JP2959721B2 - Image data interpolation device - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to an image data interpolation device that estimates data at a desired position from original data by interpolation.

[Problems to be Solved by the Related Art and the Invention] Image data obtained by ultrasonic diagnosis or the like is data sampled at regular intervals, and data at a desired position is obtained by interpolation from the obtained original data. May be estimated. When performing the interpolation, conventionally, for example,
The data of a point between the two points has been obtained by a numerical operation using the data of the point. For example, when the data of two adjacent points is xn, Xn + 1, the data of the center point between the two points is obtained as an average value (xn + Xn + 1) / 2.

However, with interpolation using only two points of data,
In particular, when the amount of change in data before and after the position determined by interpolation is large, interpolated data far from the actual value may be obtained. Therefore, it is conceivable to obtain the amount of change in the data and perform a strict numerical operation using the data to obtain more appropriate interpolation data.However, since the numerical operation takes too much time, it is necessary to obtain the interpolation data in real time. However, there is a problem that the original data and the interpolated data cannot be displayed in real time.

The present invention has been made in view of the above circumstances, and an object of the present invention is to provide an image data interpolating apparatus capable of obtaining interpolation data expected to be close to an actual value at a high speed.

[Means and Actions for Solving the Problems] An image data interpolation device according to the present invention includes an image data holding unit that holds four image data at positions adjacent to each other, and an image data holding unit that is adjacent to each other among the four image data. First calculating means for calculating a data change amount between the first and second image data from the first and second image data;
A second calculating means for calculating interpolated image data for interpolating between the two image data from the second and third image data adjacent to each other among the four image data; Third calculating means for calculating the data change amount between the two image data from the third and fourth image data to be processed, and fuzzy processing based on the data change amount calculated by the first and third calculating means. Fuzzy operation means for inferring and calculating a correction value; and interpolated image data corrected based on the interpolated image data obtained by the second operation means and the correction value obtained by the fuzzy operation means. Means.

Embodiment An embodiment of the present invention will be described below with reference to the drawings.

FIGS. 1 to 5 relate to an embodiment of the present invention.
FIG. 2 is a block diagram showing the configuration of an ultrasonic diagnostic apparatus including an image data interpolation apparatus, FIG. 2 is an explanatory view showing the positions and sizes of original data and interpolation data, and FIG. 3 is an explanatory view showing an outline of fuzzy inference. FIG. 4 is an explanatory diagram showing fuzzy rules and membership functions, FIG. 5 (a) is an explanatory diagram showing original data, and FIG. 5 (b) is an explanatory diagram showing original data and interpolation data by a conventional interpolation method. FIG. 5C is an explanatory diagram showing the original data and the interpolated data by the interpolation method of the present embodiment.

This embodiment is an example in which the present invention is applied to data interpolation in an ultrasonic diagnostic apparatus.

As shown in FIG. 1, the ultrasonic diagnostic apparatus 1 includes an ultrasonic probe 2 and a signal processing circuit 3 connected to the ultrasonic probe 2. The signal processing circuit 3 transmits an ultrasonic pulse to the ultrasonic probe 2, receives an echo signal from the ultrasonic probe 2, and performs logarithmic compression, S
Signal processing such as TC is performed. Further, the ultrasonic probe 2 performs scanning such as linear, sector, and radial.

The image data from the signal processing circuit 3 includes latches 4, 5,
The data is input to the image memory 8 through the steps 6 and 7. The latches 4, 5, 6, and 7 store image data at positions (scan lines) adjacent to each other. That is, as shown in FIG.
Image data of two points n−1, n, n + 1, n + 2 are respectively
Assuming that xn−1, xn, xn + 1, xn + 2, the latch 4 becomes xn−1
Latch 5, xn, Latch 6, xn + 1, Latch 7,
xn + 2 are respectively stored.

Each data stored in the latches 5 and 6 is input to an arithmetic circuit 10, and the arithmetic circuit 10 calculates (xn + xn + 1) / 2 as an interpolation value of a center point between the points n and n + 1. I have. Each data stored in the latches 4 and 5 is input to an arithmetic circuit 11, and the arithmetic circuit 11
The amount of change (slope) Δx ₁ = xn−xn−1 between data is calculated. Each data stored in the latches 6 and 7 is input to an arithmetic circuit 12, and the arithmetic circuit 12
Data change amount (slope) between +1 and n + 2 Δx ₂ = xn + 2-x
n + 1 is calculated.

