JPH0736818B2 - Ultrasound diagnostic image processor - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a digital image processing technique for an ultrasonic image processing apparatus, and in particular, an ultrasonic image of a diseased part is displayed based on an echo signal from a subject and a diseased part image is displayed. Relates to an ultrasonic diagnostic image processing apparatus for edge enhancement.

[Prior Art] In the medical field, an ultrasonic diagnostic imaging apparatus that transmits and receives ultrasonic waves to and from an object such as a living body to perform image diagnosis in the object is well known.

This type of device transmits a pulsed ultrasonic wave output at a constant repetition period into the subject, detects an echo signal reflected from the subject, and by performing this for a predetermined region, The image information for one screen is obtained.

Then, this information is displayed on a CRT display as an image, and a tomographic image of the abdominal organ such as the gallbladder or the heart can be observed.

Image processing in such an ultrasonic diagnostic apparatus includes one in which contour enhancement of an echo signal is performed in an analog manner.

In other words, by applying differential processing to the echo signal after detection and converting this differential signal to the original signal,
The contour information in the echo signal is emphasized, so that the shape and structure inside the subject can be clearly displayed as an image.

However, since such contour enhancement is a fixed process, there is a problem that noise and speckles are also enhanced, and the connection as a two-dimensional image is lacking.

Therefore, the applicant of the present application has proposed and filed an ultrasonic image processing apparatus that solves all the above problems due to analog (Japanese Patent Application No. 63-317918).

In this contour emphasizing method, the contour emphasizing is performed more accurately than the conventional method, and the two-dimensional processing in which information is connected between pixels is performed.

That is, in this contour enhancement processing, when the target pixel is larger than the pixel average value, it is set larger than the target pixel value,
When the pixel of interest is smaller than the pixel average value, it is set smaller than the pixel value of interest.

As a result, the diagnostic image becomes darker in the dark part and lighter in the light part, and the contour is expressed very clearly.In particular, when the variance value is smaller than the reference variance value, the contour enhancement process is performed. Instead, since the pixel value is output as it is, signals such as noise and speckles are not emphasized, and an image can be displayed.

[Problems to be Solved by the Invention] However, in such a contour emphasizing processing method, the region to be contour-enhanced originally is an unclear portion of the contour, and the clear-edge portion should not be emphasized. On the contrary, there was a problem that the part where the contour was clear was emphasized too much.

That is, the reason is that the variance value of the periphery including the pixel of interest is obtained, and when the variance value exceeds only one reference value, it is determined to be the boundary. This is because the target pixel is emphasized in the direction of emphasis.

Therefore, conventionally, there is a disadvantage that the image is overemphasized if the contrast is large, and it may be difficult for an observer to see the affected area image.

As a result, the appropriate image contour enhancement processing has not been performed, and the conventional contour enhancement processing has not been performed with high accuracy.

OBJECT OF THE INVENTION The present invention has been made in view of the above-described conventional problems, and an object thereof is predetermined without emphatically emphasizing a portion having a small variance such as noise or speckle and a portion having a clear contour from the beginning. An object of the present invention is to provide an ultrasonic diagnostic image processing apparatus that performs edge enhancement based on two reference reference variance values in correspondence with the variance values.

[Means for Solving the Problems] To achieve the above object, according to the present invention, a pixel average value calculating means for selecting a pixel group centering on a pixel of interest and calculating a pixel average value of the pixel group. A pixel variance value computing means for computing a variance value σ for the pixel group based on each pixel value of the pixel group and the pixel average value; and the variance value σ.
With two different reference variance values σ ₁ , σ
₂ (however, σ ₁ <σ ₂ ) and a pixel value emphasizing means for performing contour emphasizing processing on the pixel value of the pixel of interest at an emphasizing degree according to the magnitude of the variance value σ,
And the pixel value enhancing means includes the σ ₁ to the σ ₂
It has the characteristic that the degree of emphasis is gradually increased over
_It has a characteristic that the degree of enhancement is gradually weakened in a region larger than ₂ , and excessive enhancement of an ultrasonic image is suppressed when the variance value σ of the pixel group is larger than σ ₂ .

Further, it is characterized in that the contour enhancement processing is performed in parallel and sequentially by a pipeline processing method in real time.

[Operation] With the above-described configuration, according to the present invention, one is a criterion (σ ₁ ) as to whether or not to perform contour enhancement processing that does not enhance noise and speckles, and another is overemphasized. By providing a reference (σ ₂ ) that does not want to emphasize the contour, the variance value σ of the pixel of interest exceeds _one reference variance value σ ₁ and the other reference variance value σ ₂ is also set. If it exceeds, the degree of contour enhancement can be reduced according to the exceeded size.

