JPH07184877A - Method for image processing and apparatus therefor - Google Patents

Abstract

PURPOSE:To automate detection of a point of intersection of a magnetic mark in an MR image with the magnetic mark and a contour of an object in the image by moving a plurality of discrete points to converge them so as to minimize the sum total of elastic energy, image energy and outer energy. CONSTITUTION:When an image as an object for detecting a magnetic mark is two dimensional data, data for an MR image with the magnetic mark as an object is inputted in an image data inputting part 1 and the initial values of the positional coordinates of a plurality of discrete points are set in an initial value setting part 2. Then, a plurality or the discrete points are moved in a discrete point moving part 3 in such a way that the total energy expressed as the sum of elastic energy, image energy and outer energy is made minimum. Successively, in a convergence judging part 4, whether the condition of convergence could be satisfied is judged on every movement of the discrete points and when the condition of convergence is not satisfied, the discrete points are moved again by the discrete point moving part 3 and the result is displayed on a display from an outputting part 5.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing method and apparatus for detecting a magnetic marker portion in a two-dimensional or three-dimensional MR image with a magnetic marker.

[0002]

2. Description of the Related Art In recent years, due to an increase in the number of fatalities due to ischemic heart disease and the like, analysis of the movement of the heart wall by images has come to be performed for the diagnosis of these diseases.

The MRI apparatus is capable of noninvasively obtaining in-vivo information as an image and, unlike X-ray CT, can directly obtain an image of an arbitrary cross section or a three-dimensional image. Moreover,
In recent years, improvements in imaging technology have made it possible to obtain moving images in a realistic time.

When obtaining motion information of a tissue in a living body using a moving image, it has been a problem how to make correspondence between positions of images. In the case of MRI, there is a method of putting a magnetic marker in the image as a method of solving it.

The magnetic marker is one of MRI imaging techniques, and a linear or lattice marker is inserted in an image by selectively changing the phase of an output signal with respect to a certain tissue section. It is a thing. Since the magnetic marker moves with the movement of the tissue, it is possible to make correspondence between images of different time phases, which is effective for motion analysis of the dynamic part such as the heart. FIG. 2 is an example of a two-dimensional magnetically labeled MR image of the heart. Reference numeral 6 is the heart wall of the left ventricle, and 7 is a magnetic marker radially inserted.

Conventionally, a motion analysis using a magnetically labeled MR image is performed by using, for example, a heart shape model (AAY
oung, et.al: “Three-dimensional Motion and Deforma
tionof the Heart Wall: Estimation with Spatial Mod
ulation of Magnetization- A Model-based Approach
, ”, Radiology, vol.185, pp.241-247, 1992), but the detection of the intersection of magnetic markers and the contour of the heart wall is done manually. Therefore, the analysis takes time and labor. There is a point.

[0007]

SUMMARY OF THE INVENTION As described above, conventionally,
Since the detection of the magnetic marker is performed manually, it takes time and labor, and there is a problem that the error is large.

Therefore, the present invention is directed to the intersection of the magnetic marker in the MR image with magnetic marker and the contour of the object in the image, or
There is provided an image processing method and an apparatus thereof for the purpose of automating the detection of the intersection of a plurality of magnetic markers in the image.

[0009]

An image processing method according to a first aspect of the present invention is directed to an intersection of a magnetic marker in an MR image with magnetic markers and a contour of an object, or an intersection of a plurality of magnetic markers in the image. Of the MR image with a magnetic label, a plurality of discrete points are arranged in the image, the elastic energy of the plurality of discrete points is obtained, and the plurality of discrete points have The image energy determined by the characteristics of the magnetic marker portion and the characteristics of the portion other than the magnetic marker is obtained, and the plurality of discrete points are moved to converge so that the sum of at least the elastic energy and the image energy is minimized. Then, the position of the intersection is detected by the position of the converged discrete point.

