JP2895414B2 - Ultrasonic volume calculator - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an ultrasonic volume calculator for calculating the volume of a living tissue based on three-dimensional information of the living tissue obtained by transmitting and receiving ultrasonic waves.

[0002]

2. Description of the Related Art A plurality of scanning planes are formed by three-dimensionally scanning an ultrasonic beam, and the three-dimensional echo data captured thereby is analyzed to obtain a three-dimensional image of a living tissue. Diagnostic devices have been proposed. In recent years, there has been proposed an ultrasonic volume computing device that calculates a volume of a specific part in a living body based on three-dimensional echo data, for example, for observing a developmental state of a fetus or a temporal change in the size of a tumor. .

[0003] As the ultrasonic volume computing device, there have been proposed the following, for example.

(1) The measurer traces the contour of the measurement object in each tomographic image of the three-dimensional image, and synthesizes the contour traced for each tomographic image over the entire tomographic image, thereby obtaining a three-dimensional image of the measurement object. A device that extracts an image and determines its volume.

(2) Echo level between the measurement object and the surrounding biological tissue (image density when image is displayed)
Focusing on the difference, the threshold of the echo level is determined and each echo data is binarized based on the threshold to separate the measurement object and the background, and the measurement object thus extracted A device for determining the volume.

[0006]

In the apparatus (1), since the measurer himself discriminates the object to be measured from other tissues and traces the outline thereof, the accuracy of extracting the object to be measured is high. Therefore, the required volume is also highly accurate. However, in ultrasound images, the boundaries of the tissue are not always clear, and accurate tracing requires considerable experience. Performing contour tracing on a number of tomographic images constituting a three-dimensional image is a task requiring time and labor. As described above, the above-described device (1) has a problem that a burden on a measurer is large. Also,
In this device, there is also a problem that the reproducibility of a required volume value is not good because the trace results are different even for the same measurement object when the measurement person is different.

In order to solve such a problem,
As a modified example of the above-mentioned device (1), a device has been proposed which is used when obtaining the volume of a part such as a fetus head whose cross-sectional shape can be regarded as substantially elliptical. With this device, the measurer specifies only the major and minor axes of the cross section of the measurement object in each tomographic image, and based on the instructions, the device automatically forms an approximate figure of the measurement object, and the volume is Ask. According to this device, the problems of the time and the burden on the measurer as in the device (1) described above are solved, but the problem that the accuracy of the volume value is deteriorated occurs.

On the other hand, the apparatus of the above (2) has advantages that the processing speed is high and there is no problem of reproducibility of the volume value due to the difference of the measurers since there is no processing performed by the measurer such as contour tracing. . However, in this apparatus, since the measurement target is extracted based only on the image density (echo level), if the image density is the same even in a part different from the measurement target, it is calculated as the measurement target. There is a problem that the accuracy of the volume value obtained is not good.
In addition, in this apparatus, in order to accurately determine the volume, it is necessary to use a threshold value that appropriately separates the measurement object and the background, but since such a threshold value differs depending on the image, an appropriate value corresponding to each image is required. It was difficult to set a threshold.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems, and automatically extracts a measurement target from three-dimensional echo data at high speed and with high accuracy, thereby reducing the volume of the measurement target. It is an object of the present invention to provide an ultrasonic volume computing device that can be obtained at high speed and with high accuracy.

[0010]

In order to achieve the above-mentioned object, an apparatus according to the present invention uses an ultrasonic data transmission / reception system to transmit and receive echo data for each voxel contained in a three-dimensional region in a living body. A three-dimensional echo data acquisition unit that acquires the data, a binarization processing unit that performs a binarization process on each of the echo data, and, based on the binarization processing result, each voxel included in the three-dimensional region is designated. A target area extracting unit that performs a connectivity determination operation on the extracted reference voxel, and extracts a voxel group determined to have connectivity to the reference voxel as a target area, and counts voxels included in the extracted target area. A volume calculation unit that calculates the volume of the target area based on the counting result, and has the following configuration .

That is, in the apparatus according to the present invention, a binarization threshold setting change section for sequentially changing the binarization threshold used in the binarization processing section according to a predetermined rule, and wherein the binarization threshold is changed. A change amount calculating unit that calculates a change amount of the volume operation value before and after the setting change, a comparison determining unit that compares the obtained change amount with a predetermined change amount threshold value, and a comparison result indicating the change amount. Is smaller than the change amount threshold,
While instructing the binarization threshold setting change unit to change the binarization threshold setting, instructing the target region extraction unit and the volume calculation unit to perform processing using a new binarization threshold, A volume calculation control unit that determines the volume of the measurement target based on the volume calculation value before the change of the binarization threshold setting when the change amount is equal to or greater than the change amount threshold value.

According to another aspect of the present invention, there is provided an apparatus comprising: a binarization threshold setting change unit for sequentially changing a binarization threshold used in a binarization processing unit; And, based on each of the volume operation values obtained in correspondence with each of the binarization thresholds, a boundary point calculation unit that obtains a boundary point at which the rate of change of the volume operation value with respect to the binarization threshold suddenly changes, A volume determination unit that determines the volume of the measurement target based on the corresponding volume calculation value.

Further, another configuration of the present invention is characterized in that the target area extracting section performs a connectivity judgment operation using a three-dimensional diffusion projection method.

