t , μ _n , μ _b } obtained in the previous section. FIG. 12 shows the flow of the processing.

First, images representing grades of three classes are created by fuzzy inference. That is, a "tumor" having a low luminance, a "normal tissue" around a high luminance, and a "boundary" between the two. Next, defuzzification is performed by the relaxation method, and all voxels are divided into three classes. The processing in each processing will be described below.

In the fuzzy inference step in the second stage, following the process of automatic extraction, the generated membership function ({μ _t , μ _n , μ _b } | u, {μ _t , μ _n , μ _b } | D, ｛μ
_t , μ _n , μ _b ｝ | v) and a fuzzy rule, which is a process of classifying each voxel into a predetermined number of types of regions based on a fuzzy inference process.

In the fuzzy inference step, each voxel determines the degree of “likelihood” of being “tumor,” “normal tissue,” or “boundary,” and classifies the voxels in accordance with the obtained degree of “likeness”. Execute

That is, voxels that have been subjected to a three-dimensional Gaussian-Laplace filter (three-dimensional LoG filter) and classified into three classes from the positive / negative and zero crossing points of the output include voxels classified incorrectly. there is a possibility. Therefore, by providing such a fuzzy inference process, a voxel classified as a “tumor” is subjected to dilation / shrinkage processing to obtain a “tumor” voxel or a “boundary” that is isolated within normal tissue. The voxel is removed, and when the "tumor" is connected to another low-luminance closed area, the two are separated. As a result, while gradually resolving the contradiction in the three-dimensional space, it is possible to finally execute the processing for determining the tumor region.

Next, all voxel data are classified into three regions by a defuzzification (defuzzification) process based on the relaxation method from the grade images {μ _t , μ _n , μ _b } regarding these three attributes. . In this embodiment, a “normal tissue” in contact with a tumor is defined as a “boundary”.

For this reason, in the second stage, the defuzzification step following the fuzzy inference step is processing for classifying by the relaxation method following the fuzzy inference step, and the defuzzification step based on the relaxation method is applied to each voxel. And a process of classifying each voxel into one of "tumor", "normal tissue" or "boundary" and performing a final determination of the tumor region, following the pre-processing. I have.

Here, the process of performing the defuzzification process based on the relaxation method (Relaxation Method) for each voxel in the defuzzy process is performed by using 3 times for each of the feature amounts {u, d, v} in each voxel.
Images of grades belonging to two classes {μ _t , μ _n , μ _b ｝
When performing a defuzzification process based on the relaxation method and classifying all the voxel data into three regions, a "tumor" is contacted based on the following rules (1) to (3): "Normal tissue" is defined as "boundary."

Rule (1) If the voxel of interest is "tumor", it is defined that it does not come into contact with a voxel classified as "normal tissue".

Rule (2) If the voxel of interest is a “boundary”, it must be a “tumor”
Is defined as touching the voxel of "normal tissue".

Rule (3) If the voxel of interest is “normal tissue”, it does not come into contact with a voxel classified as “tumor”. As a result, all voxel data can be classified into three regions by the defuzzification (defuzzification process) process based on the relaxation method from the grade images {μ _t , μ _n , μ _b } regarding the three attributes.

A local constraint rule used in the parallel repetition processing is defined by the following rule (formula (2-11)) expressed in an if then else conditional statement format.

R1: if N _t > 1 and N _b ≧ 2 and N _n = 1 then μ _t ↓, μ _n ↓, μ _b ↑, R2: else if N _n > 0 and N _b ≧ 1 and N _t = _{1 then μ t ↑, μ n} ↓, μ b ↓, R3: else if N t = 0 and N b ≧ 1 and N n ≧ 1 then μ t ↓, μ n ↑, μ b ↓, R4: else if N _{t> N n +12 then μ t} ↑, μ n ↓, μ b ↓, R5: else if N n> N t +12 then μ t ↓, μ n ↑, μ b ↓, R6: else then μ t ↓, μ _n ↓, μ _b ↑, ... Equation (2-11) (However, A 定 数 is obtained by adding a constant C to the value of A (A
+ C), A ↓ is an operator meaning subtracting a constant C from the value of A (AC). Also, a constant 12 in R3 and R5 is a 3 × connected to a voxel of interest. This is a numerical value indicating that the difference between N _n and N _t is considerably large in a 3 × 3 region, and this value itself is obtained by trial and error so that a human can optimally see the defuzzy result. In the application processing of the tumor extraction processing according to the present embodiment, the increase / decrease constant C of the symbol ↑ ↓ in the equation (2-11) is set to 0.25. 0.0 if less than 0
I have to.

Specifically, for each voxel, R1 to R
6 are processed in order, and if any of them is satisfied, the subsequent rules are ignored. Also, this defuzzification process is repeatedly performed in parallel, and μ _t ,
The process ends when the sum of the change amounts of μ _n and μ _b becomes less than a certain threshold. At this point, μ _t , μ _n ,
The element with the largest value in μ _b is determined as the attribute of that voxel, and as a result, each voxel is eventually assigned one of the elements “tumor”, “normal tissue”, or “boundary”. Become.

In addition, one of the elements “tumor”, “normal tissue”, and “boundary” of the surrounding voxels is assigned. Therefore, for example, even if the grade μ _b belonging to the “boundary” of a certain voxel is initially large, if only voxels of a large grade belonging to the “tumor” (or “normal tissue”) exist around the voxel, The grade “ _t” (or μ _n ), which is a “tumor” (or “normal tissue”), is large due to iterative processing;
Grade μ _b belonging to “normal tissue” (or “tumor”)
And μ _n (or μ _t ) are changed to be smaller (see R3 and R5 in equation (2-11)).