The image processing method of the second invention is a method for detecting an intersection of a magnetic marker in an MR image with magnetic markers and a contour of an object, or an intersection of a plurality of magnetic markers in the image. , Inputting an MR image with magnetic markers, arranging a plurality of discrete points in the image, obtaining elastic energies possessed by the plurality of discrete points, and determining some of the plurality of discrete points The image energy consisting of the energy determined by the characteristics of the magnetic marker portion and the characteristics of the portion other than the magnetic marker and the energy determined by the characteristics of the portion other than the magnetic marker possessed by other discrete points is obtained, and at least the elasticity The plurality of discrete points are moved until they converge so that the total sum of energy and the image energy is minimized, and the position of the intersection is detected by the position of the converged discrete points. That.

[0011]

[Operation] The first invention relates to a method for detecting an intersection of a magnetic marker in an MR image with magnetic markers and a contour of an object in an image processing method, or an intersection of a plurality of magnetic markers in the image. explain.

An MR image with magnetic markers is input.

A plurality of discrete points are arranged in the input image.

Elastic energy possessed by the plurality of discrete points is obtained.

Image energies determined by the characteristics of the magnetically labeled portions of all the discrete points and the characteristics of the portions other than the magnetically labeled portions are obtained.

The plurality of discrete points are moved until they converge so that at least the sum of the elastic energy and the image energy is minimized.

The positions of the intersections are detected by the positions of the converged discrete points. As the discrete points to be used, all of the plurality of discrete points arranged in the image may be used, but only some of the discrete points are used. May be used.

In the image processing method of the second invention, in place of the image energy in the first invention, the characteristics of the magnetic marker part of some of the discrete points among the plurality of discrete points and the features other than the magnetic markers. The image energy is used, which consists of the energy determined by the features of the part and the energy determined by the features of the part other than the magnetic label of the other discrete points.

[0019]

EXAMPLES First Example Hereinafter, a first example of the present invention will be described with reference to FIGS.

[When the target image for magnetic label detection is two-dimensional data] FIG. 1 is a block diagram and a flow chart of this embodiment. Using this figure, from a two-dimensional MR image of the heart containing radial magnetic labels as shown in Figure 2,
A case of finding the intersection between the magnetic marker and the contour of the heart wall will be described.

(1) Image data input unit 1 In the image data input unit 1, a magnetic tag M with a target is attached.
Input the data of R image.

(2) Initial value setting unit 2 The initial value setting unit 2 sets initial values of the position coordinates of a plurality of discrete points. An example of the initial value is shown in FIG.

(3) Discrete Point Moving Unit 3 In the discrete point moving unit 3, the plurality of discrete points are moved so that the total energy represented by E _all in [Equation 1] is minimized.

[0024]

[Equation 1] i is the number of a plurality of discrete points shown in FIG. w _tag
(I) and w _ext (i) are weighting factors.

E _int (i) is elastic energy,
In this embodiment, the following equation is used.

[0026]

[Equation 2] However, v _i (x _i , y _i ) is the position coordinates of the discrete points.

E _tag (i) is the image energy that represents the characteristics of the magnetic label, and the sum of the image density and the weight of energy that represents the edge strength of the image is used. That is,

[Equation 3] Becomes

However, I _image (v _i ) is the image density, and E _edge (i) is the energy representing the edge strength.

Here, the above equation is used as the image energy representing the characteristic of the magnetic marker because the image density of the magnetic marker portion changes sharply as compared with other portions in the image, and This is because the edge strength is high at the intersection.

As E _edge (i), in this embodiment,

[Equation 4]

Is used.

E _ext (i) is external energy added as needed. In the present embodiment, [Equation 5] is energy that moves a discrete point away from a certain point v ₀ or approaches v ₀ . Whether to move away or to move closer depends on whether the coefficient k is positive or negative.

[0033]

[Equation 5]

For example, a variational method is used as a solution for finding the positions of discrete points that minimize the total energy. According to the variational method, in order to minimize [Equation 1], simultaneous equations such as [Equation 6] must be established.

[0035]

[Equation 6] However, the weight of each term is a constant,

[Equation 7] And

Now, considering only the x-coordinate, the simultaneous equations of [Equation 6] can be expressed by the following expression in vector notation.