[0014]

According to the present onset bright, the echo data for each voxel of the three-dimensional region of first in vivo acquired by the three-dimensional echo data acquiring unit, binarization processing in the binarization processing unit is performed. Here voxel (voxel = volu
The term “me cell” refers to a small cube serving as an image display unit in a three-dimensional image space, and is a concept corresponding to a pixel in a two-dimensional image. Then, the target area extraction unit receives the binarization processing result, and extracts a target area for volume calculation while checking the connectivity between the voxels. That is,
For example, when the measurer designates any one voxel in the measurement object as a reference voxel, the target region extraction unit performs a connectivity determination operation on each voxel with respect to the reference voxel, and determines that the voxel has connectivity with the reference voxel. The extracted voxel group is extracted as a target area. Then, the volume calculation unit calculates a volume calculation value of the target region based on the number of voxels included in the extracted target region. According to this configuration, since the target region is extracted in view of the connectivity between the voxels, even when a part of the background region has the same value as the measurement target when the binarization process is performed, this is not the case. Are excluded because they have no connectivity with the reference voxel. Therefore, according to this configuration, only the region corresponding to the measurement target can be extracted, and thus the volume of the measurement target can be accurately determined.

[0015] Then, in the case of performing the binarization processing using the binarization threshold, the secondary and the binarization threshold by the value of the threshold setting changing unit sequentially set to change according to a predetermined rule, each of the secondary which is the setting change For each binarization threshold, a binarization processing unit and a target region extraction unit extract a target region, and a volume calculation unit calculates a volume calculation value of the target region. The change amount calculation unit calculates a change amount of the volume operation value before and after the setting change each time the binarization threshold is changed. The obtained change amount is compared with a predetermined change amount threshold by the comparison determination unit. When the change amount is smaller than the change amount threshold in response to the comparison result in the comparison determination unit, the volume calculation control unit instructs the binarization threshold setting change unit to change the setting of the binarization threshold and sets the target area. Instructs the extraction unit and volume calculation unit to perform processing using the new binarization threshold, and when the change amount exceeds the change amount threshold value, measures based on the volume calculation value before the binarization threshold setting change Determine the volume of the object. According to this configuration, it is possible to automatically search for a binarization threshold capable of accurately extracting the measurement target, and to accurately determine the volume of the measurement target based on the threshold.

Further, in another configuration of the present invention, when the binarization processing is performed using the binarization threshold, the binarization threshold is sequentially changed by the binarization threshold setting change unit, and the setting change is performed. For each of the binarization thresholds, a target area is extracted by the binarization processing unit and the target area extraction unit, and a volume calculation value of the target area is calculated by the volume calculation unit. The boundary point calculation unit is configured to calculate, based on each of the binarization thresholds and each of the volume operation values obtained for the binarization thresholds, a boundary point at which the rate of change of the volume operation value with respect to the binarization threshold changes rapidly. Ask for. The volume determination unit determines the volume of the measurement target based on the calculated volume value at the binarization threshold. According to this configuration, it is possible to automatically search for an appropriate binarization threshold for accurately extracting the measurement target, and to accurately determine the volume of the measurement target based on the threshold.

Further, in still another configuration of the present invention, the target region extracting unit performs the connectivity judgment operation using the three-dimensional diffusion projection method, so that the operation time of the connectivity judgment operation can be reduced.

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of an ultrasonic volume calculating apparatus according to the present invention will be described below with reference to the drawings.

First Embodiment FIG. 1 is a block diagram showing the configuration of a first embodiment of the ultrasonic volume calculating apparatus according to the present invention. In FIG. 1, an ultrasonic diagnostic apparatus 10 transmits and receives an ultrasonic wave into and from a living body while sequentially moving a scanning plane in parallel, and acquires echo data for each part in the living body. Here, interpolation processing and the like are further performed on the echo data obtained by scanning the ultrasonic beam, and echo data is obtained for all voxels in the three-dimensional region where the ultrasonic beam is scanned. The obtained echo data of each voxel is stored in the three-dimensional echo data memory unit 12.

The binarization processing section 14 binarizes the echo data of each voxel read from the three-dimensional echo data memory section 12 by comparing it with a predetermined binarization threshold. Here, the echo level of each voxel is binarized. Note that, when an image is displayed, the echo level is expressed as the density of the image (for example, 256 gradations). For the sake of simplicity, the following description uses the term image density instead of the echo level. Binarize image density "
And so on.

For example, in the case where the volume of the stomach of a fetus is determined, since the amniotic fluid is filled in the stomach of the fetus, the image density inside the stomach is considerably lower than the surroundings. Therefore, if an appropriate binarization threshold is set, it is possible to discriminate the inside and surroundings of the stomach from a portion having a higher image density and a portion having a lower image density than the binarization threshold.

However, in this case, amniotic fluid having no difference in image density from the inside of the stomach exists around the fetus, and if this is included in the ultrasonic scanning area, this is used in the binarization processing. Amniotic fluid also has the same value as the stomach. Further, if there is a shadow peculiar to the ultrasonic wave in the scanning area, the shadow portion has the same value as the stomach in the binarization process. As described above, it is difficult to separate and extract only the fetal stomach to be measured only by the binarization processing.

Therefore, in the present embodiment, by subjecting the binarization processing result to the connectivity determination operation by the target area extraction unit 16, the voxels having the same value as a result of the binarization are measured. Only voxels of an object (eg, fetal stomach) are separated and extracted.

The connectivity judgment operation is an operation process for obtaining a connected area in an image, and has connectivity to a specific voxel from a voxel group having the same value by binarization. This is to extract only those that are. That is, one voxel inside the object to be measured is designated as a reference voxel, voxels around the reference voxel are sequentially examined, and only voxels having the same value as the reference voxel and connected to the reference voxel are selected. , Only voxels corresponding to the measurement object are extracted. For example,
Since the body tissue of the fetus intervenes between the amniotic fluid part around the fetus and the inside of the fetal stomach, even if the binarization result values are the same, they are not connected to each other. By the determination operation, the inside of the stomach of the fetus can be separated and extracted from the amniotic fluid part around the fetus.

