JP3538260B2 - Ultrasonic measuring device - 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 measuring apparatus such as a medical ultrasonic diagnostic apparatus, and more particularly, to a method for detecting a waveform distortion of a transmitted ultrasonic wave or a received ultrasonic wave due to non-uniform sound velocity in a subject. The present invention relates to a signal processing method for correcting.

[0002]

2. Description of the Related Art In recent years, the diameter of an electroacoustic transducer array for transmitting and receiving ultrasonic waves can be made relatively large with the advancement of the performance of ultrasonic diagnostic apparatuses.

[0003] However, it has been pointed out that even when an image is taken of a real human body using a large-diameter array, the resolution required for the aperture width may not be obtained depending on individual differences. The cause of the reduction in resolution is due to the non-uniformity of the human body tissue.

Many current ultrasonic diagnostic apparatuses geometrically calculate a delay time difference between received signals on the assumption that the speed of sound in a human body is almost constant near water.

[0005] Among the tissues constituting the human body, the subcutaneous fat tissue has a sound velocity of about 1400 m / s and the other tissues have a velocity of 1500 m / s.
It is known that the speed of sound is remarkably slow as compared with about / s.

[0006] For this reason, it is presumed that this is the main cause of aberration (sound wave propagation time difference) when focusing by transmitting and receiving ultrasonic waves.

For this reason, a method of assuming a two-layer structure of a layered fat layer and another uniform layer and calculating a time difference by a cross-correlation operation between received signals to correct aberrations only for the phase is known, for example. It is disclosed in, for example, JP-A-63-51846.

[0008]

When it is considered that the cause of aberration caused by the fat layer is diffraction at a spatially varying boundary surface, the waveform not only changes in propagation time due to a change in sound speed but also changes in amplitude. Both the phase and the phase change for each frequency component.

Only when the change in amplitude and the change in frequency of the wavefront that has undergone diffraction at such a boundary surface is extremely small in relation to the sound source (reflection source) and the receiving point, the above publication (Japanese Unexamined Patent Publication No. Sho 63 (1988)). -51846), it is possible to correct only by the phase.

However, in general, not only such a suitable boundary surface exists, but a correction method capable of correcting the amplitude and phase that change for each frequency component in consideration of the diffraction phenomenon is disclosed. Needed.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems of the prior art. It is an object of the present invention to provide an ultrasonic measuring apparatus which estimates a sound velocity distribution in a subject to determine the amplitude and phase. It is an object of the present invention to provide a technique capable of correcting waveform distortion of a transmitted ultrasonic wave or a received ultrasonic wave which changes for each frequency component.

The above and other objects and novel features of the present invention will become apparent from the description of the present specification and the accompanying drawings.

[0013]

Means for Solving the Problems Of the inventions disclosed in the present application, the outline of a representative one will be briefly described.
It is as follows.

(1) A plurality of electroacoustic transducers are provided, ultrasonic waves are transmitted into a subject, reflected echoes are received by the plurality of electroacoustic transducers, and a received signal is signal-processed. In an ultrasonic measurement apparatus that obtains information on the inside of a subject, a plurality of boundary shape curves when the subject is considered as a two-layer medium structure having different sound speeds are assumed, and a plurality of reference points provided in the subject and the plurality of electric points are provided. First means for determining a transfer function between the plurality of boundary shape curves and a transfer function with respect to each of the plurality of boundary shape curves, and further determining an inverse transfer function from the determined transfer function;
The convolution operation in the time domain or the frequency domain between the inverse transfer function obtained by the means and the actually received signal in a specific time section is performed, and the reflected signal for each reference point is converted to the plurality of electric signals. Estimating means for estimating in relation to the acoustic transducer, and selecting means for selecting an optimal curve from the plurality of boundary shape curves based on the reflection signal at each reference point estimated by the estimating means, Based on the optimal boundary shape curve selected by the selecting means, the inverse transfer function between the focal point of the ultrasonic wave transmission or reception provided in the subject and the plurality of electro-acoustic transducers, respectively. And a signal processing means for performing convolution signal processing of the determined inverse transfer function and the transmitted signal or the received signal among the determined inverse transfer functions.