[0037]

[Equation 8] However,

[Equation 9] Becomes However, a = β, b = −α−4β, c = 2α + 6
β, f _i = −f (v _i ).

This simultaneous equation is solved by the successive approximation method using the Jacobi method. A is divided into a diagonal matrix D and the other part F as in the following equation.

[0039]

[Equation 10] If the nth approximate solution is x ⁿ , the n + 1th approximate solution x
^{n + 1} is

[Equation 11] Becomes However, γ is a convergence speed parameter.

An example of the structure of the discrete point moving unit 3 is shown in FIG.

In the image energy calculation unit 9 and the external energy calculation unit 10, E _tag in [ _Equation 1] is calculated.
(I), E _ext (i) are calculated.

The parameter storage unit 12 stores the weighting coefficient of each energy, the values of the elements of the matrices D and F in [Equation 11], γ, and the like.

The next coordinate calculation unit 11 calculates the next coordinate of each discrete point using [Equation 11] and moves the discrete point.

(4) Convergence determination unit 4 The convergence determination unit 4 determines whether or not the convergence condition is satisfied each time the discrete point is moved. If the convergence condition is not satisfied, the discrete point moving unit 3 again returns the discrete point. Move
Repeat until the convergence condition is satisfied.

As the convergence condition, in the present embodiment, [Equation 1]
It is assumed that the energy change amount ΔE _all ^{n in the} n-th repetition becomes smaller than a certain fixed amount as represented by [2].

[0046]

[Equation 12] FIG. 6 shows a configuration example of the convergence determination unit 4.

In the energy calculation unit 13, [Equation 1]
Calculate E _all in. The value of the nth time E _all ⁿ
Is stored in the storage unit 14.

Then, the energy comparison unit 15 obtains the absolute value ΔE _all ⁿ of the difference from the ^n−1-th value E _all ⁿ⁻¹ stored in the storage unit 14 in advance, and this is a constant amount ε. If it is smaller, it is determined that it has converged, and if it is larger, it is determined that it has not converged. Here, the convergence state that satisfies the convergence condition is
All discrete points gather at the intersection of the magnetic marker and the contour of the heart wall. Therefore, the coordinates of this intersection can be easily obtained from the coordinates after the movement of each discrete point.

(5) Output unit 5 The above result is output from the output unit 5. As a method of outputting the result, there is a method of recording it in a file as data or displaying it on a display. In this embodiment, it is displayed on the display.

[When the Target Image for Magnetic Label Detection is Three-Dimensional Data] Next, the case where the target image for magnetic label detection is three-dimensional data consisting of voxel data or a plurality of slice data will be described.

When the three-dimensional data is targeted, FIG.
Instead of such discrete points, for example, discrete points are arranged on a grid as shown in FIG. 7, and closed contours are used as shown in FIG. In this case, the lattice spacing need not be constant.

For the total energy, the following equation similar to [Equation 1] is used.

[0053]

[Equation 13] w _tag (i, j) and w _ext (i, j) are weighting factors, i,
j represents the coordinates of each discrete point on the grid as shown in FIG.

The following equation is used as the elastic energy E _int (i, j).

[0055]

[Equation 14] However, v _{i, j} = (x _{i, j} , y _{i, j} , z _{i, j} ). As the energy E _tag (i, j) representing the characteristic of the magnetic label, the following equation similar to [Equation 3] is used.

[0056]

[Equation 15] I _image (v _{i, j} ) is the image density and E
_edge (i, j) is energy representing edge strength.

The following equation is used as E _edge (i, j).

[0058]

[Equation 16] Similarly to [Equation 5], the following equation is used as E _ext (i, j).

[0059]

[Equation 17] As in the case of the two-dimensional case, except that [Equation 13] is used instead of [Equation 1] as the total energy, magnetic points are detected by moving the discrete points so as to minimize the total energy. To do.

[0060] The second implementation example Next, a second embodiment of the present invention will be described with reference to FIG. 9-13.