The change amount Δx ₁ obtained by the arithmetic circuits 11 and 12,
Δx ₂ is input to a fuzzy operation unit 13, which uses the fuzzy inference to calculate the interpolation value (xn + xn
A correction value k for correcting +1) / 2 is obtained. The interpolation value (xn + xn) obtained by the arithmetic circuit 10
+1) / 2 and the correction value k obtained by the fuzzy calculation unit 13 are added by an adder 14, and the added value is input to the image memory 8 as interpolation data x. I have. The image memory 8 stores original data and interpolation data, and the image data read out from the image memory 8 is input to a monitor 16, and the monitor 16 stores an ultrasonic data such as an interpolated ultrasonic tomographic image. Sound wave information is displayed.

Here, an outline of fuzzy inference will be described with reference to FIG.

The fuzzy theory was proposed in 1965 by Professor LAZadeh of the University of California, and in 1974 EHMamdani of the University of London showed its feasibility, after which various realization methods were proposed. I have. Fuzzy inference using this fuzzy theory is inference using fuzzy rules (fuzzy inference rules) expressed in ambiguous words used by humans in everyday life.
As shown in FIG. 3, the fuzzy rule is “if A = BIG
and B = NORMAL then X = SMALL "(" If A is large and B is normal, X is small "). Similarly, the other fuzzy rules are "if B
= VERY BIG and C = SMALL then X = NORMAL ("If, B
If X is very large and C is small, X is normal ")
Can be described as follows. Here, A, B, and C are input variables, and X is an output variable. The part where the conditions for establishing the rule are written, such as “if A = BIG and B = NORMAL then X = SMALL”, is called the antecedent part, and the conclusion part “X = SMALL” is called the consequent part. In the fuzzy inference, each input variable is converted into a value of 0 to 1 for calculation, and this conversion is defined by a membership function (antecedent membership function). As shown in FIG. 3, the membership function is defined for each proposition (BIG, NORMAL, SMALL, etc.) handled by the fuzzy rule. Then, the degree to which the input variable satisfies each proposition is calculated with reference to the membership function. When there are a plurality of propositions connected by "and" in the antecedent part, the minimum value of the degrees at which each input variable satisfies each proposition is obtained. This is called minimum value (MIN) operation. Next, the membership values for each rule are combined. This is done by comparing the consequents of each rule, taking the maximum value and creating a new membership function. This is called a maximum value (MAX) operation. The value of the position of the center of gravity of the combined membership function is the inference result (output value), and the subsequent control is performed based on the result. Note that the above inference method is a typical example, and the present invention is not limited to this.

Next, fuzzy inference in the fuzzy operation unit 13 will be described.

In the present embodiment, the data change amount (slope) Δx ₁ = xn−xn−1 between points n−1 and n, and the data change amount (slope) Δx ₂ = xn + 2−xn + 1 between points n + 1 and n + 2 are calculated as follows: Expressed as a fuzzy scale as follows:

Large positive ++ Small positive + No change 0 Small negative-Large negative --- Here, 0≤xi≤1 (where i = ..., n-1, n, n + 1, n + 2,
..), −1 ≦ Δx ₁ ≦ 1 −1 ≦ Δx ₂ ≦ 1.

The following nine rules are adopted as fuzzy rules in this embodiment. In the following, in the consequent part, "..." in "the interpolation data is calculated from (xn + xn + 1) / 2"
Only the part is described.

1. If Δx ₁ = ++ and Δx ₂ = −−, increase the value.

2. If Δx ₁ = ++ and Δx ₂ = −, slightly increase.

3. If Δx ₁ = + andΔx ₂ = −, increase the value slightly.

4. If Δx ₁ = + andΔx ₂ = −−, make it slightly larger.

5. If Δx ₁ = −and Δx ₂ = ++, make it a little smaller.

6. If Δx ₁ = −and Δx ₂ = +, reduce a little.

7. If Δx ₁ = −− and Δx ₂ = +, make it a little smaller.

8. If Δx ₁ = −− andΔx ₂ = ++, decrease the value.

9. If Δx ₁ = 0 and Δx ₂ = 0, do not change.

It is assumed that "slightly" is greater than "slightly".

FIG. 4 shows the above rules using a membership function.

The fuzzy calculation unit 13 applies the above nine rules simultaneously to the input values Δx ₁ and Δx ₂ , calculates the degree of coincidence of the antecedent of each rule with respect to the input by MIN calculation, and calculates
The weight is applied to the consequent part of the rule. And MAX
The consequent part of each rule is synthesized by calculation, and for example, a barycenter value of the obtained membership function is set as an output value. This output value is used as a correction value k for increasing or decreasing the interpolation data from (xn + xn + 1) / 2. For example, in FIG. 4, when there is an input value as indicated by a vertical line traversing the membership function of the antecedent part of each rule, rules 3 and 9 are applied with weights indicated by oblique lines, respectively. As a result, a membership function of the inference result is obtained. The execution of each fuzzy rule and the calculation of the inference result as described above can be performed at high speed by a fuzzy chip.