Accordingly, in the present embodiment, the measurer first looks at the binarized image (display of the binarized processing result as an image) or the original three-dimensional ultrasonic image before binarization. The reference voxel setting unit 20 sets a reference voxel in the measurement object. Then, the target region extracting unit 16 determines the connectivity of each voxel to the reference voxel, and as a result of the connectivity determination, determines the voxel group determined to have connectivity to the reference voxel, corresponding to the measurement object. Is extracted as a target area to be processed.

Examples of the method of extracting the target area by the connectivity judgment operation include, for example, a three-dimensional thickening method, a three-dimensional projection method, and a three-dimensional diffusion projection method.

Hereinafter, these three methods will be described in detail.

The three-dimensional thick method This method is an extension of the thick methods used in the two-dimensional image processing three-dimensionally.

In this method, a search range is set around a reference voxel, and connectivity is checked and extracted for all voxels in the search range. The survey area is set as a cube centered on the reference voxel, and is gradually expanded ("thicker").

That is, at first, only the reference voxel is the investigation range, and this reference voxel is extracted because it is included in the measurement object. Hereinafter, the extracted voxel is referred to as “extracted voxel”.

Next, the investigation range is "thickened", and a 3 × 3 × 3 voxel cube centered on the reference voxel is defined as the investigation range. Then, connectivity is checked for all voxels in the 3 × 3 × 3 search range. The connectivity is examined using the (a) 6 neighborhood mask and the (b) 26 neighborhood mask shown in FIG.

That is, first, one voxel included in the survey range is selected, and this is designated as a voxel of interest. Then, apply the mask to the voxel of interest,
Only when the value of the target voxel (binarization result value) is the extraction target value (binarization result value of the voxel in the measurement object to be extracted) and there is at least one extraction voxel in the mask, It is determined that the target voxel is connected to the measurement object, and the target voxel is extracted. When a 6-neighbor mask is used, the presence / absence of an extracted voxel is checked for six adjacent voxels at the top, bottom, left, right, front and back of the voxel of interest 100. In the case where a 26-neighbor mask is adopted, voxels in the oblique direction are further added to the six-neighbor mask, and 26 voxels obtained by removing the voxel 100 of interest from the cube of 3 × 3 × 3 voxels centered on the voxel 100 of interest. Is used as a mask, and the presence or absence of an extracted voxel in this mask is checked. Then, all the voxels in the survey range are sequentially designated as the target voxel 100, and the same determination / extraction process is performed.

When the voxel of interest is no longer extracted no matter how much the voxel of interest is changed, the investigation range is enlarged one voxel at a time in all directions. Therefore, when the survey range is extended n times, the survey range at that time is (2n
+1) Become a ^3- voxel cube. Then, the same determination / extraction process is repeated for the expanded investigation range, and voxels are extracted.

The determination / extraction process and the expansion of the investigation range are repeated in this manner. When no new voxels are extracted even if the investigation range is expanded, all the regions connected to the reference voxel are extracted. It is determined that the processing has been performed, and the processing ends.

FIG. 3 shows an example of the flow of the target area extraction by the three-dimensional thickening method, and shows how the area to be extracted expands for each step.
Although the three-dimensional thickening method originally targets a three-dimensional area, the basic procedure of the thickening method is the same in two dimensions and three dimensions. Will be explained.

In FIG. 3, the shaded area indicates the area where the binarized data is the extraction target value.
In step 1, first, a reference voxel is set inside a region to be extracted (that is, a region of an extraction target value voxel), and the reference voxel is extracted. In step 2, the cube near one voxel of the reference voxel is set as the investigation range (in FIG. 3, inside the bold line frame), the connectivity determination operation is performed on all the voxels included in this investigation range, and the connected part is extracted as the extracted voxel. I do. Hereinafter, the survey range is sequentially expanded, and the connectivity of all voxels included in the expanded survey range is checked to perform an extraction process (steps 3 and 4). Then, when the number of voxels to be extracted no longer increases even if the investigation range is expanded, the extraction process is terminated (step n). As described above, according to the three-dimensional thickening method, all the voxels connected to the reference voxel and having the extraction target values are extracted.

As shown in FIG. 4, even if there is an area 300 where the binarized data is an extraction target value in addition to the area 200 corresponding to the measurement object, the three-dimensional thickening method does not
Since the extracted region spreads around the reference voxel, the non-extraction target value portion between the region 200 and the region 300 breaks the connectivity between the two, and the region 200 ends without being extracted. Thus, according to the three-dimensional thickening method, only the voxel group corresponding to the measurement target can be extracted as the target region.

The three-dimensional projection method This method is an extension of the projection method which is used in the image processing two-dimensional three-dimensionally.

In this method, for all voxels in the three-dimensional image area, 6 from top to bottom, from bottom to top, from left to right, from right to left, from front to back, and from back to front.
The connectivity is checked for the direction, and extraction processing is performed. That is, this method is, for example, a method of sequentially irradiating a three-dimensional image region with light sequentially from six directions of up, down, left, right, and front, and sequentially extracting a shadow portion of an extracted voxel that has already been extracted. is there.

More specifically, first, for example, the voxel at the upper right front corner of the three-dimensional image area is set as the first voxel of interest, and the voxel of interest is moved from top to bottom to sequentially apply the mask shown in FIG. I will apply it. The mask shown in FIG. 5 is for extracting the voxel immediately before the voxel of interest 100 in the traveling direction of the voxel of interest 100. Only when the value of the voxel of interest is the value to be extracted and the voxel extracted by the mask is an extracted voxel, it is determined that the voxel of interest has connectivity, and the voxel of interest is extracted. This extraction process is performed for all voxels of one line from top to bottom. When the extraction processing for one line is completed, the same processing is performed by changing the line, and this processing is repeated for all the lines in the three-dimensional image area.