[When the Target Image for Magnetic Label Detection is Two-Dimensional Data] FIG. 9 shows the contour of the target in the image extracted in advance from the image targeted for magnetic label detection or an image similar to the target image. FIG. 3 is a diagram that also serves as a block diagram and a flow chart when the result is used for detection of a magnetic marker. Based on this figure, the MR with a two-dimensional magnetic marker of the heart as shown in
A case where the intersection between the magnetic marker and the contour of the heart wall is obtained from the image will be described.

As the image similar to the target image for magnetic label detection, either a moving image in which the magnetic label is not present in the same phase as the target image, or an image in which the magnetic label is not present in a time phase close to the target image. To use.

(1) Outline Extracting Unit 16 First, the outline extracting unit 16 extracts the outline of the object. FIG. 10 shows a configuration example in the case of extraction by energy minimization similar to the magnetic label detection in the first embodiment.

First, the image data input section 1 inputs an image for contour extraction, and the next initial value setting section 2 sets initial coordinate values of a plurality of discrete points.

Then, the discrete point moving section 3 carries out a process of moving the discrete points so that the total energy becomes the same as in the case of the magnetic marker detection in the first embodiment.

However, E _all of the following equation is used as the total energy.

[0067]

[Equation 18]

However, w _edge (i) and w _ext (i) are weighting factors. E _int (i) is elastic energy, and [Equation 2] is used in this embodiment. E _ext (i) is external energy added as necessary, and in this embodiment, [Equation 5] is used. E _edge (i) is the image energy,
In this embodiment, the energy representing the edge strength of [Equation 4] is used.

(2) Discrete Point Arrangement Unit 17 After extracting the contour, the discrete point arrangement unit 17 arranges a plurality of discrete points on the extracted contour. However, when contour extraction is performed by energy minimization as described above, it is not necessary to arrange new discrete points because discrete points have already been arranged. I do. Specifically, when adjacent discrete points overlap, one of them is removed, and when the distance between adjacent discrete points is too large, interpolation is performed.

(3) Magnetic Mark Corresponding Point Determining Section 18 Next, the magnetic marker corresponding point determining section 18 determines the magnetic marker corresponding points from the arranged discrete points. The arranged discrete points are on the contour line, but if some of these discrete points are made to correspond to the magnetic markers and the other points are made to correspond to the contour line, then these plural discrete points become It is possible to detect the magnetic marker portion without deviating from the contour.
Such points corresponding to the magnetic markers will be referred to as magnetic marker corresponding points.

As a method of determining the magnetic mark corresponding points, FIG.
Since the image shown schematically in Fig. 6 is the main image, the image density of the magnetic marker part 7 is low. Therefore, the vicinity of each point is searched along the contour line, and if there are many places where the image density is low, it is handled as a magnetic marker. The method of making points is adopted. It should be noted that the low image density means a place where the brightness is dark on the image. With this method, when the discrete points after the contour extraction are as shown in FIG. 11, the magnetic marker corresponding points 20 shown by x in FIG. 12 can be determined. The magnetic marker corresponding point 20 is located in the vicinity of the intersection of the desired magnetic marker and the contour of the heart wall, but it is not an accurate position, and therefore the magnetic marker detection unit 1 is further provided.
Determine the exact position at 9.

(4) Magnetic marker detecting section 19 After the magnetic marker corresponding points are determined, the magnetic marker detecting section 19
Accurately detects the intersection of the magnetic marker and the contour of the heart wall.

FIG. 13 shows an example of the structure of the magnetic marker detecting section.

This is the same as the processing after the discrete point moving unit 3 in FIG.

First, if the target image for magnetic label detection is not the same as the contour-extracted image, the same image data is input to the image data input unit 1. In the case of the same image data, the image data input by the contour extraction unit 16 is used.

The discrete point moving section 3 carries out a process of moving the discrete points. However, as the image energy, [Equation 1] is specially given to the magnetic label corresponding point, and [Equation 18] is used to the other points. That is, the total energy is expressed by the following equation.