As a result, it is possible to effectively emphasize only the area where the contour is unclear, and it is possible to perform appropriate contour emphasis processing without over-emphasizing the area where the shade is intense.

Therefore, it is easy for the diagnostician to make a diagnosis because an appropriate affected part image can be obtained by the contour enhancement processing.

[Embodiment] A preferred embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a circuit block diagram showing an embodiment of an ultrasonic diagnostic image processing apparatus according to the present invention.

A characteristic of the present invention is that the variance value of the pixel of interest is obtained in the process of the edge enhancement, and two different reference variance values σ
₁ and σ ₂ are compared and determined, and based on this, optimum contour enhancement is performed on the image display.

That is, in the present embodiment, the variance value of the neighborhood including the pixel of interest is obtained, and the pixel of interest is subjected to predetermined processing according to the magnitude of the variance value.

This process variance sigma ₁ as the two _reference, the sigma ₂ is provided, thereby, if the dispersion value of a target pixel is not processed is smaller than sigma _1, it exceeds the sigma _1, size exceeding of Then, the degree of edge enhancement is adjusted by increasing or decreasing the pixel value of interest.

First, the basic principle of the present invention will be described below.

Assuming that the pixel value of interest is z, and the average value of n * n neighbors including this pixel is the pixel value after the edge enhancement processing is zz, the processing of the present invention is expressed by the following equation.

zz = z + K * (z-) However, K = 0 ... ・ ・ ・ ・ ・ σ <σ1 (1) K = KE * {1- (σ1 / σ)} ... σ1 ≦ σ ≦ σ2 ( 2) K = KE * (σ1 / σ) ... σ2 <σ (3) Further, σ = ^2- () ² where KE = contour enhancement constant (1 to 5), σ ₁ , σ ₂ = Any constant as the standard dispersion value (σ ₁ <σ
₂ ), ² = root mean square value in the vicinity of the target pixel value z.

Here, K in the above equation, zz = z + K * (z-)
And the relationship between σ ₁ and σ ₂ are shown in FIG.

That is, FIG. 2 is a characteristic diagram in which the degree of emphasis on the vertical axis is determined with respect to changes in the dispersion value σ on the horizontal axis (lower limit value σ ₁ and upper limit value σ _{2 of the} reference dispersion value). As described above, when the variance value σ is σ ₁ to σ _2, the contour emphasis processing is substantially proportional to the σ, and when the variance value σ is σ ₂ or more, the contour emphasis processing is substantially inversely proportional to the σ. .

According to this characteristic, the contour emphasis processing is performed according to the degree of light and shade (when the variance value σ is small), and not too much for the shade (when the variance value σ is large). The contour enhancement processing can be performed.

Here, FIG. 1 shows a specific circuit configuration of an ultrasonic diagnostic image processing apparatus based on such a principle.

In FIG. 1, the ultrasonic diagnostic image processing apparatus of the present invention is
A probe 10, a memory 12, a variance value calculation circuit 16 including an average value calculation circuit 14, a comparison / calculation circuit 18, and a constant setting circuit 20.
And a pixel value emphasizing circuit 22.

That is, a predetermined ultrasonic signal is transmitted from the probe 10 to the subject.
The received and transmitted echo signals are serially stored in the memory 12 for one screen as image information.

In the memory 12, the grayscale information of each pixel is stored as a two-dimensional digital image.

Then, the image density information stored in the memory 12 is output from the memory 12 as the pixel value of interest z and the average value Zn of the regions in the n * n neighborhood including this pixel.

Therefore, the variance value calculation circuit 16 first selects a pixel group (pixel value n × n) in a predetermined area centered on the pixel of interest by the average value calculation circuit 14, and calculates the pixel average value and the pixel 2 of the pixel group. The arithmetic mean value ² is calculated and the variance value σ is calculated.

Here, the variance value σ is obtained based on the average value of the pixel groups n * n in the predetermined region and the pixel mean square value ² , which is represented by σ = ^2- () ² .

Then, from the dispersion value σ, σ ₁ , σ ₂ , which are predetermined by the constant setting circuit 20 by the comparison / arithmetic circuit 18,
Based on KE, the three states K of the above equations (1), (2), and (3) are calculated.

That is, in the comparison / arithmetic circuit 18, the variance value σ is compared with _two different predetermined reference variance values σ ₁ and σ ₂ , and the variance value σ is smaller than the respective reference variance values σ ₁ and σ _2. It is determined whether it is large or not.

Specifically, K in the case of σ <σ ₁ , σ ₁ ≦ σ ≦ σ ₂ , and σ ₂ <σ is calculated as in the above equations (1), (2), and (3). .