(2) In the means of the above (1), the first means may include a communication between a plurality of reference points provided in the subject and the plurality of electroacoustic transducers by Fresnel-Kirchhoff integration. It is characterized by including means for calculating a function.

(3) In the means of (1) or (2), the selection means is a square root of a time integration value of a square after adding all the reflection signal groups estimated at the same reference point. The energy evaluation function determined by, the similarity evaluation function determined by dividing the energy evaluation function by the sum of the square root of the time integral of each square is obtained at all the reference points, and further, the evaluation function value group as a variable A means for selecting an optimum curve from the plurality of boundary shape curves by optimizing the evaluation function for all reference points.

(4) In the means of (3), the total reference point evaluation function is a standard deviation from an average value of the similarity evaluation function values, and optimization is performed with a minimum of the standard deviation. I do.

(5) In the means of (3) or (4), changing a spatial density in the subject of a reference point, which is a variable of the all reference point evaluation function, or a weighting factor for multiplying a deviation of each reference point. It is characterized in that a means for enabling is provided.

(6) In the means of (1) to (5), the signal processing means transmits or receives a plurality of ultrasonic waves provided in the subject based on the selected boundary shape curve. , And transfer functions between the plurality of electro-acoustic transducers are calculated by Fresnel-Kirchhoff integration, and a quantization delay operation and an impulse determined for each inverse transfer function obtained from the calculated transfer functions are calculated. It is a means for performing a convolution operation in a time domain between a response coefficient group and a transmission signal or a reception signal.

(7) In the means of (1) to (6), the first means is an inverse transfer function between a plurality of reference points provided in the subject and the plurality of electroacoustic transducers. To
All of the results calculated for each of the plurality of boundary shape curves are stored in the storage means, and the inverse transfer function is sequentially read from the storage means.

[0021]

A plurality of electroacoustic transducers are provided. The plurality of electroacoustic transducers are brought into contact with a subject to transmit ultrasonic waves, and reflected echoes are received by the plurality of electroacoustic transducers. In an ultrasonic measuring apparatus that obtains information about the inside of a subject by performing signal processing on a received signal, the inside of the subject can often be approximated by a two-layer medium structure in which the sound speeds of a fat layer and a main organ layer are different.

According to the above-mentioned means (1) to (3),
A plurality of unknown two-layer boundary shape curves are assumed in advance, and the transfer function between a plurality of imaginary reference points provided in the subject and the plurality of electroacoustic transducers is defined for each of the plurality of boundary shape curves. For example, Fresnel-Kirchh
It is calculated by off integration.

The Fresnel-Kirchhoff integral formula (diffraction formula) calculates the wave propagation relationship between a sound source (reflection source) and a receiving point or a transmitting point and a transmitting focal point that captures the reverse, in a form dependent on frequency. It is possible to approximately estimate how an impulse-like wave generated from one point across the layer boundary is transmitted to the other.

In order to return a received signal of a normal pulse whose band is limited to a waveform before transmission, a Fresn is calculated from a boundary shape curve which approximates the actual sound velocity distribution in the subject well.
A complex division (inverse filtering, deconvolution) operation in the frequency domain with a transfer function calculated by el-Kirchhoff integration may be performed.

This can be called a convolution operation of the transfer function with the inverse transfer function, and is hereinafter referred to as "back propagation".

A specific time section of the actually obtained received signal is back-propagated, and a reflection signal at a point (reference point) assumed in the subject is estimated for each reference point in relation to a plurality of electroacoustic transducers. I do.

At this time, if the boundary shape curve obtained by calculating the transfer function used for the backpropagation closely approximates the actual sound velocity distribution in the subject, the received waveforms after the backpropagation are very similar to each other. It becomes.

In order to evaluate the degree of similarity, for the reflection signal group estimated at the same reference point, an energy evaluation function determined by the square root of the squared time integrated value after all addition, and the time of each square A similarity evaluation function determined by dividing the energy evaluation function by the sum of the square roots of the integrated values is used.

If the received waveforms after backpropagation differ only in the amplitude within a section where time integration is performed with each other, and when the waveform shapes match, the similarity evaluation function is 1, which is generally a positive real number.

If the waveform shapes are similar to each other, the optimum condition is that the energy evaluation function after addition is maximum.