For the original data as shown in FIG.
In the case of obtaining interpolation data at the center point between +1 and (xn + xn + 1) / 2 as in the prior art, a value far from the actual value can be obtained as shown in FIG. 5 (b). There are cases. On the other hand, as in the present embodiment, the interpolation data is set to (xn + xn + 1) / 2 + k by adding the correction value k obtained by fuzzy inference by inputting the change amount of the original data, thereby obtaining a more actual value. Interpolation data expected to be close can be obtained.

In addition, since the fuzzy operation can be performed at high speed by the fuzzy chip, the interpolation data can be obtained in real time, and the original data and the interpolation data can be simultaneously displayed in real time.

[Effects of the Invention] As described above, according to the present invention, interpolation data is obtained using fuzzy inference based on the original data and the amount of change in the original data near the position where the image interpolation data is obtained. Therefore, there is an effect that image interpolation data expected to be close to the actual value can be obtained at high speed.

[Brief description of the drawings]

1 to 5 relate to an embodiment of the present invention, FIG. 1 is a block diagram showing a configuration of an ultrasonic diagnostic apparatus including an image data interpolating apparatus, and FIG. FIG. 3 is an explanatory diagram showing an outline of fuzzy inference, FIG. 4 is an explanatory diagram showing fuzzy rules and membership functions, FIG. 5 (a) is an explanatory diagram showing original data, FIG. FIG. 5B is an explanatory diagram showing the original data and the interpolated data by the conventional interpolation method, and FIG. 5C is an explanatory diagram showing the original data and the interpolated data by the interpolation method of the present embodiment. 4, 5, 6, 7 latch (image data holding unit) 10 arithmetic circuit (second arithmetic unit) 11 arithmetic circuit (first arithmetic unit) 12 arithmetic circuit (third arithmetic unit) Means 13 Fuzzy calculation unit (Fuzzy calculation means) 14 Adder (Interpolated image data generation means)

Continuation of the front page (72) Inventor Hiroki Hibino 2-43-2 Hatagaya, Shibuya-ku, Tokyo Within O-limpus Optical Industry Co., Ltd. (72) Inventor Tomohisa Sakurai 2-43-2 Hatagaya, Shibuya-ku, Tokyo O-limpus Inside Optical Industry Co., Ltd. (72) Inventor Yutaka Takahashi 2-43-2 Hatagaya, Shibuya-ku, Tokyo O-limpus Optical Industry Co., Ltd. (72) Inventor Akira Murata 2-43-2 Hatagaya, Shibuya-ku, Tokyo O-limpus Within Optical Industry Co., Ltd. (72) Nobuyuki Sakamoto 2-43-2 Hatagaya, Shibuya-ku, Tokyo O-limpus Optical Industry Co., Ltd. (72) Yoshihiro Kosaka 2-43-2 Hatagaya, Shibuya-ku, Tokyo O-limpus Inside Optical Industry Co., Ltd. (72) Inventor Koichi Matsui 2-43-2, Hatagaya, Shibuya-ku, Tokyo O In-Limpus Optical Industry Co., Ltd. (72) Inventor Shoichi Gotanda 2-43-2, Hatagaya, Shibuya-ku, Tokyo O Departed from Linpass Optical Co., Ltd. (72) Person Kazunori Kobayashi Within O-limpus Optical Industry Co., Ltd. 2-43-2, Hatagaya, Shibuya-ku, Tokyo (72) Inventor Yoshiji Koda 2-43-2 Hatagaya, Shibuya-ku, Tokyo Within O-lympus Optical Industries Co., Ltd. (56) References JP-A-2-93942 (JP, A) JP-A-64-8439 (JP, A) JP-A-1-126779 (JP, A) (58) Fields investigated (Int. Cl. ⁶ , DB name ) G06F 17/17 G06F 9/44 554 G06T 3/40

Claims (1)

(57) [Claims] 1. An image data holding means for holding four image data at positions adjacent to each other, and a first image data and a second image data adjacent to each other among the four image data, First calculating means for calculating a data change amount between the first and second image data, and interpolated image data for interpolating between the two image data from the second and third image data adjacent to each other among the four image data. A second calculating means for calculating, from third and fourth image data adjacent to each other among the four image data, a data change amount between the two image data; Fuzzy calculating means for performing a fuzzy inference based on the data change amount calculated by the first and third calculating means to calculate a correction value; and interpolating image data obtained by the second calculating means. Based on the correction value obtained by data and the fuzzy operation means, the image data interpolation apparatus comprising: the means for generating a corrected interpolated image data, further comprising: a.