[0217] Conversely, two or more around even grade mu _b of "boundary" ness initially smaller "boundaries" of the voxel, grade of one or more "tumor" ness mu _t
If if is large voxel grades mu _n grades mu _b and "normal tissue" ness of "boundary" ness, the "boundary" ness grade mu _b is larger value of the voxel, "tumor" ness grade mu grade mu _n grades mu _b and "normal tissue" ness of _t and "boundary" ness is gradually updated to a smaller value (see R1 in formula (2-11)).

For the voxel data classified into three classes, only the tumor region is extracted by the voxel integrity. ROI (Region of inter) of voxel data processed in the tumor extraction processing of the present embodiment
est) is a closed area surrounded by “normal tissues”, and the area may include “normal tissues” classified into an incorrect class. Therefore, the area of the tumor at the target point of interest ROI is determined by the following rule.

Processing (1) Assuming that the entire voxel data is surrounded by “normal tissue”, starting from any one voxel of the “normal tissue” surrounding the voxel data, one “tumor” To search for voxels.

Processing (2) “Connected to voxel of one searched“ tumor ”
Search for all tumor "voxels" and label those voxels as "tumor1".

Processing (3) The processing is returned to processing (1), and “tumor 1” and “tumor 2”
.., Labeling the area of the tumor as “tumor n”.
When this process ends, the process moves to the next process.

Processing (4) Labeled “tumor k” (k = 1, 2, 3,..., N)
If the volume is arranged in order from large to small, and if it is assumed to be a sphere, those with a diameter of 2 mm or less are discarded as noise, and the center of gravity of the remaining "tumor" is the center of the ROI. The closest one is finally determined as "tumor".

Small size (about 2.8 × 2.2 × 1.5 cm ³ )
By attaching the lightweight (17.0 g) AC magnetic field position sensor 131 to the ultrasonic ultrasonic probe 121,
While the examiner manually scans the ultrasonic probe 121 in the same manner as daily diagnosis, the three-dimensional ultrasonic image data 15b
Was able to collect. When applied to 16 cases of malignant tumors and 11 cases of benign tumors (fibroadenomas), all of them successfully extracted tumors, suggesting the effectiveness of this method.

As described above, in summary of the tumor extraction processing of this embodiment, the feature value calculation is performed for each of the three-dimensional voxels, on the three-dimensional voxel data 16a, on the tumor region, the surrounding normal tissue, and the boundary between the two. The membership function (｛μ _t , μ _n , μ _b ｝ | u, ｛μ _t , μ _n , μ _b ｝ | d, ｛μ
_t , μ _n , μ _b ｝ | v) are processed as a probability distribution of [0, 1] division. In addition, the membership function automatic generation process includes a membership function ({μ _t , μ _n , μ _b ｝ | u,
Based on the fuzzy inference process with ｛μ _t , μ _n , μ _b ｝ | d, ｛μ _t , μ _n , μ _b ｝ | v) Execute the process of automatically extracting the area of. As a result, the three-dimensional Gaussian Laplace filter (three-dimensional LoG filter) is applied to the voxel data, and the output is classified into three classes from the positive / negative and zero crossing points of the output. Membership function (｛μ
_t , μ _n , μ _b ｝ | u, ｛μ _t , μ _n , μ _b ｝ | d, ｛μ _t ,
μ _n , μ _b ｝ | v) can be obtained. 1
A kind of second-order differential filter or band-pass filter. In image processing, a simple three-dimensional LoG (Laplaceof Ga
ussian function) filter, resulting in voxel
At the position of the data boundary, the second derivative becomes zero, or the output of the filter changes between positive and negative, that is, the zero crossing (Zer
o crossing) appears. Connecting the zero crossings of these outputs forms a boundary. Therefore, 3
When a zero crossing point of the output of the three-dimensional Gaussian Laplace filter (three-dimensional LoG filter) is connected, it becomes a boundary surface of the target region to be extracted. Therefore, in addition to the extraction of the "boundary" by zero crossing, the voxel is classified into three classes by utilizing the fact that the positive value of the output is "tumor" with low luminance and the negative value is "normal tissue" with high luminance. Be able to classify. By executing such a simple class classification, a membership function (｛μ _t , μ) capable of performing calculation (image processing) in a short time (high speed) with a small computer resource having a calculation capability of about PC. _n ,
μ _b ｝ | u, ｛μ _t , μ _n , μ _b ｝ | d, ｛μ _t , μ _n , μ _b ｝
| V) can be automatically created. The membership function (フ ァ μ _t , μ _n , μ _b ｝ | u, ｛μ _t , and the membership function generated when the fuzzy inference step executes the membership function automatic generation step obtained in the membership function automatic generation step μ _n , μ _b ｝
| D, ｛μ _t , μ _n , μ _b ｝ | v), and a process of classifying each voxel into a predetermined number of types of regions based on a fuzzy inference process including fuzzy rules. Execute. In the Defuzzy process, a defuzzification process based on the relaxation method is performed on each voxel, and after this process (process of the defuzzification process), each voxel is replaced with “tumor”, “normal tissue”. A process is performed to classify the tumor area as either “or” or “boundary” to make a final decision on the tumor region. By providing such a fuzzy inference process and a defuzzy process, a simple fuzzy inference that can be calculated (image processed) in a short time (high speed) with a small computer resource having a calculation capability of a personal computer (PC) is provided. Using voxel and relaxation methods, each voxel can be classified into three classes: "tumor", "normal tissue", and "boundary" between them. As a result, P
It can be calculated (image processed) in a short time (high speed) with a small computer resource having a calculation capability of about C.
Tumors (especially mammary tumors) from simple ultrasound three-dimensional images from which low-brightness areas other than tumors such as artifacts such as speckle noise and acoustic shadows, missing borders, and muscle layers have been removed "Tumor") with high accuracy,
It becomes possible to automatically extract a judgment of a malignant tumor (cancer) with good reproducibility.