[0077]

[Formula 19]

However, w _tag (i), w _ext (i), w
(I), w _edge (i), and w ′ _ext (i) are weighting factors. When the convergence determining unit 4 determines that they have converged, the output unit 5 outputs the result. In the converged state where the convergence condition is satisfied, all the magnetic marker corresponding points are accurately gathered at the intersections between the magnetic markers and the contour of the heart wall. Therefore, the coordinates of this intersection can be easily obtained from the coordinates after the movement of each magnetic marker corresponding point.

By using the contour extraction result for magnetic marker detection in this way, it is possible to accurately detect the intersection of the contour line of the object and the magnetic marker in the image.

[When the target image for magnetic label detection is three-dimensional data] Even when the target image for magnetic label detection is three-dimensional data composed of voxel data or slice data of a plurality of slices, on the grid points in FIG. By arranging discrete points in, and instead of using [Equation 18] for the total energy in contour extraction,

[Equation 20] To calculate the total energy for magnetic label detection [Equation 19]
Instead of,

[Equation 21]

May be used. However, E _int (i, j)
As [ _Equation 14], and E _tag (i, j) as [Equation 1]
5], E _ext (i, j) is [Equation 17], E
[Equation 16] is used as _edge (i, j).

[0082] A third implementation example Next, a third embodiment of the present invention will be described with reference to FIG. 14.

[When the target image for magnetic label detection is two-dimensional data] FIG. 14 is a block diagram showing a case in which contours are extracted in advance when motions are detected using images of a plurality of time phases. It is a figure which also served as a flowchart.

The process until the magnetic label detection unit 19 detects the magnetic label in the initial image is the same as that in FIG. 9 of the second embodiment. First, the contour extracting unit 16 extracts a contour.
Next, the discrete point placement unit 17 places discrete points. The magnetic marker corresponding point determination unit 18 determines the magnetic marker corresponding point, and the magnetic marker detection unit 19 detects the magnetic marker.

(1) Motion amount measuring unit 21 Next, in the motion amount measuring unit 21, the difference between the coordinates of each magnetic marker corresponding point after detection of the magnetic marker and the corresponding coordinates of the magnetic marker corresponding point in the previous image is calculated. Calculated as the amount of movement.

If the point sequence of the magnetic marker corresponding points in the l-th image is v _l and the point sequence of the magnetic marker corresponding points in the l + 1-th image is v _{l + 1} , the l-th image to the l + 1-th image The movement amount Δv _l, which is the motion vector of each magnetic marker corresponding point, is represented by [Equation 22].

[0087]

[Equation 22] The calculated movement amount Δv _l is stored in a memory or the like in the storage unit 22.

(2) End determination section 23 In the end determination section 23, it is determined whether or not there is an image, and if there is an image, the procedure returns to the magnetic marker detection section 19 to process the next image. If there are no more images, the output unit 5 outputs the result and the process ends. Here, as an output method, it is possible to display the movement amount Δv _l as a motion vector or to display it numerically.

By repeating these processes, the motion can be quantitatively detected from the images of a plurality of time phases.

[When the target image for magnetic label detection is three-dimensional data] Even when the target image for magnetic label detection is three-dimensional data consisting of voxel data or slice data of a plurality of slices, for example, as shown in FIG. If discrete points are arranged on such a grid and the three-dimensional equations are used as the energy equations as shown in the first and second embodiments, the detection of the magnetic marker and the measurement of the motion amount are similarly performed. be able to.

[0091] implementation of the fourth Next, a fourth embodiment of the present invention will be described with reference to FIG. 15-19.

The present embodiment is an example of detecting an intersection of a plurality of magnetic markers in a two-dimensional MR image with magnetic markers. The intersection between the plurality of magnetic markers is shown at 25 in FIG.

FIG. 16 is both a block diagram and a flow chart for detecting such an intersection between a plurality of magnetic markers.

(1) Image Data Input Unit 1 First, the image data input unit 1 inputs the data of the target MR image with a magnetic marker.

(2) Initial value setting unit 2 Next, the initial value setting unit 2 sets initial values of the position coordinates of a plurality of discrete points. An example of the initial value is shown in FIG.

As a method of setting the initial value, after extracting the heart wall contour by energy minimization as shown in the second embodiment, the magnetic marker corresponding points are determined and the magnetic marker corresponding points are initialized. Used as a value.