When the extraction process from the upper side to the lower side is completed as described above, the same process is sequentially performed in the other five directions. By repeating this, the process ends when there is no more voxel to be newly extracted.

FIG. 6 is a diagram showing the flow of target area extraction by the three-dimensional projection method. Here, as in the case of the three-dimensional thickening method, description will be made using a two-dimensional diagram. In addition,
In FIG. 6, the arrow shown below the figure of each step indicates the traveling direction of the voxel of interest.

In FIG. 6, first, in step 1, a reference voxel is set inside a region to be extracted, and the reference voxel is extracted as an extraction voxel. In step 2, extraction is performed while moving the voxel of interest from top to bottom. As a result, each voxel below the reference voxel is extracted. Less than,
Steps 3, 4, and 5 are performed in this order while sequentially changing the traveling direction of the target voxel from bottom to top, left to right, and right to left. In this manner, when the change in the traveling direction of the target voxel has completed one cycle, the steps 6-1 are performed while sequentially changing the traveling direction of the target voxel again.
The extraction work is performed as shown by 0, and the extraction process ends when there is no more voxel to be newly extracted (step n) regardless of how the traveling direction is changed.

Also in this three-dimensional projection method, only a portion having connectivity to the reference voxel can be extracted as a target region.

3D Diffusion Projection The object to be measured can be reliably extracted by using either the 3D thickening method or the 3D projection method described above. However, in the three-dimensional thickening method, every time the investigation range is expanded, the connectivity determination operation is performed on all voxels within the investigation range. Therefore, when the image area becomes large, the time required for the processing becomes enormous. Further, in the three-dimensional projection method, the number of voxels extracted in the initial stage of the extraction process is small, so that the processing efficiency is low, and the processing time is long.

On the other hand, the three-dimensional diffusion projection method to be described below can extract a target area in a considerably short processing time even when the image area is large.

The three-dimensional diffusion projection method is an improvement of the three-dimensional projection method described above, and is roughly divided into the following two steps.

The first stage does not aim to completely extract the object to be measured, but to extract as many voxels as possible at high speed. The processing of the first stage is similar to the above-described three-dimensional thickening method in that the connectivity is examined using a mask for each voxel of interest within the investigation area while sequentially expanding the investigation area. The major difference between the two is that in the three-dimensional thickening method, the entire cube centered on the reference voxel was set as the investigation range, and all the voxels in this cube were considered as voxels of interest, while the connectivity was examined for each, In the first stage of the original diffusion projection method, only the outermost voxel of the cube centered on the reference voxel is set as the target voxel, and the connectivity determination operation is performed only on these voxels.

That is, in this first stage, one voxel of the outermost shell of the cube centered on the reference voxel is set as the search range, and voxels included in this search range are sequentially designated as the target voxel, and On the other hand, the connectivity is determined by applying the mask shown in FIG. Then, only when the binarization result value of the target voxel 100 is the extraction target value and at least one extracted voxel exists in the mask, it is determined that the target voxel is connected to the measurement target, and the target voxel is determined. Extract voxels.

When the extraction processing is completed for all the voxels within the investigation range, the investigation area is expanded and the same extraction processing is performed. In other words, a cube centered on the reference voxel is expanded one voxel at a time in the up, down, left, and right directions, and extraction processing is performed using the outermost shell of the cube as a survey range.

An example of the flow of the first stage processing is shown in steps 1 to m of FIG. As shown in the figure, first, in step 1, a reference voxel is set and extracted. Next, in step 2, a voxel outside one of the reference voxels is set as a survey range (in FIG. 8, a portion between two squares drawn with a thick line), and connectivity is determined for each voxel within the survey range. Can be examined. Hereinafter, in steps 3 and 4, the survey area is sequentially moved outward.
The connectivity is checked for each voxel within the investigation range and extraction processing is performed. Then, when the number of extracted voxels no longer increases even if the survey range is moved, the processing of the first stage is ended (step m).

As described above, in the first stage, the processing time can be significantly reduced by limiting the investigation range to only the outermost voxels of the cube that is sequentially expanded. However, in the processing of the first stage, since the investigation range is limited to the outermost shell of the cube, complete extraction cannot be expected when the measurement target has a complicated shape. That is, using the example of FIG. 8, the connectivity of the area 400 to which the dot is applied in step 4 of FIG. 8 is checked because it is included in the investigation range in step 4, but is not extracted in step 4. Then, in the subsequent steps, the region 400 is outside the investigation range itself, and is not extracted again. Thus, in the first stage of the three-dimensional diffusion projection method, unlike the three-dimensional thickening method, once an extraction leak occurs, it remains until the end, so it is necessary to completely extract a complex shape. Can not.

Therefore, in the second stage of the three-dimensional diffusion projection method, portions not extracted in the first stage are extracted. The processing in the second stage is exactly the same as the three-dimensional projection method except that the mask shown in FIG. 7 is used. That is, the first
Starting from the extraction result at the stage, the step (m
As shown in +1) and thereafter, the extraction process is performed while sequentially changing the extraction direction (ie, the traveling direction of the voxel of interest), such as from above to below and from below to above. Then, no matter how the traveling direction is changed, the extraction process ends when there are no more voxels to be newly extracted (step n). Since the processing in the second stage is the same as the three-dimensional projection method, the object to be measured can be completely extracted without any omission.

As described above, according to the three-dimensional diffusion projection method,
After many voxels are quickly extracted by the first-stage process, the same process as the three-dimensional projection method is performed in the second stage. Can improve the evil,
As a whole, the speed of the extraction process can be improved.