[Automatic Malignant Tumor Identification Process] Next, an embodiment of an automatic malignant tumor identification process (automatic identification algorithm for mammary gland tumor), which is a post-process of the breast cancer screening system 10, will be described with reference to the drawings. FIG. 13 is a processing flow of the malignant tumor automatic identification processing of the present embodiment.

The automatic malignant tumor identification process of the present embodiment is a process following the tumor extraction process which is a post-process of the breast cancer screening system 10. The process is performed on a three-dimensional tomographic image such as an MRI image or an ultrasonic image of a living body. A method applicable to a pathological diagnosis support system for extracting a boundary between tissues based on voxel data 16a, which is a form of image data, and finding a breast cancer tissue from normal tissues. It has a logical configuration centered on the integer quantification step (Step P10) and the tumor diagnosis step (Step P20), and includes a CP such as a PC constituting the breast cancer screening system 10 as described above.
It is described by a program code executable by U.

In the automatic malignant tumor identification processing of the present embodiment,
The area of the tumor obtained by the automatic extraction method of the breast tumor described above is displayed three-dimensionally, and the geometrical irregularities on the surface are quantitatively measured and evaluated. The three-dimensional representation of the tumor surface is performed using surface rendering, which allows the shading of the tumor surface to be similar to the light shading that occurs when we look at an object. As a result, it becomes possible to observe subtle changes in the unevenness of the tumor surface shape, and it becomes possible to make an understanding and judgment based on common recognition.

The most distinctive feature of breast tumor diagnosis is that it involves benign or malignant discrimination, and among them, discriminative diagnosis of small breast cancers of 1 cm or less is the most important. In addition, a benign tumor (fibroadenoma) has a "regular and smooth" contour shape such as a circular or elliptical shape.

On the other hand, a malignant tumor (cancer) has an irregular and irregular shape such as a crab shape or a star shape. In normal ultrasonography, such geometrical features in tomographic images are often used, but if the tumor is still small, observing the tomographic image alone will determine whether the tumor is malignant or benign. Sometimes it is difficult to know what is.

Therefore, the present embodiment utilizes the fact that the surface shape of a malignant tumor (cancer) is more irregular than that of a benign tumor (fibroadenoma), and that the surface area is larger than that of a benign tumor (fibroadenoma) having the same volume. I do. Therefore, in order to quantify the irregularity of the surface shape (step P10), a parameter (tumor shape determination parameter γ) using the ratio of the extracted surface area and volume of the tumor is defined (step P13).
Tumor shape determination parameter γ (= Surface) of the ratio of surface area to volume normalized in the case of a sphere (Step P14)
³ / Volume ² ratio, S / V ratio)
Means that malignant tumor (cancer) shows higher value than benign tumor (fibroadenoma).

For this purpose, in the tumor irregularity quantification step (step P10), as shown in FIG. 13, the surface area S (step P11) and the volume V (step P12) of the tumor are used as indices for measuring the degree of irregularity of the tumor surface shape. ) Is used to calculate the tumor shape determination parameter γ (step P13).

Here, the tumor shape determination parameter γ is defined by the following equation: γ = (S ³ / V ² ) / κ, κ = constant (3-1) The constant κ is γ when the tumor is spherical.
Is normalized to be 1 (step P14), specifically, 36 / π (step P14).

In order to assist the breast cancer screening system 10 in diagnosing breast tumors, the extracted tumors are three-dimensionally displayed and their geometrical irregularities are evaluated. The irregularity of the surface shape of the tumor is quantified using S / V ratio, which is the ratio of the cube of the surface area of the extracted tumor to the square of the volume, as a parameter. In addition, the S / V ratio for both benign and malignant
Using the (predetermined threshold value), the two groups are classified (step P21). Specifically, the malignant tumor (cancer) 16 obtained by extracting the tumor shape determination parameter γ by clinical application
Example, the tumor shape determination parameter γ was set to S / V ratio ≒ 4 (= predetermined threshold) based on both the evaluation results applied to the results of 11 cases of benign tumors (fibroadenoma) and benign or malignant. (Step P21).

For this purpose, by executing the tumor irregularity irregularity quantification step (step P10), the tumor shape determination parameter γ (γ = (γ = ( S ³ / V ² ) /
κ) is calculated. If the surface of the tumor is smooth, the parameter γ for determining the shape of the tumor will be small, and if the surface has irregularities, the parameter γ for determining the shape of the tumor will have a large value.

Therefore, as shown in FIG. 13, by executing the shape judging step (step P21) following the tumor irregularity quantifying step (step P10), the calculated tumor shape judging parameter γ becomes a predetermined ( S / V rat
io) When it is less than the threshold value, it is determined that the surface of the tumor is smooth (step P23), while the threshold value (S / V rat) is determined.
io) When it is equal to or more than that, it is determined that the surface of the tumor has irregularities (step P22).