(3) Discrete Point Moving Unit 3 Next, the discrete point moving unit 3 moves the discrete points so as to minimize the total energy. As the total energy, the one represented by [Equation 1] is used. However, [ _Equation 2] is used as E _int (i), and [ _Equation 3] is used as E _tag (i). In FIG. 17, in order to detect an intersection outside the initial value, the external energy E _ext (i) is used so that the energy of each discrete point decreases toward the outside. In this embodiment, v ₀ in [Equation 5] is the center of gravity of all discrete points as represented by [Equation 23], and the coefficient k is a negative value.

[0098]

[Equation 23] However, N indicates the number of discrete points.

In this case, the value of k is E which contracts the discrete points.
Adjust so that E _ext (i) to be expanded is larger than _int (i).

(4) Convergence determination unit 4 Then, the convergence determination unit 4 determines convergence.

Since the image density of the magnetic marker portion changes more rapidly than the other portions in the image, the edge strength is large at the boundary portion of the magnetic marker. Further, since the edge strength does not change much along the boundary portion of the magnetic sign, the amount of change in [Equation 3] is small.

Therefore, in [Equation 1], E
_{The term of ext} (i) becomes the dominant state. Therefore, in the discrete point moving unit 3, each discrete point moves outward along the boundary portion of the magnetic marker.

However, when the discrete points reach the intersections of a plurality of magnetic markers, [Equation 3] temporarily becomes low, and the total energy represented by [Equation 1] takes a minimum value. Therefore, it can be converged by the condition that the amount of change in total energy and the amount of movement of discrete points, which are represented by ΔE _all ⁿ in [Equation 12], are smaller than a certain value or the number of repetitions. FIG. 18 shows an example of the movement of discrete points. The direction of the arrow indicates the direction in which the total energy decreases, that is, the outside, compared to the initial value. Due to the above-mentioned convergence condition, the discrete points move a and b along the boundary portion of the magnetic marker, but stop at c when reaching the intersection.

The movement of the discrete points is repeated until the convergence condition is satisfied.

(5) End Judgment Unit 26 When the convergence condition is satisfied, this time the end judgment unit 26 judges whether or not to end.

The determination of the end is made by presetting the number of intersections between a plurality of magnetic markers to be detected and ending when the number is reached. Alternatively, it is assumed that the discrete point ends when it goes out of the image.

By repeating the convergence calculation a plurality of times in this way, it is possible to detect a plurality of intersections. When the end determination unit 26 finishes, the output unit 5 outputs the result.

The present invention is not limited to the examples given here.

In the first embodiment, as the convergence condition in the convergence determination unit 4, the method based on the amount of change in energy is mentioned. However, when the amount of movement of a plurality of discrete points becomes smaller than a certain value, or In consideration of the fact that there may be cases where convergence does not occur under such conditions, it is possible to make the convergence occur by the number of iterations.

In the second embodiment, the contour extraction section 16 shows an example of a method of energy minimization as a contour extraction method, but the result of contour extraction by hand may be used. This also applies to the third embodiment.

In the third embodiment, as a method of detecting the magnetic label in the magnetic label detector 19, only the magnetic label corresponding points are used, that is, the magnetic label corresponding points are attracted to the magnetic label portion as the image energy. There is also a method of using only such energy. In this case, [Equation 1] is used as the total energy.

In the fourth embodiment, when detecting the intersection between a plurality of magnetic markers, the initial value may be manually set.

Further, the convergence condition in the convergence determination unit 4 can be that the amount of change in the edge direction of the image is larger than a certain value. The following equation is an example of the amount of change in the edge direction obtained by the gradient.

[0114]

[Equation 24]

[Equation 25] FIG. 19 shows an example of Δθ ⁿ (i).

The convergence condition is the change amount Δθ ⁿ (i) in the edge direction at all or some of the discrete points.
, When it exceeds a certain value, that is, when the condition of the following equation is satisfied,

[Equation 26] Alternatively, it is possible to use a method as shown in the following formula, in which the average value exceeds a certain value.