This is based on the assumption that the most correct backpropagation should result in the lowest received energy loss in the entire image.

The two evaluation function value groups are obtained at all reference points which are determined in advance by a predetermined position and a predetermined number, and the above-mentioned two evaluation function value groups are used as variables to optimize the entire reference point evaluation function. An optimum curve can be selected from a plurality of boundary shape curves.

On the basis of the obtained optimal boundary shape curve, an inverse based on a transfer function between a plurality of focal points of transmitting or receiving ultrasonic waves and a plurality of electroacoustic transducers provided in the subject. The transfer function is obtained, a predetermined inverse transfer function is selected from the transfer functions, and the transmitted signal or the received signal is convolved with the convolution signal processing to correct the waveform distortion of the transmitted ultrasonic wave or the received ultrasonic wave. It becomes possible.

According to the means of (4), since all the reference point evaluation functions are standard deviations from the average value of the similarity evaluation function values, optimization is performed by minimizing the standard deviation, thereby making it possible to optimize the evaluation function within the subject. It is possible to prevent a biased correction effect from appearing in a specific area of the reference points distributed in the region.

According to the means of (5), there is provided means for changing the spatial density of the reference point, which is a variable of the total reference point evaluation function, within the subject or the weight coefficient by which the deviation of each reference point is multiplied. Therefore, when the actual sound velocity distribution in the subject cannot be sufficiently approximated over the entire measurement region using only the layer boundary shape model of the present invention, the specific region of interest of the reference points distributed in the subject is mainly used. Correction can be performed.

According to the above-mentioned means (6), a plurality of ultrasonic wave transmitting or receiving focal points provided in the subject and a plurality of electroacoustic transducers are determined based on the obtained optimum boundary shape curve. Between the transfer functions calculated by the Fresnel-Kirchhoff integral, and the time domain convolution of the quantization delay operation and the impulse response coefficient group determined for each inverse transfer function obtained from the calculated transfer function. Therefore, the number of impulse response coefficients required for the back propagation operation can be realized by a convolution operation over a relatively small time width, whereby the scale of a circuit provided for each received signal can be effectively optimized. .

According to the means (7), the inverse transfer function between the plurality of reference points provided in the subject and the plurality of electroacoustic transducers is calculated for each of the plurality of boundary shape curves. Is read from the storage unit that stores all
Since a portion requiring the most calculation circuit scale and the most calculation time in the correction operation can be prepared in advance, the time until the measurement based on the result of correcting the received signal is achieved can be significantly reduced.

[0038]

Embodiments of the present invention will be described below in detail with reference to the drawings.

In all the drawings for explaining the embodiments, parts having identical functions are given same symbols and their repeated explanation is omitted.

FIG. 11 shows Fresnel-Kirchh
It is a figure for explaining transfer function calculation based on off diffraction model.

First, referring to FIG. 11, Fresnel
The transfer function calculation based on the Kirchhoff diffraction model will be described.

The object layer is composed of two layers (L1, L1) having different sound speeds.
L2).

[0043] In FIG. 11, a first layer (L1) represents the fat layer, the sound velocity c _r is 1400 [m / s] before and after addition, Layer 2 (L2) organs such as the liver It mainly represents the substantial layer, and its sound speed c _R is 1540 [m /
s].

In the second layer (L2), a reference point group 81 for performing correction is assumed.

Here, the reference point groups 81 do not need to be arranged at equal intervals, but are appropriately set according to the purpose of correction.

The electroacoustic transducers (101 to 10c) are in contact with the subject in order to transmit and receive ultrasonic waves.

When the minute boundary element e on the layer boundary 82, the reference point T, and the center point of the electroacoustic transducer 10i are taken up,
Three space vectors are calculated.

A vector R is a displacement vector from the reference point T to the boundary element e, a vector n is a normal vector of the minute boundary element e toward the inside of the second layer (L2), and a vector r is an electroacoustic signal from the minute boundary element e. It is a displacement vector to the center of the conversion element 10i.

The components are R = (Rx, Ry), r =
(Rx, ry), n = (nx, ny).