As shown in FIG. 13, by executing the shape determination step (Step P21), a tumor determined to have a smooth surface is diagnosed as a benign tumor (fibroadenoma), and the surface is determined to have irregularities. Is diagnosed as a malignant tumor (cancer).

As described above, the three-dimensional display of the tumor surface in the breast cancer screening system 10 is performed by the shading method called surface rendering, and it is possible to easily observe subtle changes in the three-dimensional shape of the surface. become able to.

[0238] Tumor unevenness quantification step (step P1
0) executes a tumor volume calculation step (step P12) of calculating the tumor volume as the sum of voxels constituting the finally extracted tumor, as shown in FIG.

According to the calculated tumor volume, a tumor shape determination parameter γ (γ = (S ³ /
V ² ) / κ) is calculated (step P13). Following the tumor irregularity quantification step (Step P10), the above-described shape determination step (Step P21) is executed, and the smoothness of the tumor surface and the nature of the irregularities are determined according to the calculated tumor shape determination parameter γ ( Step P21). Following the shape determination step (Step P21), a tumor diagnosis step (Step P20) is executed to determine whether the tumor is benign or malignant according to the surface determination result (Step P21 → P2).
2, or P21 → P23).

The tumor volume calculation step (step P12)
Since calculations are performed using voxels as basic units, a breast cancer screening system 10 that can be automatically executed at high speed with computer resources such as a compact and inexpensive PC and that is effective as an initial diagnosis of breast tumors and a mass screening can be constructed.

In addition, the irregularity quantification step (step P
10) is a first step (step P) as shown in FIG.
111) and a second step (step P112) as a basic configuration are performed in a tumor surface area calculation step (step P11).

Tumor Surface Area Calculation Step (Step P11)
Is a step of creating a combination of three adjacent voxels with respect to the voxel labeled as the outline of the tumor, and using the sum of the areas of the triangles formed by the three combinations as the surface area.

That is, by providing a tumor volume calculation step (Step P12) in the tumor irregularity quantification step (Step P10), the tumor volume is calculated as the sum of voxels constituting the finally extracted tumor. In addition, by providing a tumor surface area calculation step (Step P11) in the tumor irregularity quantification step (Step P10), a combination of three adjacent voxels is created for a voxel labeled as a contour of a tumor. .
By executing the second step (Step P112) following the first step (Step P111), the total sum of the areas of the triangles formed by the combination of three adjacent voxels is calculated as the surface area of the tumor (Step P1).
1). In addition, according to the volume of the tumor, a tumor shape determination parameter γ using the surface area and volume of the tumor is calculated as an index for measuring the degree of unevenness of the tumor surface shape (step P13). Subsequent to the tumor irregularity quantification step (step P10), the above-mentioned shape determination step (step P2)
1) is executed, and the smoothness and unevenness of the surface of the tumor are determined according to the calculated tumor shape determination parameter γ (step P21). Steps P22 and P23 are executed according to the result of the shape determination step (step P21), and it is diagnosed whether the tumor is benign or malignant according to the surface determination result.

In such a tumor surface area calculation step (step P11) including the first step (step P111) and the second step (step P112), since the calculation is performed using voxels as basic units, a compact and inexpensive PC can be used. Breast cancer screening system 1 that can be automatically executed at high speed with such computer resources, and is effective as an initial diagnosis and mass screening of breast tumors
0 can be constructed.

Tumor unevenness quantification step (step P1)
0) follows the tumor diagnosis step (step P20) in FIG.
As shown in the figure, tumor irregularity quantification step (step P1)
If the calculated tumor shape determination parameter γ calculated in step (0) is less than a predetermined threshold (<in step P21), it is determined that the surface of the tumor is smooth. It has a logical configuration that determines that the surface has unevenness.

In the tumor diagnosis step (step P20), a tumor whose surface is determined to be smooth is diagnosed as a benign tumor (fibroadenoma) in accordance with the result of the shape determination step (step P21) (step P23). Is diagnosed as a malignant tumor (cancer) (step P
22).

In the automatic malignant tumor identification processing of the present embodiment,
In order to assist the diagnosis of breast tumors, the extracted tumors were three-dimensionally displayed in surface shape and their geometrical irregularities were evaluated. The three-dimensional display of the tumor surface was performed by a shadow method called surface rendering, and it was possible to easily observe a subtle uneven change in the three-dimensional shape of the surface. In addition, irregularities in the surface shape of the tumor were quantified using the tumor shape determination parameter γ of the ratio (S / Vratio) of the cube of the surface area of the extracted tumor to the square of the volume (step P10).

The evaluation results obtained by applying the tumor shape determination parameter γ to the results of 16 cases of malignant tumor (cancer) and 11 cases of benign tumor (fibroadenoma) extracted by this clinical application, The shape determination parameter γ is S
It is shown below that the two groups can be grouped (step P21) at / V ratio ≒ 4.

In order to verify the effectiveness of the automatic malignant tumor identification processing of this embodiment, 16 cases of malignancy shown in FIG. 14 (case (1) to case (1
6)) and 11 benign cases (case (17) to case (27) in the figure) were used (determination of malignant / benign is described in the “symptom” column in the figure). As shown in FIG. 14, the age structure of the subject (“age” in the figure) is 25 to 77 years
Age. In addition, the size of the tumor (“size [c
m] ") is 0.5 cm to 1.8 cm.