[0116]

[Equation 27] However, N indicates the number of discrete points.

[0117]

According to the first and second inventions, the intersection of the magnetic marker in the MR image with magnetic markers and the contour of the object in the image, or the intersection of the plurality of magnetic markers in the image. Can be easily detected, so that the time and labor for the detection can be shortened and the movement of each part of the object in the image can be quantified accurately and easily.

[Brief description of drawings]

FIG. 1 is a configuration diagram of a first embodiment of the present invention.

FIG. 2 is an example (schematic diagram) of a magnetically labeled MR image.

FIG. 3 is an example of initial values.

FIG. 4 is a configuration example of a discrete point moving unit.

FIG. 5 is a diagram (two-dimensional) representing numbers of a plurality of discrete points.

FIG. 6 is a configuration example of a convergence determination unit.

FIG. 7 is a diagram (three-dimensional) showing numbers of a plurality of discrete points.

FIG. 8 is a diagram (three-dimensional) showing an arrangement of a plurality of discrete points.

FIG. 9 is a configuration diagram of a second embodiment of the present invention.

FIG. 10 is a configuration example of a contour extraction unit.

FIG. 11 is an example of a contour extraction result.

FIG. 12 is an example of magnetic marker corresponding points.

FIG. 13 is a configuration example of a magnetic label detection unit.

FIG. 14 is a configuration diagram of a third embodiment of the present invention.

FIG. 15 is an example of an MR image with magnetic labels.

FIG. 16 is a flow chart of processing when detecting an intersection of a plurality of magnetic markers.

FIG. 17 is an example of initial values.

FIG. 18 is an example of movement of discrete points.

FIG. 19 is an example of the amount of change in the edge direction.

[Explanation of symbols]

 1 Image Data Input Section 2 Initial Value Setting Section 3 Discrete Point Moving Section 4 Convergence Judgment Section 5 Output Section 6 Left Ventricle Heart Wall 7 Magnetic Labeling 8 Discrete Points (Part Represented by O) 9 Image Energy Calculation Section 10 External Energy Calculation unit 11 Secondary coordinate calculation unit 12 Parameter storage unit 13 Energy calculation unit 14 Storage unit 15 Energy comparison unit 16 Contour extraction unit 17 Discrete point arrangement unit 18 Magnetic marker corresponding point determination unit 19 Magnetic marker detection unit 20 Magnetic marker corresponding point (× 21) Motion amount measuring unit 22 Storage unit 23 End determination unit 24 Magnetic marker 25 Intersection of a plurality of magnetic markers 26 End determination unit 27 Initial discrete point

─────────────────────────────────────────────────── ───

[Procedure amendment]

[Submission date] January 7, 1994

[Procedure Amendment 1]

[Document name to be corrected] Drawing

[Name of item to be corrected] Figure 2

[Correction method] Change

[Correction content]

[Fig. 2]

[Procedure Amendment 2]

[Document name to be corrected] Drawing

[Name of item to be corrected] Figure 3

[Correction method] Change

[Correction content]

[Figure 3]

[Procedure 3]

[Document name to be corrected] Drawing

[Name of item to be corrected] Fig. 11

[Correction method] Change

[Correction content]

FIG. 11

[Procedure amendment 4]

[Document name to be corrected] Drawing

[Name of item to be corrected] Fig. 12

[Correction method] Change

[Correction content]

[Fig. 12]

[Procedure Amendment 5]

[Document name to be corrected] Drawing

[Correction target item name] Figure 15

[Correction method] Change

[Correction content]

FIG. 15

[Procedure correction 6]

[Document name to be corrected] Drawing

[Name of item to be corrected] Fig. 17

[Correction method] Change

[Correction content]

FIG. 17

[Procedure Amendment 7]

[Document name to be corrected] Drawing

[Name of item to be corrected] Fig. 18

[Correction method] Change

[Correction content]

FIG. 18

[Procedure Amendment 8]

[Document name to be corrected] Drawing

[Name of item to be corrected] Fig. 19

[Correction method] Change

[Correction content]