Since the difference between the sound velocities of the two (L1, L2) is relatively small, all waves generated from the reference point T in the second layer (L2) and reaching the layer boundary 82 are reflected without being reflected. It is approximated that the light is propagated in the first layer (L1).

In the Kirchhoff integration formula that satisfies the boundary integration condition of the wave equation, it is considered that the distance between the reference point T and the boundary and the distance between the center of the electroacoustic transducer 10i and the boundary are sufficiently larger than the wavelength. And omit the small term, the next well-known Fresnel-Ki
The rchoff integral equation (Equation 1) is obtained.

[0052]

(Equation 1)

Here, (r · n) and (R · n) are inner products of vectors, the lengths of R and r are represented by | R | and | r |, respectively, and the imaginary unit is represented by j.

The layer boundary 82 is divided by a primary boundary element having an equal length and shorter than the wavelength of the signal band upper limit frequency of the ultrasonic pulse signal at the time of imaging, and the boundary integral of the continuous function of the above equation is divided into discrete elements. If the constant term that does not depend on each frequency is removed after replacing with the sum of the following, the following discretization equation (Equation 2) is obtained.

[0055]

(Equation 2)

According to this equation, the transfer function relation between the reference point T in the subject and one element 10i in the electroacoustic transducer array can be calculated for each angular frequency ω. , The transfer function の of the phase and the amplitude in the case where there is a single boundary having different sound speeds for all the frequencies of the signal band of.

In general, the ultrasonic pulse generated at the reference point T passes through the boundary surfaces having different sound speeds and the electroacoustic transducer 10i.
The waveform is distorted while reaching the center point of, but if the shape of the layer boundary can be correctly estimated as a model, the inverse of the transfer function Ψ in the frequency space is used as an inverse filter.
Waveform distortion and time delay can be accurately recovered.

Hereinafter, in this specification, the operation of calculating the original signal at the transmission point from the signal at the reception point using the reciprocal of the transfer function Ψ in the frequency space as an inverse filter will be referred to as “back propagation”. I do.

When the ultrasonic pulse generated at the reference point T is received by each of the electroacoustic transducers (101 to 10c),
If the transmission source (reflection source) is only the reference point T, it is considered that the waveforms after the back propagation are all equal.

Further, when the transmission sources (reflection sources) are distributed around the reference point T, the spatial distribution of the transmission sources (reflection sources) viewed from each element in the electroacoustic transducer array changes. Therefore, the waveforms after the back propagation are not all the same, but the correlation between adjacent reception points is very high.

A waveform similarity coefficient ρ is defined by the following equation (Equation 3) as such an evaluation function.

[0062]

[Equation 3]

Here, s (t) is the waveform after the received waveform of each element in the electroacoustic transducer array (101 to 10c) is back-propagated by the model.

If the backpropagation is performed completely correctly, the waveforms after the backpropagation of each element are completely coincident, so that the denominator term and the numerator term are completely equal, and ρ = 1.

However, because of the error of the model for the back propagation and the distribution of the transmitting source (reflecting source) near the reference point, the weakening due to the interference between the signals occurs at the stage of the summation in the numerator term. <1.

It is also considered that the energy in a predetermined time width of the waveform after summation (phasing addition) of the respective received waveforms after back propagation increases.

An energy function ε is defined by the following equation (Equation 4) as such an evaluation function.

[0068]

(Equation 4)

Since the waveform similarity coefficient ρ merely evaluates the similarity of the waveform, the wave source (reflection source)
It can be evaluated by the energy function ε that the loss of energy of the wave reaching the electroacoustic transducer array from is minimized.

In advance, Fresnel-Kirch is performed while sequentially changing the shape model of the layer boundary which cannot be directly measured.
Backpropagation is performed using a transfer function based on the Hoff integration, and the waveform similarity coefficient ρ and the energy function ε are optimized at each reference point as a focus evaluation function, thereby finally obtaining a most suitable layer boundary shape model. .

Next, a measurement method for correcting the waveform distortion of the transmitted ultrasonic wave or the received ultrasonic wave in the ultrasonic measuring apparatus of the present embodiment based on such consideration will be described with reference to FIG.

FIG. 10 is a flowchart showing a processing procedure of the ultrasonic measuring apparatus according to the present embodiment.