Although the results of application to the clinical data shown in FIG. 14 were all successfully extracted, 5 cases of malignant and 5 cases of benign were shown in FIGS. 15 to 19 and FIG. 24 to FIG.

FIGS. 15A to 24A are images in which the tomographic image of one section on the zx plane of the voxel data 16a and the boundary of the extracted tumor are displayed in a superimposed manner. It can be seen that a better tumor region is detected from each of FIGS. 15A to 24A.

Each of FIGS. 15 to 24 is a three-dimensional display image of a tumor surface shape by a shading technique called surface rendering.

From FIGS. 15 (b) to 19 (b), irregularities on the surface of the tumor, which are characteristic of malignant tumors (cancer), are well recognized, and infiltration around the cancer is well observed.

Further, it can be seen from FIGS. 20 to 24 that the smooth surface of the tumor, which is a characteristic of a benign tumor (fibroadenoma), is clearly extracted. As described above, the region of the mammary gland tumor is favorably detected by the automatic malignant tumor identification processing of the present embodiment, and the benign and malignant geometric features can be easily observed from the three-dimensional display image of the tumor surface. In other words, the discriminating diagnosis of tumor benign and malignant can be performed more objectively.

As described above, if the surface of the tumor is smooth, the parameter γ for determining the shape of the tumor becomes small, and if the surface has irregularities, the parameter γ for determining the shape of the tumor has a large value. FIG. 25 shows the calculation result obtained by applying Expression (3-1) to the tumor extracted in the above-described tumor extraction processing.

Also, the calculated tumor shape determination parameter γ
FIG. 26 shows the relationship between the tumor shape determination parameter γ and the obtained tumor volume. According to this, the tumor shape determination parameter γ has significantly different values (malignant tumor shape determination parameter γ ≧ 4, benign tumor shape determination parameter γ <4) for benign (fibroadenoma) and malignant (cancer). The benign and malignant are grouped according to the tumor shape determination parameter γ ≒ 4 (step P21). Thus, it can be seen that quantitative determination of tumor benign or malignant has been made by the present tumor shape determination parameter γ.

In summary, by performing the tumor irregularity irregularity quantification step (step P10), the tumor shape determination parameter γ using the surface area and volume of the tumor as an index for measuring the degree of irregularity of the tumor surface shape is calculated. (Step P13). Subsequent to the tumor irregularity quantification step (step P10), normalization is performed in the case of a sphere (step P1).
4) Parameter S / V rat of ratio of surface area to volume
Note that io shows higher values for malignant tumors (cancers) than for benign tumors (fibroadenomas).
0), the calculated tumor shape determination parameter γ is smaller than the predetermined threshold (Step P21)
In <>, it is determined that the surface of the tumor is smooth, while when it is equal to or larger than the threshold (≧ in step P21), it is determined that the surface of the tumor has irregularities (step P2).
1). By executing the tumor diagnosis step (Step P20), a tumor whose surface is determined to be smooth is diagnosed as a benign tumor (fibroadenoma) (Step P23), and a tumor whose surface is determined to have unevenness is determined. Is diagnosed as a malignant tumor (cancer) (step P22).

As a result, a region of a mammary gland tumor having irregularities on the tumor surface, which is a feature of a malignant tumor (breast cancer), can be detected well, and a benign or malignant geometric feature can be easily detected from a three-dimensional display image of the tumor surface. This makes it possible to more objectively make a diagnosis of benign or malignant tumor.

That is, it is possible to three-dimensionally display a tumor region obtained by an automatic breast tumor extraction system (pathological diagnosis support system) and quantitatively measure and evaluate the geometrical irregularities on the surface. become able to. The three-dimensional display of the tumor surface in the breast cancer screening system 10 is as follows:
It uses surface rendering, which can give the surface of the tumor a shading distribution similar to the shadows of light that occur when we look at an object. As a result, it becomes possible to observe subtle changes in the unevenness of the tumor surface shape, and it becomes possible for the examiner and the patient to understand and judge based on common recognition. As a result, the extracted tumor is converted into a three-dimensional image, and the geometrical irregularities of the tumor surface shape are quantified (step P).
10) By doing so, it becomes possible to support breast breast cancer diagnosis.

The calculation of the tumor shape determination parameter γ can be automatically performed at high speed with a computer resource such as a compact and inexpensive PC, and a breast cancer screening system 10 effective as an initial diagnosis of breast tumor and a mass screening can be constructed.

[0261]

According to the first aspect of the present invention, in order to quantify the irregularity of the surface shape, a parameter using the ratio of the surface area to the volume of the extracted tumor is defined. For this reason, by performing the unevenness quantification step, the surface area of the solid as an index for measuring the degree of unevenness of the three-dimensional surface shape,
Shape determination parameter using volume (γ = (S ³ / V ² )
/ Κ) is calculated. Subsequently, by performing the shape determination step, it is determined that the three-dimensional surface is smooth when the shape determination parameter (γ) calculated in the unevenness quantification step is less than a predetermined threshold, while the shape determination parameter (γ) is greater than or equal to the threshold. In some cases, it is determined that the surface of the solid has irregularities.