Note that the three-dimensional diffusion projection method basically includes the above-described two-stage processing. However, when the measurement object has a simple shape such as a sphere, it can be completely extracted only in the first stage. It is possible, and in such a case, the processing may be terminated only in the first stage.

The masks used in the first and second steps are not limited to those shown in FIG. For example, in the first stage, the mask shown in FIG.
At the stage, the mask of FIG. 5 may be used.

In the foregoing, various methods for extracting a target area in the target area extraction unit 16 have been described. When the target region is extracted by the target region extracting unit 16 in this manner, the volume calculation unit 18 counts voxels in the target region and multiplies the count result by the volume of one voxel, for example. Is calculated, and the calculated volume value is output as the volume of the measurement object.

As described above, according to the present embodiment, by performing the connectivity judgment operation on the binarization processing result, only the voxel corresponding to the measurement object can be extracted. The volume can be determined accurately.

The present embodiment is applicable not only to the case where the three-dimensional image is subjected to the binarization processing with respect to the echo level, but also to the case where the three-dimensional image is subjected to the binarization processing by, for example, texture analysis.

Second Embodiment Next, a second embodiment of the present invention will be described. This second
The second embodiment is an improvement of the first embodiment. In the case where binarization processing of each voxel is performed by comparing magnitudes with thresholds, such as binarization of an image density value (echo level), the second embodiment is not applicable. The optimum threshold value (binary threshold value) for the binarization process is automatically determined, and the volume of the measurement target is determined with higher accuracy.

That is, for example, since the image density values of the object to be measured and the background are different for each image, the binarization threshold value cannot be set as a fixed value. Must be set to an appropriate value. As a method of setting the binarization threshold, a method in which the measurer appropriately sets the binarization threshold while looking at the ultrasonic image is also conceivable.However, since the ultrasonic image has a complex image density,
It is generally difficult to directly determine the optimal binarization threshold from this. Also, even if the operator can determine the optimal binarization threshold, if the operator is different, the value of the binarization threshold will be different even for the same image, so this method has a problem in terms of reproducibility. The problem remains.

Therefore, in the second embodiment, an appropriate binarization threshold value is automatically determined by the ultrasonic volume computing device, and the volume of the measurement object is more accurately determined based on the threshold value. Hereinafter, the principle of the present embodiment will be described with reference to FIG. 9 taking an example of obtaining the volume of the stomach of a fetus.

FIG. 9 shows an image density value (echo level) along a certain direction in an ultrasonic image of a stomach of a fetus, for example.
Is shown. In the image density value distribution shown in FIG. 9, since the amniotic fluid is seen inside the stomach of the fetus, the reflection of ultrasonic waves is small, and therefore the image density value is also low. On the other hand, a portion of the body tissue around the stomach (hereinafter referred to as a surrounding tissue) has a higher image density value than that of the stomach. The boundary between the stomach and the surrounding tissue has a high echo level due to a large reflection of the ultrasonic wave due to a difference in acoustic impedance, and therefore has a peak image density value. Usually, in an ultrasonic image, the boundary between different tissues is blurred to some extent, and therefore, the rise of the image density value near the boundary is not vertical but rather gentle.

In such an image density value distribution, the stomach portion can be accurately determined only by appropriately setting a binarization threshold value between the image density value of the fetal stomach and the image density value of the surrounding tissue. The accuracy of the calculated stomach volume is not very good. This is because the rise of the image density value is not vertical in the vicinity of the boundary between the stomach and the surrounding tissue, so that the boundary determined by the binarization threshold deviates from the true boundary between the stomach and the surrounding tissue (in FIG. 9, Inside the true boundary).

Therefore, in such an example, in order to accurately extract the part of the stomach of the fetus from the three-dimensional ultrasonic image,
It is necessary to set the binarization threshold so that the boundary between the stomach and the surrounding tissue determined by the binarization threshold is as close as possible to the true boundary. Therefore, in the present embodiment, as a method of obtaining such an appropriate binarization threshold, the volume calculation value is calculated using the method of the first embodiment for various binarization thresholds while sequentially changing the binarization threshold. From the correlation between the obtained binarization threshold and the volume operation value, a binarization threshold that is a boundary point at which the volume operation value changes abruptly is obtained, and the binarization threshold is set to an appropriate binarization threshold. Is adopted.

The principle of this method will be described with reference to FIG. From the characteristics of the ultrasonic image, it is considered that the true boundary between the stomach of the fetus and the surrounding tissue in the image density value distribution shown in FIG. Here, if the binarization threshold is set to a small value like x in FIG. 9, since the stomach is the part of the image density value smaller than the binarization threshold, only the part considerably smaller than the actual stomach is extracted. Not extracted. Therefore, as the binarization threshold is increased, the portion extracted as the stomach gradually increases and approaches the actual size of the stomach. Here, for example, assuming that the binarization threshold value is y in FIG. 9, there is a portion where the image density value is smaller than y in the surrounding tissue. However, since these two are not connected, only the stomach portion can be separated and extracted by performing the target region extraction by the connectivity determination operation. In this way, only the stomach portion can be separated and extracted by the connectivity determination operation until the binarization threshold value reaches z.

However, in FIG. 9, when the binarization threshold becomes larger, the image density values of the peaks A and B become smaller than the binarization threshold. The values become the same and they are connected to each other, so that even if the connectivity determination operation is performed, only the stomach portion cannot be separated. As a result, the portion of the surrounding tissue, which has been removed by the connectivity determination operation, is added to the extraction region, so that the region extracted as the stomach is much larger than the actual stomach. For this reason, the calculated value of the calculated volume of the stomach greatly changes before and after the binarization threshold value z.