Then, the pixel value emphasizing circuit 22 inputs the K to calculate z + K * (z-), performs edge emphasizing processing, and outputs zz.

That is, as shown in FIG. 2, the variance value σ is σ ₁ ≦
When it is determined that σ ≦ σ ₂ , contour enhancement processing that is substantially proportional to the variance value σ is performed, and the variance value σ is σ ₁ <σ ₂ <σ
If it is determined that the contour enhancement processing is performed, the contour enhancement processing is approximately inversely proportional to the variance value σ.

As a result, in the pixel value emphasizing circuit 22, if the target pixel value is larger than the pixel average value, it is set larger,
When the pixel average value is smaller than the average value, it is possible to perform the contour emphasis processing which is set smaller.

Next, an example of specific contour image processing will be described in detail with reference to FIGS.

In FIG. 3, first, the outline emphasis process when the gradation difference is gentle (when the variance value σ is small) will be described.

First, assume that a pixel group (n × n) is selected in a predetermined area S of the subject 24 in the image display 26.

Here, for example, if the pixel group is selected by 3 × 3, it becomes as shown in (a) in the case of the conventional unprocessed contour coordination, and the gray level is 24, 26, 28.

However, the conventional contour enhancement processing having only the lower limit level standard dispersion value σ ₁ described above (in the case of Japanese Patent Application No. 63-317918).
Then, as shown in (b), the gray level is 24 and 26 except 28 as compared with (a), and it can be seen that the contour enhancement processing is performed.

On the other hand, according to this embodiment having _two different reference variance values σ ₁ and σ ₂ , the result is as shown in (c), which is 24 and 26, which are the same as the conventional gray level shown in (b) above. I understand.

Therefore, when the variance value σ in the 3 × 3 pixel group is within the range of σ ₁ ≦ σ ≦ σ ₂ and the grayscale difference is relatively gentle, the contour enhancement processing is performed at the same grayscale level as in the conventional case (b). This is performed, and the three states of (a), (b), and (c) are compared, as shown in FIG.

As a result, when the gradation difference is gentle, unprocessed (a)
It is understood that, in comparison with (b) and (c), appropriate contour enhancement processing is performed.

However, as shown in FIG. 4, when the gray level difference is large, the gray level 2 of the unprocessed (a) in the 3 × 3 pixel group is
In contrast to 4,24,40, in the conventional contour enhancement processing (b), the grayscale level 24,0,64 is extremely sharply emphasized, and excessive contour enhancement processing is performed.

That is, in (b) above, since there is only one reference variance value σ _{1 and} there is no reference value that limits the upper limit level, the variance value σ ₁
This is because the contour enhancement processing that is approximately proportional to is always performed.

On the other hand, in the contour emphasizing process of the present embodiment, as shown in FIG. 4 (c), the gray level is 24, 21, 43, and excessive emphasis is suppressed as compared with the above (b), and it is appropriate. It is understood that various contour enhancement processing is performed.

This means that a reference variance value σ ₂ that defines the upper limit level is set, and when the variance value σ in the 3 × 3 pixel group is σ ₁ <σ ₂ <σ, contour enhancement processing that is approximately inversely proportional to the variance value σ is performed. It is done by things.

Therefore, as is clear from the characteristics shown in FIG. 6, when the difference in gray level is large, the difference in edge enhancement level is more obvious than in the conventional edge enhancement process (b), and excessive gray level enhancement occurs. It can be seen that the values are suppressed (however, the characteristics shown in FIGS. 5 and 6 show constants KE = 5, σ ₁ = 6, σ ₂ = 20 as an example).

Next, in this embodiment, two different reference variance values σ
_{Although the case where 1} and σ ₂ are fixed to predetermined values has been shown, it is of course possible to arbitrarily change σ ₁ and σ ₂ respectively, and the contour emphasis level can be changed according to this change. An example of this is shown in FIG.

FIG. 7 shows a specific internal configuration of the constant setting circuit 20 shown in FIG.
Are variable resistors 30 and 32 that variably set the reference dispersion values σ ₁ and σ _2, and A / D converters 34 and 36 that convert analog signals from the variable resistors 30 and 32 into digital signals. And a switching unit 38 for switching the variable set values of the reference dispersion values σ ₁ and σ ₂ .

The variable resistors 30 and 32 can arbitrarily set reference dispersion values σ ₁ and σ ₂ , and the observer adjusts the variable resistors to change σ ₁ and σ ₂ while observing the image state to enhance contour enhancement. You can change the level.