After the start of correction (step 70), the model parameters are changed (step 710).

The model parameters generally refer to variables for describing the layer boundary shape.

For example, when the curve function of the boundary shape in the x-axis direction assumed in FIG. 11 is divided into a plurality of sections, and each section is approximated by a quadratic function, the coefficient of each order term, an orthogonal function such as a Fourier series, etc. The amplitude, phase, and the like for each spatial frequency when developed in a system can be considered.

The model parameters are initialized at the correction start stage. For example, a B-mode tomographic image before correction is observed, and the boundary between the fat layer and another tissue is artificially determined from the image to determine its shape. May be used to calculate the initial values of each model parameter.

Next, a layer boundary curve according to the model parameter indicated by 82 in FIG. 11 is divided as a boundary element (step 720).

For example, division points are determined so that each element has the same length, for example, from one end of an assumed layer boundary curve.

After this is completed, a reference point in the subject to be a transmission source (reflection source) indicated by 81 or T in FIG. 11 is selected (step 730), and a point in the electroacoustic transducer array to be a reception point is selected. Is selected (step 740).

Next, a transfer function Ψ ′ is calculated according to (Equation 2) (step 75).

As a result, the change of the amplitude and the phase for each angular frequency can be calculated, and the signal of the specific time window section [T1, T2] of the received signal is back-propagated by the inverse filter based on the change (step 76). .

The inverse filter at this time is simply 1 / Ψ '(ω)
Rather than using, Wiene considering the presence of noise
An r filter may be considered.

Next, in step 741, the processes in steps 740 to 76 are repeated until the back propagation calculation is completed for all the receiving points.

However, since iterative processing in this process can be parallelized depending on the implementation means, it is possible to reduce the number of repetitions or perform processing without repetition depending on the degree of parallelization.

In step 741, when the back-propagation calculation at all the receiving points is completed, the focus evaluation function such as the waveform similarity coefficient ρ and the energy function ε described above is calculated (step 77).

Further, in step 731, the processing in steps 730 to 77 is repeated until the calculation of the focus evaluation function for the reference point in the space is completed.

It is needless to say that the processing speed can be similarly increased by this parallelization in this iterative process.

Next, after the calculation of the focus evaluation function for the reference point in the space is completed in step 731, in step 711, the processing in steps 710 to 731 is repeated for all combinations of model parameters assumed in advance.

At step 711, when the processing at steps 710 to 731 is completed for all combinations of model parameters assumed in advance, the result of the focus evaluation function calculation 77 for the reference points in each space for all combinations of model parameters Is used to determine which combination of model parameters is optimal (step 78).

As the evaluation function for making this determination, there is a method of taking the sum of the waveform similarity coefficients ρ or the sum of the energy functions ε of the reference points in each space, and selecting a condition for maximizing the sum.

When the optimum boundary shape curve is selected in step 78, based on the optimum boundary shape curve, the focal points of a plurality of ultrasonic waves transmitted or received in the subject and the plurality of electric waves are detected. Calculate the transfer function between the acoustic transducer, respectively, further obtain the inverse transfer function based on the transfer function, from the respective inverse transfer function obtained, from the actual transmission of ultrasonic waves or An inverse transfer function corresponding to the focal point of the received wave is selected, and convolution signal processing of the inverse transfer function and the received signal is performed.

Here, if the actual focal point of the transmission or reception of the ultrasonic wave does not coincide with the focal point of the transmission or reception of the ultrasonic waves provided in the subject, the adjacent subject is transmitted. The inverse transfer function corresponding to the focal point of the transmission or reception of the ultrasonic wave provided in the plurality is used.

An ultrasonic measuring apparatus according to an embodiment of the present invention will be described with reference to FIGS.

FIG. 1 is a block diagram showing a schematic configuration of an ultrasonic measuring apparatus according to one embodiment of the present invention.

In the ultrasonic measuring apparatus shown in FIG. 1, the electroacoustic transducers (101, 102 to 10c) are connected to the transmission / reception wave separation circuits (111, 112 to 11c), and the transmission / reception wave separation circuits (111, 112 to 110c). 11c) includes a receiving circuit (1
31, 132 to 13c) and the wave transmitting circuit 12 are connected.