The three-dimensional region obtained as a result can be displayed three-dimensionally, and the geometric unevenness of the surface can be quantitatively measured and evaluated. The three-dimensional representation of a solid surface uses surface rendering that allows the shading of the solid surface to be similar to the shadows of light that occur when we look at an object. Thereby, it becomes possible to observe a subtle unevenness of the three-dimensional surface shape.

The calculation of the three-dimensional shape determination parameter γ can be automatically executed at high speed with a computer resource such as a compact and inexpensive PC, and is an effective means for three-dimensional shape diagnosis.

For this reason, according to the second aspect of the present invention, in addition to the effect of the first aspect, in the unevenness quantification step, first, a constant normalization step is performed to make the three-dimensional shape spherical. After normalizing κ such that the value of the shape determination parameter becomes 1 at the time of, the shape determination parameter (γ = (S ³ /) using the surface area and volume of the solid as an index for measuring the degree of unevenness of the solid surface shape V ² ) / κ) is calculated. Subsequently, the above-described shape determination step is performed.

As described above, the three-dimensional display of the three-dimensional surface is performed by the shading method called surface rendering, and it is possible to easily observe a subtle unevenness of the three-dimensional shape of the surface.

According to the third aspect of the present invention, the first aspect
In addition to the effects described in the above, the volume calculation step in the unevenness quantification step calculates the volume of the solid as the sum of the finally extracted voxels of the solid.

In such a volume calculation step, since a calculation is performed using voxels as basic units, a compact and inexpensive PC is used.
It can be automatically executed at high speed with computer resources such as the above, and is an effective means for initial diagnosis of breast solids and mass screening.

According to the invention set forth in claim 4, according to claim 1
In addition to the effects described in the above, in the uneven irregularity quantification step,
First, by performing a constant normalization step, κ is normalized so that the value of the shape determination parameter becomes 1 when the solid shape is a sphere, and then the first step is executed to label the solid as a solid contour. 3 for the voxel
One voxel combination is created, and the second step is executed following the first step, and the sum of the areas of the triangles formed by the combination of three adjacent voxels is defined as the surface area of the solid. Based on this, a shape determination parameter (γ = (S ³ / V ² ) / κ) using the surface area and volume of the solid as an index for measuring the degree of unevenness of the solid surface shape is calculated. Subsequently, the above-described shape determination step is performed.

The surface area calculation step including the first step and the second step can be automatically executed at high speed with a computer resource such as a compact and inexpensive PC because the calculation is performed using voxels as basic units. This is an effective means for initial diagnosis of three-dimensional objects and mass screening.

[0270] The greatest feature of breast tumor diagnosis is that it involves differentiation between benign and malignant tumors. Among them, the discrimination and diagnosis of small cancers having an outer diameter of 1 cm or less is the most important. Benign tumors are "regular and smooth", such as circular or elliptical in contour.

On the other hand, a malignant tumor (cancer) has an irregular and irregular shape such as a crab shape or a star shape. In normal ultrasonography, such geometrical features in tomographic images are often used, but if the tumor is still small, observing the tomographic image alone will determine whether the tumor is malignant or benign. Sometimes it is difficult to know what is.

Therefore, the invention described in claim 5 utilizes the fact that the surface shape of a malignant tumor (cancer) is larger than that of a benign tumor having the same volume because the surface shape is more irregular than that of a benign tumor. Therefore, in order to quantify the irregularity of the surface shape, a parameter using the ratio of the surface area to the volume of the extracted tumor is defined. For this purpose, by performing a tumor irregularity irregularity quantification step, a tumor shape determination parameter using the surface area and volume of the tumor is calculated as an index for measuring the degree of irregularity of the tumor surface shape. Subsequent to this tumor irregularity quantification step, the parameter S / V ratio of the surface area to volume ratio normalized in the case of a sphere is determined by the malignant tumor (cancer)
Paying attention to showing a higher value than a benign tumor, by performing the shape determination step, when the calculated tumor shape determination parameter is less than a predetermined threshold, while determining that the surface of the tumor is smooth, In this case, it is determined that the surface of the tumor has irregularities. By performing a tumor diagnosis step following the shape determination step, a tumor determined to have a smooth surface in the shape determination step is diagnosed as a benign tumor, and a tumor determined to have a surface having irregularities is identified as a malignant tumor. Diagnose a tumor (cancer).

As a result, a region of a mammary gland tumor having irregularities on the surface of a tumor, which is a feature of a malignant tumor (cancer), can be detected well, and a benign or malignant geometric feature can be easily detected from a three-dimensional display image of the tumor surface. This makes it possible to more objectively make a diagnosis of benign or malignant tumor.

That is, it is possible to three-dimensionally display a tumor region obtained by an automatic breast tumor extraction system (pathological diagnosis support system) and quantitatively measure and evaluate the geometrical irregularities on the surface. become able to. The three-dimensional representation of the tumor surface uses surface rendering that allows the shading of the tumor surface to be similar to the light shading that occurs when we look at an object. As a result, it becomes possible to observe subtle changes in the unevenness of the tumor surface shape, and it becomes possible for the examiner and the patient to understand and judge based on common recognition. As a result, breast cancer diagnosis can be supported by making the extracted tumor into a three-dimensional image and quantifying the geometric unevenness of the tumor surface shape.

The calculation of the tumor shape determination parameter can be automatically performed at high speed with a computer resource such as a compact and inexpensive PC, and is an effective means for initial diagnosis and mass screening of breast tumors.