FIG. 19

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Office reference number FI technical display location 9061-5L G06F 15/70 350 B

Claims (7)

[Claims] 1. A method for detecting an intersection between a magnetic marker and an outline of an object in an MR image with a magnetic marker, or an intersection between a plurality of magnetic markers in the image, wherein an MR image with a magnetic marker is input. Then, a plurality of discrete points are arranged in the image, the elastic energy possessed by the plurality of discrete points is obtained, and the characteristic of the magnetic marker portion and the characteristic of the portion other than the magnetic marker are possessed by the plurality of discrete points. Obtaining the determined image energy, moving the plurality of discrete points until they converge so that at least the sum of the elastic energy and the image energy is minimized, and the position of the intersection is determined by the position of the converged discrete points. An image processing method characterized by detecting. 2. A method of detecting an intersection between a magnetic marker and an outline of an object in an MR image with a magnetic marker, or an intersection between a plurality of magnetic markers in the image, wherein an MR image with a magnetic marker is input. Then, a plurality of discrete points are arranged in the image, the elastic energy possessed by the plurality of discrete points is obtained, and among the plurality of discrete points, a feature of a magnetic marker portion which some of the discrete points have Image energy consisting of the energy determined by the characteristics of the portion other than the magnetic marker and the energy determined by the characteristics of the portion other than the magnetic marker possessed by other discrete points is obtained, and at least the sum of the elastic energy and the image energy The image processing method is characterized in that the plurality of discrete points are moved until they converge so as to minimize, and the position of the intersection is detected by the position of the converged discrete points. 3. An outline of an object in the image is extracted from the MR image or an image similar to the MR image, and a plurality of discrete points are arranged on the extracted outline. The image processing method according to claim 1 or 2, wherein the position of the point in the image is used as an initial value, and the intersection is detected by using a part or all of the plurality of discrete points. 4. An image of a different time phase is detected by detecting an intersection of a magnetic marker in an image and a contour of an object in the image for each different time phase, or an intersection of a plurality of magnetic markers in the image. According to the detection result of the intersection point in
The image processing method according to claim 1 or 2, wherein a motion vector at the intersection is detected. 5. An image energy composed of energy determined by a magnetic marker portion having the discrete points and energy determined by a portion other than the magnetic marker,
3. The image processing method according to claim 1, wherein the sum of the weights of the energy representing the image density and the energy representing the edge strength of the image is used. 6. An apparatus for detecting an intersection of a magnetic marker and an outline of an object in an MR image with a magnetic marker, or an intersection of a plurality of magnetic markers in the image, the MR image with a magnetic marker being detected. Input means for inputting, discrete point arranging means for arranging a plurality of discrete points in the image, elastic energy calculating means for obtaining elastic energy of the plurality of discrete points, and a plurality of discrete points Image energy calculating means for obtaining image energy determined by the characteristics of the magnetic marker portion and the characteristics of portions other than the magnetic marker, and the plurality of discrete points so as to minimize the sum of at least the elastic energy and the image energy. Is moved until it converges, the intersection point detecting means detects the position of the intersection point based on the position of the discrete point converged by the discrete point moving / converging means. The image processing apparatus according to symptoms. 7. An apparatus for detecting an intersection between a magnetic marker and an outline of an object in an MR image with a magnetic marker, or an intersection between a plurality of magnetic markers in the image, wherein the MR image with a magnetic marker is detected. Input means for inputting, discrete point arranging means for arranging a plurality of discrete points in the image, elastic energy calculating means for obtaining elastic energy of the plurality of discrete points, and among the plurality of discrete points , Image energy consisting of the energy determined by the characteristics of the magnetic marker portion having some discrete points and the characteristics of the portion other than the magnetic marker, and the energy determined by the characteristics of the portion other than the magnetic marker having other discrete points Image energy calculating means for obtaining the image energy, and moving until the plurality of discrete points are converged so as to minimize the total sum of the elastic energy and the image energy. A discrete point moving converging means for an image processing apparatus characterized by comprising further an intersection detecting means for detecting the position of the intersection point by the position of the discrete points converged by the discrete points moving the converging means.