The receiving circuits (131, 132 to 13c)
The input from the transmission / reception separation circuit (111, 112 to 11c) is amplified according to time and output to the analog-digital converter (141, 142 to 14c).

The analog-digital converter (141, 1
42 to 14c) (digital signal output) are output to the inverse filter operation circuit 3 and the back propagation circuit (151, 152 to 152c).
15c).

Back propagation circuit (151, 152 to 15c)
Is a storage circuit (161i,
162i-16ci, 161q, 162q-16cq)
And complementarily.

Storage circuits (161i, 162i to 16c)
i) are added by an adder 171 to obtain an added output 18i of the real components, and the storage circuit (16qi, 1
62q to 16cq) are added by the adder 172 to obtain an added output 18q of the imaginary components.

The signal components (18i, 18q) are input to a signal processing circuit for measuring a subsequent B-mode image and Doppler (not shown).

By the boundary model generation circuit 1, Fresn
The boundary element parameter for el-Kirchhoff integration is Fresnel-Kirchhoff integration circuit 2
Is output to

The boundary element parameters are a vector R from the reference point in the space to the boundary element, a normal vector of the boundary element, and a vector r from the boundary element to the reception point.

These values are set in parallel with a predetermined number of sets of Fr.
Output to the esnel-Kirchhoff integration circuit 2.

The Fresnel-Kirchhoff integration circuit 2 outputs the transfer function value for each frequency to the inverse filter operation circuit 3 for each complex component with the degree of parallelism required from the angular frequency accuracy required for the back propagation calculation.

The inverse filter operation circuit 3 performs a fast Fourier transform (FFT) on the outputs of the analog-digital converters (141, 142 to 14c) and a Fresnel-K
Back propagation (convolution by an inverse filter) is performed by performing complex multiplication with the reciprocal of the transfer function in the frequency domain obtained from the output of the irchhoff integration circuit 2.

The operation result for each frequency is calculated by the evaluation function circuit 7
And the output of the evaluation function circuit 7 (EVOUT)
Becomes a command to the boundary model generation circuit 1 again.

The evaluation function circuit 7 evaluates the result of the back-propagation operation, performs repetition of the back-propagation operation under a new model parameter condition, and specifies the optimal model parameter condition.

When the optimal model parameters are specified by the evaluation function circuit 7, the transfer function for each set of focal points and reception points in space based on the model is again converted to the boundary model generation circuit 1 and Fresnel-Kirchhoff integration circuit 2 And is input to an impulse response generation circuit 4 connected via a branch in the inverse filter operation circuit 3.

The impulse response generation circuit 4 has an output group CI to the back propagation circuits (151, 152 to 15c) given for each received signal.

These are back propagation circuits (151, 152 to 152).
15c) Coefficients of a finite length impulse response (FIR) filter of length d to be convolved with each received signal inside.

The impulse response generation circuit 4 is based on the output of the inverse filter operation circuit 3 and has a quantization delay suitable for accurately realizing the impulse response of the inverse filter with a finite time length, and an impulse determined simultaneously and complementarily. Separate into response coefficient groups.

In this method, a time section having a large coefficient value is detected in the impulse response of the inverse filter, the quantization delay output 6 is determined from the group delay in the section, and the phase of the quantization delay is determined from the impulse response of the original inverse filter. This is realized by removing the rotation.

The delay control circuit 5 receives the quantization delay output 6 as an input, and stores the read address AD of the received signal in the storage circuit (161).
i, 162i to 16ci, 161q, 162q to 16c
q).

Storage circuits (161i, 162i to 16c)
i, 161q, 162q to 16cq) sequentially store the outputs of the back-propagation circuits (151, 152 to 15c) and perform the output according to the read address AD commanded by the delay control circuit 5.

Storage circuits (161i, 162i to 16c)
i, 161q, 162q to 16cq) are dual-port devices capable of simultaneously performing writing and reading in parallel.
A delay is realized by using a RAM (Dual-Port RAM) or the like, and making the address for writing and the address for reading different.

Also, the impulse response of the inverse filter can be applied to the focal point of the transmission of the ultrasonic wave.
The impulse response generating circuit 4 also has an output 8 to the transmitting circuit 12.