In the invention according to claim 6, in order to assist the diagnosis of a breast tumor, the surface shape of the extracted tumor is three-dimensionally displayed, and its geometric unevenness is evaluated. The irregularity of the surface shape of the tumor is quantified using a parameter of the ratio S / V ratio of the cube of the surface area of the extracted tumor to the square of the volume of the tumor. In addition, the S / V ratio (predetermined threshold) is used to classify both good and malignant groups.

Therefore, according to the invention of claim 6, in addition to the effect of claim 5, by performing the tumor irregularity irregularity quantification step, the tumor is used as an index for measuring the degree of irregularity of the surface shape of the tumor. A tumor shape determination parameter (γ = (S ³ / V ² ) / κ) using the surface area and volume of the target is calculated.
If the surface of the tumor is smooth, γ will be small, and if the surface has irregularities, γ will have a large value. Therefore, by performing the shape determination step following the tumor irregularity quantification step, it is determined that the surface of the tumor is smooth when the calculated tumor shape determination parameter is less than a predetermined (S / V ratio). On the other hand, the threshold (S / V rate
o) If it is equal to or more than the above, it is determined that the surface of the tumor has irregularities. By performing a tumor diagnosis step following the shape determination step, a tumor determined to have a smooth surface in the shape determination step is diagnosed as a benign tumor, and a tumor determined to have a surface having irregularities is identified as a malignant tumor. Tumor (cancer)
Diagnose.

As described above, the three-dimensional display of the tumor surface is performed by the shadow method called surface rendering, and it is possible to easily observe a subtle change in the three-dimensional shape of the surface.

According to the invention of claim 7, claim 5
In addition to the effects described in the above, since the tumor volume calculation step performs the calculation using voxels as basic units, it can be automatically executed at high speed with a computer resource such as a compact and inexpensive PC, and performs the initial diagnosis of breast tumor, This is an effective means for mass screening.

According to the invention described in claim 8, claim 5
In addition to the effects described in the above, the tumor surface area calculation step including the first step and the second step performs calculation using voxels as basic units, so that it is automatically performed at high speed with computer resources such as a compact and inexpensive PC. It can be implemented and is an effective tool for initial diagnosis and mass screening of breast tumors.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a basic configuration of a breast cancer screening system of the present invention.

FIG. 2 Fasttrack tracking system 3
It is a dimensional space coordinate system.

FIG. 3 is a flowchart illustrating an embodiment of a three-dimensional image acquisition process performed by the breast cancer screening system of FIG.

FIG. 4 is a block diagram for explaining an algorithm of a tumor extraction process executed in the breast cancer screening system of FIG. 1;

FIG. 5 is a conceptual diagram of a distance between a luminance center and a geometric center of the luminance and a geometric center.

FIG. 6 is a flowchart showing a procedure for creating a membership function.

FIG. 7 is a graph of a Rician function.

FIG. 8 shows a histogram of average brightness values of voxels belonging to “tumor”, “normal tissue”, and “boundary” and an approximated membership function.

FIG. 9 shows a histogram of a distance between the center of gravity of a luminance and a geometric center in a feature amount and an approximated membership function.

FIG. 10 shows a histogram of luminance variance in a feature amount and an approximated membership function.

FIG. 11 is a graph of a Rayleigh distribution function.

FIG. 12 is a second-stage processing flow for extracting a tumor region by fuzzy inference using a membership function.

FIG. 13 is a processing flow of a malignant tumor automatic identification process of the present embodiment.

FIG. 14 shows various cases for verifying the effectiveness of the automatic malignant tumor identification processing of the present embodiment.

FIG. 15A shows an image in which a tomographic image of one section on the zx plane of voxel data and a boundary of an extracted tumor are displayed in a superimposed manner in the first case shown in FIG. FIG. 4B is a three-dimensional display image of the tumor surface shape by a shadow technique called surface rendering.

FIG. 16A is an image in which a tomographic image of one section on the zx plane of voxel data and a boundary of an extracted tumor are displayed in a superimposed manner in the fourth case shown in FIG. FIG. 4B is a three-dimensional display image of the tumor surface shape by a shadow technique called surface rendering.

17 (a) is an image in which a tomographic image of one section on the zx plane of voxel data and a boundary of an extracted tumor are displayed in a superimposed manner in the sixth case shown in FIG. FIG. 4B is a three-dimensional display image of the tumor surface shape by a shadow technique called surface rendering.

FIG. 18 (a) is an image in which the tomographic image of one section on the zx plane of the voxel data and the extracted tumor boundary are superimposed and displayed in the eighth case shown in FIG. FIG. 4B is a three-dimensional display image of the tumor surface shape by a shadow technique called surface rendering.

FIG. 19A is an image in which a tomographic image of one section on the zx plane of voxel data and a boundary of the extracted tumor are displayed in a superimposed manner in the tenth case shown in FIG. FIG. 4B is a three-dimensional display image of the tumor surface shape by a shadow technique called surface rendering.

FIG. 20 (a) is an image in which a tomographic image of one section on the zx plane of voxel data and a boundary of an extracted tumor are displayed in a superimposed manner in the nineteenth case shown in FIG. FIG. 4B is a three-dimensional display image of the tumor surface shape by a shadow technique called surface rendering.

FIG. 21A is an image in which a tomographic image of one section on the zx plane of voxel data and a boundary of an extracted tumor are displayed in a superposition manner in the twentieth case shown in FIG. FIG. 4B is a three-dimensional display image of the tumor surface shape by a shadow technique called surface rendering.