In FIG. 7, the setting of the reference dispersion values σ ₁ and σ ₂ can be arbitrarily adjusted by the variable resistor, but the present invention is not limited to this, and the σ ₁ and σ ₂ can be changed stepwise by a predetermined level. You can also

In addition, the switching unit 38 is 2 × σ ₁ = σ by switching ON / OFF of the switch unit 40 from the operation panel (not shown).
It is possible to switch whether to establish the relationship of ₂ .

By this switching, it is possible to obtain a curve which is substantially inversely proportional when the dispersion value σ shown in FIG. ₂ is σ ₂ or more (σ ₁ <σ ₂ <σ).

That is, by the switching unit 38, the curve of (a) shown in FIG. ₂ can be obtained when 2 × σ ₁ = σ ₂ , and in other cases, it is set by the variable resistors 30 and 32. The curve (a) or, for example, the curve (c) having the relationship of σ ₁ / σ ₂ = 6/20 can be used.

Thereby, for example, the contour emphasis level can be changed according to the affected part when the difference in density is large.

Of course, in the contour enhancement processing according to the present embodiment shown in FIGS. 1 and 7, the obtained variance value σ is σ ₁ ≦ σ ≦ σ
_In the case of ₂ , only substantially proportional contour enhancement processing is performed, and σ ₁ <σ
It is also possible not to perform the substantially inversely proportional contour enhancement processing when ₂ <σ, which is achieved by, for example, the constant setting circuit 2
0, performed by the pixel value emphasizing circuit 22.

Further, K in σ ₂ <σ may be φ, and the same effect can be obtained.

According to the present embodiment, specifically, the above constant set values are set as KE = 2, σ ₁ = 300, σ ₂ = 600, experimentally computer-simulated, and the conventional contour enhancement processing is performed. It can be confirmed that there is a difference and a remarkable effect.

In this way, in the contour enhancement processing according to the present embodiment, the observer observes the image display and adjusts the reference variance value σ according to the affected area.
₁ and σ ₂ can be varied so that the contour emphasis state can be set to an optimum value, and the contour emphasis processing according to the affected area can be easily performed.

[Effects of the Invention] As described above, according to the ultrasonic diagnostic image processing apparatus of the present invention, the pixel value emphasizing unit does not emphasize an extremely large shade area, and the variance value σ is set to two. It is possible to perform comparison and determination based on different reference variance values σ ₁ and σ ₂ , and based on this, optimum edge enhancement processing can be performed on the image display.

Further, by making each of σ ₁ and σ ₂ variable,
It is possible to emphasize any shaded area according to the affected area.

As a result, the diagnostician can observe the image display state of the affected area that has been subjected to the contour enhancement processing properly, and thus the diagnosis can be easily performed.

[Brief description of drawings]

FIG. 1 is a circuit block diagram showing an example of an ultrasonic diagnostic image processing apparatus according to the present invention, FIG. 2 is an operation explanatory view of the pixel value emphasis circuit shown in FIG. 1, FIG. 3 and FIG. Is an explanatory view showing a specific emphasis processing state in a predetermined pixel region of interest, FIGS. 5 and 6 are explanatory views showing the emphasis processing state in a graph, FIG. 7 is FIG. 3 is a circuit block diagram showing an example of the constant setting circuit of FIG. 12 ... Memory 14 ... Average value arithmetic circuit 16 ... Dispersion value arithmetic circuit 18 ... Comparison / arithmetic circuit 20 ... Constant setting circuit 22 ... Pixel value emphasis circuit 30,32 ... Variable resistors 34,36 ... A / D converter 38 ... Switching unit σ ... Variance value σ ₁ , σ ₂ ... Reference variance value

Claims (2)

[Claims] 1. A pixel average value calculating means for selecting a pixel group centering on a pixel of interest and calculating a pixel average value of the pixel group, and based on each pixel value of the pixel group and the pixel average value,
Pixel dispersion value calculation means for calculating a dispersion value σ for the pixel group, and the dispersion value σ are compared with two different predetermined reference dispersion values σ ₁ , σ ₂ (where σ ₁ <σ ₂ ). A pixel value emphasizing unit that performs contour emphasizing processing on the pixel value of the pixel of interest with an emphasizing degree according to the magnitude of the variance value σ, wherein the pixel value emphasizing unit includes the σ ₁ To σ ₂ have a characteristic of gradually increasing the degree of emphasis, and have a characteristic of gradually weakening the degree of emphasis in a region larger than σ ₂ , and when the variance value σ of the pixel group is larger than σ _2, An ultrasonic diagnostic image processing apparatus characterized by suppressing excessive emphasis of an image. 2. The ultrasonic diagnostic image processing apparatus according to claim 1, wherein the contour enhancement processing is sequentially performed in parallel by a pipeline processing method.