The binarization threshold value z, which is the boundary point of the change in the volume operation value, is located at a position substantially equal to the peak in the image density distribution. The closest region can be extracted.

As described above, in this embodiment, the state of the change of the volume operation value accompanying the change of the binarization threshold value is obtained, and from this, the binarization threshold value which is the boundary point at which the volume operation value changes rapidly is obtained. The volume operation value at the binarization threshold is defined as the volume of the measurement object (here, the fetal stomach). As a result, the target area extracted by the binarization processing and the target area extraction processing can be considerably close to the area of the actual measurement target, and the accuracy of the required volume can be improved.

Next, a specific configuration of the second embodiment will be described.

FIG. 10 is a block diagram showing one configuration example of the second embodiment. 10, the same members as those in FIG. 1 are denoted by the same reference numerals.

In FIG. 10, the configuration from the ultrasonic diagnostic apparatus 10 to the reference voxel setting unit 20 is the same as that of the first embodiment shown in FIG. The volume calculation value is obtained by performing extraction.

In this configuration, the binarization processing unit 14
The echo level of each voxel, that is, the image density value is binarized using a binarization threshold. The binarization threshold used here is sequentially changed by the binarization threshold setting change unit 28 according to a predetermined rule (for example, increment by a predetermined value). That is, in the present embodiment, for the same three-dimensional echo data information stored in the three-dimensional echo data memory unit 12, the binarization threshold is sequentially changed, and the binarization processing is performed for each binarization threshold. The volume calculation value is obtained by performing the region extraction process and the volume calculation value calculation process.

Each time the binarization threshold value is changed, the change amount calculation unit 22 calculates the change amount of the volume calculation value before and after the setting change. That is, the change amount calculation unit 2
2 obtains a difference between the volume operation value obtained for the binarization threshold value before the setting change and the volume operation value obtained for the binarization threshold value after the setting change, and outputs this as a change amount.
The change amount calculation unit 22 can be configured using, for example, a memory for holding the volume operation value before the change of the binarization threshold setting and a subtraction circuit. In this case, the change amount calculation unit 2
When a new volume calculation value is input from the volume calculation unit 18 to the volume calculation unit 2, the difference between the input volume calculation value and the volume calculation value held in the memory is obtained by a subtraction circuit, and the difference is calculated. And thereafter, the content of the memory is updated and the input volume calculation value is held in the memory.

The comparison / determination unit 24 compares the change amount output from the change amount calculation unit 22 with a preset change amount threshold value. Then, the volume calculation control unit 26 controls the entire calculation process based on the comparison result. That is, when the change amount is smaller than the change amount threshold value, a signal for instructing the binarization threshold value setting change unit 28 to change the setting of the binarization threshold value is issued to each processing unit after the binarization processing unit 14. Then, arithmetic processing using the new binarization threshold is performed. And
When the change amount becomes larger than the change amount threshold value, it is determined that the volume calculation value has suddenly changed, and the volume calculation value immediately before this sudden change,
That is, the volume calculation value obtained for the binarization threshold value before the setting change is output as the volume of the measurement target.

FIG. 11 is a flowchart showing the flow of the volume calculation process using the apparatus shown in FIG. 10. Referring to FIG. 11, the volume calculation of the fetal stomach will be described as an example. The procedure will be described.

In the calculation of the volume of the fetal stomach, first, the ultrasound diagnostic apparatus 10 acquires echo data of a three-dimensional region including the measurement object (fetal stomach) and stores it in the three-dimensional echo data memory unit 12 (S500). ). Next, the values used in each processing unit are initialized (S502). That is, for example, initialization of the set value of the binarization threshold in the binarization threshold setting change unit 28 and the volume operation value used when the change amount calculation unit 22 calculates the amount of change is performed. In this example, since the image density of the stomach of the fetus is lower than that of the surrounding tissue, and the image density has a distribution as shown in FIG. 9, the binarization threshold is initially set to a small value.

When the initialization is completed, the binarization processing unit 14
The echo data is read from the three-dimensional echo data memory unit 12, and a binarization process is performed (S504). Then, the target area extraction unit 16 performs a connectivity determination operation on the result of the binarization processing to extract a target area (S50).
6). The volume calculation unit 18 counts voxels included in the extracted target area, and calculates a volume calculation value based on the counting result (S508). The calculated volume operation value is
At the same time as being provided to the change amount calculation unit 22, it is also provided to the volume calculation control unit 26.

The change amount calculating section 22 obtains a difference between the volume calculated value input from the volume calculating section 18 and the volume calculated value before the change of the binarization threshold setting, and outputs this difference as a change amount (S
510). In the first loop, since there is no volume calculation value before the setting change, a change amount from the initial value of the volume calculation value set in S502 is obtained instead. And
The change amount of the calculated volume value before and after the change of the binarization threshold setting thus obtained is compared with a predetermined change amount threshold value (S
512).

As a result of the comparison, if the change amount is smaller than the change amount threshold value, the binarization threshold value setting changing unit 28 increases the binarization threshold value by a predetermined value (S 514). And the processing of S504 to S512 is repeated. Then, the series of processing is repeated while sequentially increasing the binarization threshold by a predetermined amount until the change amount becomes larger than the change amount threshold value in the process of S512.

That is, as can be seen from the image density distribution shown in FIG. 9, when the binarization threshold is gradually increased from a small value, the calculated volume calculation value is also gradually increased, and the stomach in the target area extraction processing result is increased. The volume calculation value increases gradually until the and the surrounding tissue are connected. Therefore, it is considered that only the stomach portion is extracted before the change amount exceeds the change amount threshold value. Moreover, during that time, as the binarization threshold increases, the boundary between the stomach and the surrounding tissue separated by the binarization threshold approaches the true boundary, so that the volume calculation value gradually increases to the true value of the stomach. Approaching the volume value.