FIG. 22 (a) is an image in which a tomographic image of one section on the zx plane of voxel data and a boundary of an extracted tumor are displayed in a superimposed manner in the twenty-fifth case shown in FIG. FIG. 4B is a three-dimensional display image of the tumor surface shape by a shadow technique called surface rendering.

FIG. 23 (a) is an image in which a tomographic image of one section on the zx plane of voxel data and a boundary of the extracted tumor are displayed in a superposition on the 26th case shown in FIG. 14, FIG. 4B is a three-dimensional display image of the tumor surface shape by a shadow technique called surface rendering.

FIG. 24 (a) is an image in which a tomographic image of one section on the zx plane of voxel data and a boundary of an extracted tumor are displayed in a superimposed manner in the twenty-seventh case shown in FIG. FIG. 4B is a three-dimensional display image of the tumor surface shape by a shadow technique called surface rendering.

FIG. 25 shows a calculation result obtained by applying equation (3-1) to a tumor extracted in the tumor extraction processing.

FIG. 26 shows the relationship between the calculated tumor shape determination parameter and the calculated tumor volume.

FIG. 27 is a block diagram showing a basic configuration of the first related art.

FIG. 28 is a block diagram showing a basic configuration of a second conventional technique.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 ... Breast cancer screening system 12 ... Probe 121 ... Ultrasonic probe 12a ... Probe data 121a ... Ultrasonic tomographic image data 13 ... 3D position sensor 131 ... AC magnetic field position sensor 13a ... Probe coordinate data 14 ... Tracking means 15 ... 3 Three-dimensional image data 15b Three-dimensional ultrasonic image data 16 Three-dimensional voxel data generating means 16a Voxel data 30 Measurement object 301 Mammary gland PC Personal computer

 ────────────────────────────────────────────────── ─── Continuing on the front page (72) Inventor Koichi Ito 3311-1 Yakushiji, Minamikawachi-cho, Kawachi-gun, Tochigi Pref. JC01 JC06 JC08 JC11 JC20 KK17 KK24 5B057 AA07 BA03 BA05 CA02 CA08 CA13 CB02 CB08 CB13 CC01 CH01 DB03 DB05 DB09 DC02 DC04 DC09

Claims (8)

[Claims] 1. A three-dimensional surface shape quantification method for determining the degree of three-dimensional surface irregularities based on voxel data, wherein three-dimensional surface area (S) and volume are used as indices for measuring the three-dimensional surface shape irregularities. (V) a shape determination parameter (γ) that calculates a shape determination parameter (γ); and if the calculated shape determination parameter (γ) is less than a predetermined threshold, the surface of the solid is determined to be smooth. A shape determining step of determining that the surface of the three-dimensional object exhibits the unevenness property when the threshold value is equal to or greater than the threshold value, wherein the shape determining parameter is γ = (S ³ / V ² ) / κ, A three-dimensional surface shape quantification method characterized by being defined by κ = constant. 2. The method according to claim 1, wherein the irregularity quantification step includes a constant normalization step of normalizing the κ such that the value of the shape determination parameter becomes 1 when the three-dimensional shape is a sphere. The method for quantifying a three-dimensional surface shape according to claim 1. 3. The solid according to claim 1, wherein the irregularity quantification step includes a volume calculating step of calculating a volume of the solid as a sum of voxels of the solid finally extracted. Surface shape quantification method. 4. The step of quantifying the irregularities in the irregularities includes the steps of:
A first step of forming a combination of three adjacent voxels, and a second step following the first step, wherein the total area of the triangles formed by the combination of the three adjacent voxels is defined as the surface area of the solid. The three-dimensional surface shape quantification method according to claim 1, further comprising a surface area calculation step including: 5. An automatic malignant tumor identification method for determining the degree of unevenness of a tumor surface shape based on three-dimensional voxel data, wherein the surface area (S) of the tumor is used as an index for measuring the degree of unevenness of the tumor surface shape; A tumor irregularity quantification step of calculating a tumor shape determination parameter (γ) using a volume (V); and if the calculated tumor shape determination parameter (γ) is less than a predetermined threshold, the surface of the tumor is smooth. Is determined, the shape determination step to determine that the surface of the tumor presents the property of irregularities when the threshold is equal to or more than, the tumor determined that the surface is smooth is diagnosed as a benign tumor, A tumor diagnosis step of diagnosing, as a malignant tumor, a tumor determined that the surface has irregularities, wherein the tumor shape determination parameter is γ = (S ³ / V ² ) / κ, κ = constant Defined by Malignancy automatic identification method, characterized in that is. 6. The tumor irregularity quantification step includes a constant normalization step of normalizing the κ such that the value of the tumor shape determination parameter becomes 1 when the shape of the tumor is a sphere. The method for automatically identifying a malignant tumor according to claim 5, wherein 7. The method according to claim 5, wherein the step of quantifying the irregularity of the tumor irregularities includes a step of calculating a volume of the tumor as a sum of voxels constituting the finally extracted tumor. 3. The method for automatically identifying a malignant tumor according to item 1. 8. The step of quantifying the irregularity of the tumor irregularities, wherein the voxel labeled as the contour of the tumor is:
A first step of forming a combination of three adjacent voxels, and a second step following the first step, wherein a surface area of the tumor is a sum of triangular areas formed by the combination of the three adjacent voxels. 6. The method for automatically identifying a malignant tumor according to claim 5, comprising a tumor surface area calculating step including: