JPH0775605B2 - Ultrasonic image processor - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention relates to an ultrasonic image forming apparatus,
In particular, it relates to boundary extraction of an organization.

[0002]

2. Description of the Related Art In recent years, an ultrasonic diagnostic apparatus for displaying a three-dimensional image has been put into practical use. In this ultrasonic diagnostic device,
The ultrasonic probe transmits and receives ultrasonic waves in a three-dimensional region in the living body. The captured echo data is once stored in the memory by the ultrasonic diagnostic apparatus (or the ultrasonic image processing apparatus), then read and sent to the three-dimensional image processing unit, where an ultrasonic three-dimensional image is formed. It

Here, the ultrasonic three-dimensional image is a three-dimensional representation of an object (tissue) included in the three-dimensional area, which is subjected to shading processing on the surface of the object. A method has also been proposed in which the three-dimensional region is cut along three orthogonal planes to display the tissue as three cross-sectional planes.

As described above, in producing an ultrasonic three-dimensional image, it is necessary to extract the surface (boundary) of the tissue. Conventionally, the following method has been adopted.

The first conventional method is a method of setting a split surface. In this method, one split surface is set in the three-dimensional area, and information before this is cut off.

The second conventional method is a threshold extraction method. In this method, data having an echo level exceeding a threshold for surface extraction is used as the surface of the tissue.

The third conventional method is an artificial tracing method. In this method, a mouse or a light pen is used, and the operator actually traces the contour of the tissue on the image.

[0008]

However, the above-mentioned first method has a problem that an accurate surface cannot be displayed depending on how the split surface is set. FIG. 10 shows a state in which the split surface 100 is set. actually,
The viewpoint is at the top of the figure. (A) is the case where the setting of the cleavage plane is too near, so the image of the fetus to be viewed is obstructed by the pelvis, and (B) the setting of the cleavage plane 100 is too far, so This is the case when it has been cut.
In any case, the image of the fetus cannot be properly captured. Although an unnecessary portion can be removed by setting a plurality of split surfaces, the processing becomes extremely complicated.

The second conventional method described above can be used between tissues having a large difference in echo level, but the placenta-amniotic fluid,
In the case of a combination in which there is not much level difference between the fetus and amniotic fluid, there is a problem that the threshold setting becomes strict and it is difficult to easily detect the boundary.

The above-mentioned third conventional method has a problem that it is not suitable for a large amount of image processing because the processing is complicated due to an artificial method.

The present invention has been made in view of the above conventional problems, and an object thereof is to provide an ultrasonic image processing apparatus capable of performing boundary extraction simply, quickly, and accurately.

[0012]

In order to achieve the above object, the invention according to claim 1 is a three-dimensional data storing echo data captured by transmitting and receiving ultrasonic waves in a three-dimensional region in a living body. An ultrasonic image processing apparatus including: a memory; a boundary extracting unit that extracts a boundary of a tissue based on the echo data; and an image forming unit that forms the ultrasonic image using the extracted boundary. The boundary extracting means includes a first dispersion value calculating means for obtaining a dispersion value for each direction of the echo data for each of a plurality of reference directions that three-dimensionally spread in a certain range while intersecting at the coordinates of interest, It is determined whether or not the coordinate of interest is a boundary point from the second variance value calculating means for calculating the boundary value of the pixel of interest by further calculating the variance value from the direction-specific variance value and the size of the boundary value. Boundary judgment Characterized in that it comprises a stage, a.

Further, in the invention according to claim 2, the input means for inputting the coordinates of the starting point of the boundary extraction processing, and the boundary value of each point along a predetermined line from the starting point are obtained, And a means for setting the processing range on the line based on the coordinates of the boundary point specified on the adjacent line as a reference, and the boundary within the processing range. It is characterized in that a value is obtained and the maximum value of the values is used as a boundary point.

[0014]

According to the structure of the first aspect, a plurality of reference directions passing through the coordinates of interest are set, and the variance value of the echo data is calculated in each reference direction. That is, variations in echo data in each reference direction are required.

Next, the boundary value is calculated by further calculating the variance value for the calculated plurality of variance values. The boundary value serves as an indicator of whether or not the coordinate is a boundary point, and the coordinate with a large boundary value is specified as the boundary point.

According to the above configuration, when the starting point is input, the boundary value is obtained along the line from the starting point, and the coordinate having the maximum boundary value is set as the boundary point. Next, a processing range is set on the line next to the line, and the boundary point is specified from the boundary value of each point within the range.

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS [Calculation of Boundary Value] In the present invention, as shown in FIG. 2, a plurality of (for example, 13) references are passed through a coordinate of interest (voxel of interest) with respect to the reference area 102 shown in FIG. The direction is set and the variance value (representative value) is calculated in each reference direction. Then, the boundary value of the coordinate of interest is calculated by further calculating the variance value of the plurality of variance values.

Here, the boundary value indicates the degree of whether or not the coordinate position of interest is a boundary. FIG. 1 shows a reference area 102 that is a part of the three-dimensional area in the living body.
This reference area is an area used to determine whether or not the center coordinate of interest is the boundary. In the example of FIG. 1, 3 × 3
A region of × 3 is shown, but in an actual device, for example, 1
A 7x17x17 area is used.

As shown in FIG. 2, with respect to the reference region 102, for example, thirteen reference directions that are evenly divided radially are set. For simplification of the drawing, FIG. 2 shows only a part of the reference directions.

A representative value, which is a dispersion value, is calculated for each reference direction by using the echo data (magnitude of echo level) in each reference direction described above. Specifically, the following calculation is executed.

D ₁ = vari {d (xn, yn, zn), ..., d (x, y, z), ..., d (x + n, y + n, z + n)} (1) D ₂ = vari {d (xn, y, zn),…, d (x, y, z),…, d (x + n, y, z + n)}… (2) D ₃ = vari {d ( xn, y + n, zn),…, d (x, y, z),…, d (x + n, yn, z + n)}… (3) D ₄ = vari {d (x, yn, zn),…, d (x, y, z),…, d (x, y + n, z + n)}… (4) D ₅ = vari {d (x, y, zn),…, d (x, y, z), ..., d (x, y, z + n)}… (5) D ₆ = vari {d (x, y + n, zn),…, d (x, y, z ), ..., d (x, yn, z + n)}… (6) D ₇ = vari {d (x + n, yn, zn),…, d (x, y, z),…, d ( xn, y + n, z + n)} (7) D ₈ = vari {d (x + n, y, zn),…, d (x, y, z),…, d (xn, y, z + n)}… (8) D ₉ = vari {d (x + n, y + n, zn),…, d (x, y, z),…, d (xn, yn, z + n) }… (9) D ₁₀ = vari {d (xn, yn, z),…, d (x, y, z),…, d (x + n, y + n, z)}… (10) D ₁₁ = vari {d (xn, y, z),…, d (x, y, z),…, d (x + n, y, z)}… (11) D ₁₂ = vari {d (xn, y + n, z),…, d (x, y, z),…, d (x + n, yn, z)}… (12) D ₁₃ = vari {d (x, yn, z),… , d (x, y, z), ..., d (x, y + n, z)} (13) With the above, 13 representative values D _{1 to} D ₁₃ are calculated.
Note that vari {} is an operator for obtaining the variance value within {}. d (x, y, z) is the echo data level (corresponding to the brightness level) at that coordinate.

Next, the variance value is further calculated for the 13 representative values as follows.

[0023] _{BL = vari (D 1, D} 2, D 3, D 4, D 5, D 6, D 7, D 8, D 9, D 10, D 11, D 12, D 13) ... (14) However, the next BL is the boundary value of the coordinates of interest.

To illustrate the nature of this boundary value, consider the region join in FIG. In FIG. 3, with respect to the point P1 in the area A having a low level and a constant level, the representative value in each reference direction has a small value, and the boundary value, which is the variance value thereof, also has a small value. For the point P2 in the area B having various levels, the representative value in each reference direction becomes a large value,
The boundary values are small because their variance values are high side by side.

On the other hand, for the point P13 on the boundary,
The representative value is small for the reference direction parallel to the boundary, while the representative value is large for the other reference directions. Therefore, the boundary value becomes large.

In short, due to the property of double dispersion, the boundary value becomes maximum when the coordinate of interest exists in two regions, and if the boundary value is sequentially obtained along a certain line, the maximum value is obtained. Coordinates with boundary values can be specified as boundary points.

That is, the present invention uses the property that the variance value becomes the largest when the high level and the low level exist with the same probability, and the variance value operation is used twice. .

In the present invention, the boundary value of the connected boundary point is larger than that of the isolated boundary point. This is because at the isolated boundary points, the representative values of the 13 directions other than one direction take almost the same value, and the variance value (boundary value) of the representative values does not become so large. For this reason,
The detected boundary points become continuous boundary surfaces. The above is
It can be said that this is an ideal feature in extracting the surface of a tissue.

[Automatic Extraction of Boundary Line (Surface)] The coordinates of each boundary point can be easily grasped by the above principle. However, it becomes complicated to perform the calculation for obtaining the boundary value for all the coordinate points in the three-dimensional area.

On the other hand, since the boundary has continuity,
It is obvious that the next boundary point exists near one boundary point.

Therefore, in the present invention, a boundary line (face) extraction method as shown in FIG. 4 is adopted. In FIG. 4, two tissue boundaries 104 are shown. In addition, each line L
In the present embodiment, indicates the line of the ultrasonic beam.

First, by the operator, in the vicinity of the tissue to be viewed,
The starting point S is input. The input is performed by specifying the three-dimensional coordinates, and the work of the operator is basically finished.

When there is an input, the line L1 on which the starting point S is placed
The boundary values are sequentially calculated on (or the closest line). This is usually done over the entire range of the line, but may be limited to a certain range.

After that, the maximum boundary value among all the boundary values on the line L1 is determined. Therefore, the coordinate having the maximum boundary value is determined as the boundary point.

Next, the processing shifts to the adjacent line L2. In this line L2, a fixed range (processing range) is set uniformly in both directions along the line, with the center position being the same as the boundary value on the line L1 (same depth in the beam direction). Then, the boundary value of each point is calculated within the processing range, and the boundary point is specified from the maximum boundary value.

Boundary plane extraction is finally realized by successively repeating such a process for each line.

However, it is possible that the boundary becomes parallel to the line. In that case, each line may be rotated by 90 degrees and the same calculation may be performed. Note that if the viewpoint is on the probe side, such processing is basically unnecessary.

As described above, quick and simple boundary extraction can be realized. It should be noted that FIG. 5 shows the states before (A) and after (B) processing.

[Structure of Ultrasonic Image Processing Device] FIG.
The configuration of an ultrasonic image processing apparatus 10 according to the present invention is shown. This apparatus 10 forms an ultrasonic image based on the echo data sent from the ultrasonic diagnostic apparatus 12. Here, the ultrasonic diagnostic apparatus 12 captures three-dimensional echo data by transmitting and receiving ultrasonic waves in a three-dimensional region in the living body.

The echo data is converted into a digital signal by the A / D converter 14 and then stored in the three-dimensional data memory 16. In storing this memory, the address of the memory is associated with the coordinates of the three-dimensional region in the living body. A boundary extraction unit 20 is connected to the bus 18. The boundary extracting unit 20 performs the above-described boundary value calculation and the boundary extraction shown in FIG. 4, and will be described in detail later.

A processed data storage memory 22 for writing echo data after the boundary extraction processing is connected to the bus 18. An image forming unit 24 that forms an ultrasonic image (three-dimensional image) based on the echo data read from the memory 22 is connected to the bus 18.

The frame buffer memory 26 temporarily stores the formed ultrasonic image. The image data read from the frame buffer memory 26 is CR.
It is sent to T28 and displayed.

The central control unit 30 performs overall control of each component, and is connected to the external input device 32. The external input device 32 is provided with, for example, three dials X, Y, Z, and the operator can use the dials to set the above-mentioned starting point.

FIG. 7 shows how to set the starting point. The diagram shown in FIG. 7 is actually displayed as an image. In-vivo three-dimensional region 106 shown in (D)
The three cutting planes X, Y, and Z orthogonal to each other are arbitrarily set. That is, the external input device 3 shown in FIG.
A dial of 2 sets these cross sections.

(A) is an image of the Z section viewed from above, and two wires are shown to show the relationship between the three cut surfaces. Similarly, in (B), the state of the cut surface Y viewed from the front is shown, and in (C), the state of the cut surface X viewed from the side is shown.

Therefore, the operator can easily adjust the dial while monitoring the image as described above to easily set the starting point at a desired position. Note that FIG.
The present applicant has previously proposed the display form shown in FIG.