When the binarization threshold value becomes larger than the peak value of the image density indicating the boundary between the stomach and the surrounding tissue, the stomach and the surrounding tissue are connected in the target area extraction processing,
As a result, the calculated volume value increases rapidly. The calculated volume value immediately before the rapid increase is the value closest to the true volume of the fetal stomach.

Therefore, in the processing procedure of this embodiment,
In the comparison determination in S512, when the change amount becomes larger than the change amount threshold value, the volume calculation control unit 26 determines that the volume calculation value at the immediately preceding binarization threshold value is the volume of the measurement object (fetal stomach). And outputs it (S516).

In order to further improve the accuracy of the required volume, the following method can be considered. That is, when the change amount becomes larger than the change amount threshold value in the comparison determination in S512, the process returns to the immediately preceding binarization threshold value, and the increase amount of the binarization threshold value in S514 is reduced, and the same processing procedure as described above is performed. It is a method of repeating. According to this method, the binarization threshold value at which the volume operation value changes abruptly can be specified with higher accuracy, so that the accuracy of the obtained volume is improved.

FIG. 12 shows a phantom made by embedding a water-filled balloon (corresponding to a fetal stomach) in agar graphite (corresponding to surrounding tissue), and using the method of this embodiment to produce a balloon. Shows the experimental results when the measurement experiment was performed. In the figure, the horizontal axis represents the value of the binarization threshold, and the vertical axis represents the volume calculation value. In this experiment, the image density was set to 256 gradations (that is, the image density value ranged from 0 to 25).
5), and the initial value of the binarization threshold was set to 5. According to the figure, when the binarization threshold is sequentially increased from the initial value 5,
It can be seen that the volume calculation value increases very slowly and the volume calculation value increases rapidly when the binarization threshold changes from 58 to 59. Therefore, when the binarization threshold value becomes 59, it is considered that the balloon and the agar graphite layer are connected in the target area extraction processing, and the volume calculation value when the immediately preceding image density value 58 is used as the binarization threshold value Is considered to be the value closest to the true volume of the balloon. Actually, the volume operation value when the binarization threshold value is 59 is much larger than the actually measured value of the balloon volume, whereas the volume operation value when the binarization threshold value is 58 is The value was the closest to the measured value of the balloon volume.

As described above, according to the present embodiment, an appropriate binarization threshold can be automatically found, and the volume of the object to be measured can be accurately obtained.

In the above example, the image density of the object to be measured is lower than the image density of the surrounding tissue. However, the present embodiment can be applied to the opposite case. That is, when the image density of the measurement target is higher than the image density of the surrounding tissue, the initial value of the binarization threshold is increased, and if the initial value is sequentially reduced, the volume can be accurately determined by the same principle as in the above example. Can be asked well. In addition, by inverting the image density value of the three-dimensional image, the volume of the measurement target can be obtained in exactly the same manner as in the above-described example.

Further, in the example described above, the binarization threshold is sequentially increased monotonically (or monotonically decreased).
It was determined each time whether the amount of change in the volume operation value before and after the change of the binarization threshold exceeded a certain threshold value, and based on this determination, the boundary point where the volume operation value suddenly changed was determined. The configuration of the embodiment is not limited to this. For example, as a configuration in which a correlation between a binarization threshold and a volume operation value is obtained by previously obtaining a volume operation value for various binarization thresholds, and a boundary point at which the volume operation value rapidly changes from the correlation is analytically obtained. Also, the same effect as in the above example can be obtained. FIG. 13 shows an example of such a configuration.

In the configuration shown in FIG.
The correlation storage unit 30 that stores the binarization threshold value and the corresponding volume operation value in association with each other is provided at a stage subsequent to the step 8. Then, while sequentially changing the binarization threshold by the binarization threshold setting change unit 28, the binarization processing unit 14,
A volume operation value corresponding to each binarization threshold is obtained by the target region extraction unit 16 and the volume operation unit 18 and stored in the correlation storage unit 30 while associating the binarization threshold and the volume operation value with each other. As a result, a table representing the correlation between the binarization threshold value and the volume calculation value is formed in the correlation storage unit 30. The boundary point calculation unit 32 calculates the rate of change of the volume operation value at each binarization threshold based on this table in the correlation storage unit 30, and obtains a boundary point at which the rate of change rapidly changes. Then, the volume determination unit 34 reads out the volume operation value corresponding to the boundary point from the correlation storage unit 30 and outputs it as the volume of the measurement target.

As described above, the correlation between the binarization threshold value and the volume operation value is obtained first, and the boundary point where the rate of change of the volume operation value with respect to the binarization threshold value abruptly changes is obtained from the correlation. It is possible to find a binarization threshold value with which the measurement target can be extracted with high accuracy, and thus the volume of the measurement target can be determined with high accuracy.

[0091]

As described above, according to the present invention,
Since the target area is extracted in view of the connectivity between voxels, even if a part of the background has the same value as the measurement target when performing the binarization processing, such a part is regarded as the reference voxel. It can be excluded because it has no connectivity.
Therefore, according to this configuration, only the region corresponding to the measurement target can be extracted, and thus the volume of the measurement target can be accurately determined.

Further, according to the present invention, when the binarization processing of the echo data is performed using the binarization threshold, the binarization threshold for automatically extracting the object to be measured is automatically searched. The volume of the object to be measured can be determined accurately and with good reproducibility based on this.

According to the present invention, high-speed processing can be performed by using the three-dimensional diffusion projection method in the target area extraction processing.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a first embodiment of an ultrasonic volume calculating apparatus according to the present invention.

FIG. 2 is a diagram illustrating an example of a mask used in the three-dimensional thickening method.

FIG. 3 is an explanatory diagram showing a flow of a target area extraction process when a three-dimensional thickening method is used.

FIG. 4 is a diagram for explaining separation of unconnected regions in the three-dimensional thickening method.

FIG. 5 is a diagram illustrating an example of a mask used in the three-dimensional projection method.

FIG. 6 is an explanatory diagram showing a flow of a target area extraction process when a three-dimensional projection method is used.

FIG. 7 is a diagram illustrating an example of a mask used in the three-dimensional diffusion projection method.

FIG. 8 is an explanatory diagram showing a flow of a target area extraction process when a three-dimensional diffusion projection method is used.

FIG. 9 is a diagram showing an example of an image density value distribution along one direction in an ultrasonic image.

FIG. 10 is a block diagram showing a configuration of a second embodiment of the ultrasonic volume calculating apparatus according to the present invention.

11 is a flowchart showing a volume calculation processing procedure when the ultrasonic volume calculation device shown in FIG. 10 is used.

FIG. 12 is a diagram showing the results of an experiment performed using the device of the second embodiment.

FIG. 13 is a block diagram showing a configuration of a modification of the second embodiment.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 10 Ultrasound diagnostic apparatus, 12 Three-dimensional echo data memory part, 14 Binarization processing part, 16 Target area extraction part, 1
8 volume calculation unit, 20 reference voxel setting unit, 22 change amount calculation unit, 24 comparison determination unit, 26 volume calculation control unit, 28 binarization threshold setting change unit, 30 correlation storage unit,
32 boundary point calculation unit, 34 volume determination unit.

Continuation of front page (58) Field surveyed (Int.Cl. ⁶ , DB name) A61B 8/00-8/15 G06T 5/00-5/50

Claims (4)

(57) [Claims] 1. An ultrasonic volume calculating apparatus for calculating a volume of a measurement object in a three-dimensional region based on echo data of a three-dimensional region in the living body obtained by transmitting and receiving ultrasonic waves to and from a living body. A three-dimensional echo data memory unit that stores echo data for each voxel of the three-dimensional region; a binarization processing unit that performs a binarization process on each of the echo data; For each voxel included in the three-dimensional region, perform a connectivity determination operation on the specified reference voxel, a target region extraction unit that extracts a voxel group determined to have connectivity to the reference voxel as a target region , voxels included in the extracted object area counted, and volume calculating unit for obtaining the volume calculated value of the target area based on the count result, used in the binarization processing unit Follow the binarization threshold to a predetermined rule
Threshold setting change unit for sequentially changing the setting by setting, and each time the setting of the binarization threshold is changed, the setting change
Change amount calculation unit that calculates the change amount of the volume calculation value before and after
And a comparison judgment for comparing the obtained change amount with a predetermined change amount threshold value.
And, as a result of the comparison, the change amount is smaller than the change amount threshold value.
In this case, the binarization threshold setting change unit notifies the binarization threshold
Instruct the user to change the setting, and execute the
Instruct the product operation unit to process using the new binarization threshold
On the other hand, when the change amount is equal to or more than the change amount threshold value
Measured based on the volume calculation value before the binarization threshold setting change
An ultrasonic volume calculation device , comprising: a volume calculation control unit that determines a volume of an object. 2. The method according to claim 1 , wherein the ultrasonic waves are transmitted to and received from a living body.
Based on the echo data of the three-dimensional region in the living body
Ultrasonic volumetric function for calculating the volume of a measurement object in a three-dimensional area
Arithmetic device, which calculates echo data for each voxel in the three-dimensional region.
A three-dimensional echo data memory unit for storing, and a binarizing process for performing a binarizing process on each of the echo data
A processing section, based on the binarization result, each included in the three-dimensional region
For voxels, the series for the specified reference voxel
Perform connectivity determination operation, and perform connectivity to the reference voxel.
Of voxels determined as having a region of interest
And a voxel included in the extracted target area.
A body for calculating a volume operation value of the target area based on the counting result
A product unit, the two-value processing unit for sequential setting change binarization threshold value used in the binarization threshold setting change unit, and the binarization threshold value is sequentially set change, each of these binarization threshold
Binarization based on each volume operation value obtained corresponding to
Find the boundary point where the rate of change of the volume operation value with respect to the threshold changes suddenly
And a volume determination unit that determines the volume of the measurement object based on a volume calculation value corresponding to the boundary point . 3. The ultrasonic volume computing device according to claim 1 , wherein said target area extracting unit is connected using a three-dimensional diffusion projection method.
An ultrasonic volume calculation device that performs gender determination calculation . 4. An ultrasonic wave obtained by transmitting / receiving ultrasonic waves to / from a living body.
Based on the echo data of the three-dimensional region in the living body
An ultrasonic volume calculation device that calculates a volume of a measurement object in a three-dimensional region , wherein echo data for each voxel in the three-dimensional region is calculated.
A three-dimensional echo data memory unit for storing, and a binarizing process for performing a binarizing process on each of the echo data
A processing section, based on the binarization result, each included in the three-dimensional region
For voxels, the series for the specified reference voxel
Perform connectivity determination operation, and perform connectivity to the reference voxel.
Of voxels determined as having a region of interest
And a voxel included in the extracted target area.
A body for calculating a volume operation value of the target area based on the counting result
And a product operation unit , wherein the target region extraction unit is centered on the reference voxel.
The outer shell of the cube is included in the survey area.
Has connectivity from each voxel to the reference voxel.
Extracting the voxels to perform
Iterating while expanding sequentially around the reference voxel
First-stage processing and connectivity to the voxels extracted in the first-stage processing
-Stage extraction of voxels with convergence by three-dimensional projection
And a floor process .