Next, the operation of the boundary extraction unit 20 and the like shown in FIG. 6 will be described. In FIG. 8, in S101,
As described above, the external input device 22 is used to set the coordinates of the starting point. This is sent to the boundary extraction unit 20 via the central control unit 30 shown in FIG.

Next, in S102, the boundary extracting section 20 detects a boundary point serving as a reference. That is, FIG.
A boundary point serving as the odor reference is defined on L1 shown in FIG. Specifically, the processing is shown in FIG. 9, and in S201, it is determined whether or not the calculation of the boundary values for all the points is completed. If NO, the target point to be calculated is 1 Then, the variance in each of the reference directions described above is calculated in S203. That is, the above (R1 to R1
The calculation of 3) is performed. Then, in S204, the boundary value is calculated based on the obtained 13 representative values. That is, the above equation (14) is executed. By repeating these steps for each point on the line, the maximum boundary value is finally determined in S205, and the coordinate having the maximum value is regarded as the boundary point.

Returning to FIG. 8 again, in S103, it is judged whether or not the extraction of the boundary points has been completed for all the necessary lines. If not completed, the line is moved to the next one in S104. . Then, in S105, the processing range on the line is set based on the boundary point obtained on the previous line.

Then, in S106, the process shown in FIG. 9 is performed again, so that the boundary point is detected in the line.

Then, the boundary surface is finally extracted by repeating the above steps for all the lines.

When the boundary surface is obtained in this way, all the data between the start point and the boundary surface is deleted, and the data is stored in the processed data storage memory 22 shown in FIG. .

Then, the image forming section 24 forms an ultrasonic three-dimensional image by performing shading on the surface of the in-vivo tissue according to the distance using such data. This is finally displayed on the CRT 28.

[0054]

As described above, according to the ultrasonic diagnostic apparatus of the present invention, the boundary between tissues can be extracted quickly and accurately. Therefore, it is possible to shorten the image processing time and to form an accurate ultrasonic three-dimensional image.

[Brief description of drawings]

FIG. 1 is an explanatory diagram showing a concept of a reference area 102.

FIG. 2 is an explanatory diagram showing 13 reference directions.

FIG. 3 is an explanatory diagram showing joining of two regions.

FIG. 4 is a principle explanatory diagram showing automatic boundary extraction.

FIG. 5 is an explanatory diagram showing states before and after processing of boundary extraction.

FIG. 6 is a block diagram showing a configuration of an ultrasonic image processing apparatus according to the present invention.

FIG. 7 is an explanatory diagram showing an example of a display image by three orthogonal tomographic planes.

FIG. 8 is a flowchart for explaining the operation of the three-dimensional image processing device according to the present invention.

FIG. 9 is a flowchart for explaining the operation of the 3D image processing apparatus according to the present invention.

FIG. 10 is an explanatory diagram showing surface extraction by a conventional split surface.

[Explanation of symbols]

 10 Ultrasonic Image Processing Device 12 Ultrasonic Diagnostic Device 16 Three-Dimensional Data Memory 20 Boundary Extraction Unit 24 Image Forming Unit 32 External Input Device

Claims (2)

[Claims] 1. A three-dimensional data memory for storing echo data captured by transmitting and receiving ultrasonic waves in a three-dimensional region in a living body, and a boundary extracting means for extracting a tissue boundary based on the echo data. An ultrasonic image processing apparatus comprising: an image forming unit that forms the ultrasonic image by using the extracted boundary, wherein the boundary extracting unit is a plurality of three-dimensionally spread while intersecting at a coordinate of interest. A first variance value computing means for finding a direction-wise variance value of the echo data for each reference direction, and a second variance value calculating means for further calculating a variance value from the plurality of direction-wise variance values, second The ultrasonic image processing apparatus, further comprising: a variance value calculating unit and a boundary determining unit that determines whether or not the coordinate of interest is a boundary point based on the size of the boundary value. 2. The ultrasonic diagnostic apparatus according to claim 1, wherein the input means for inputting the coordinates of the start point of the boundary extraction processing, and the boundary value of each point along a predetermined line from the start point Obtained, the means of using the maximum value as the boundary point, and the coordinates of the boundary point specified on the adjacent line as a reference,
An ultrasonic image processing apparatus comprising: a unit for setting a processing range on the line; and determining a boundary point within the processing range.