The transmission circuit 12 performs digital-to-analog conversion of the impulse response data output 8 calculated for the transmission focal point, and drives the electro-acoustic transducers (101, 102 to 10c) via the transmission / reception separation circuit group 11. Signal.

First, an ultrasonic wave is transmitted assuming that the sound velocity in the subject is constant, an impulse response data output 8 to be corrected is obtained from the received signal, and the corrected ultrasonic wave is transmitted.

The model parameters are optimized again according to the corrected received signal based on the transmission of the ultrasonic wave, and the output of the impulse response data is made more accurate.

After the optimization of the model parameters is completed at an appropriate number of iterations, the corrected phasing addition signal (18i, 18q) is output from the optimized final output (CI, 6, 8) of the impulse response generation circuit 4. Get)

FIG. 2 shows the boundary model generation circuit 1 shown in FIG.
FIG. 2 is a block diagram showing a schematic configuration of the embodiment.

In the boundary model generation circuit 1 shown in FIG. 2, the control circuit 20 generates boundary model parameters according to a built-in control program 21, and divides the obtained boundary curve into s elements.

A space vector R, r between each element and a preset reference point in space and a reception point, and a normal vector n of the element are calculated, and the storage circuit group (231, 232 to 232) is calculated.
23s) to hold the result.

Each of the storage circuit groups (231, 232 to 23s) is composed of six storage circuits (not shown) which can be read in parallel.

The write permission command (WEN), read permission command (REN), and address signal (RWAD) are common between the storage circuit groups (231, 232 to 23s) and the individual storage circuits constituting the same.

The data stored in these individual storage circuits are all generated in advance by the control circuit 20 through calculation, and are written via the bus selection of the data bus selector 22 at the initial stage of the start of correction.

When the processing according to the back propagation operation is started, each of the storage circuit groups (231, 232 to 23s) stores the components (nx1, ny1, rx1, ry) of the vectors R, r, and n.
1, Rx1, Ry1), (nx2, ny2, rx2, r
y2, Rx2, Ry2) to (nxs, nys, rxs,
rys, Rxs, Rys) are output simultaneously according to a common address signal.

The simultaneously output data is temporarily held in the latch circuit group 24, and then is stored in the Fresnel-Kirchh.
The signal is output to the off integration circuit 2.

All of these operations are performed in synchronization with a clock signal (not shown).

FIG. 3 shows the Fresnel-Ki shown in FIG.
FIG. 2 is a block diagram illustrating a schematic configuration of a rchoff integration circuit 2;

In FIG. 3, the frequency independent coefficient calculation unit (3
11, 312 to 31 s) are the outputs (nx1, ny1, rx1, ry1, Rx1, Ry) of the boundary model generation circuit 1.
1), (nx2, ny2, rx2, ry2, Rx2, R
y2) to (nxs, nys, rxs, rys, Rxs,
Rys), the amplitude coefficient ai and the delay τi are calculated.

These are coefficients of the frequency-independent part of (Equation 2) and are calculated according to the following (Equation 5) and (Equation 6).

[0133]

(Equation 5)

[0134]

(Equation 6)

The components corresponding to the angular frequencies (ω1, ω2 to ωu) are calculated for each set of coefficients (ai, τi) for i = 1 to s according to the following equation (Equation 7).

[0136]

(Equation 7)

In the circuit configuration shown in FIG. 3, the frequency dependent calculation units (321, 322 to 32s) and the real part S for each angular frequency ω
ik and an imaginary part signal Sqk are added to adders (331i, 332i).
i to 33ui) and adders (331q, 332q to 3)
3uq).

However, (331i, 332i to 33u)
The subscript i of i) represents the real part, (331q, 332q to 33
The subscript q of uq) means an imaginary part.

FIG. 4 is a block diagram showing a schematic configuration of the inverse filter operation circuit 3 and the impulse response generation circuit 4 shown in FIG.

In the inverse filter operation circuit 3 shown in FIG. 4, the analog-digital converter (141, 141) shown in FIG.
The outputs 411 of 142 to 14c) are simultaneously stored in the waveform buffer 412.

The waveform buffer 412 outputs the time-series received signals to the FFT calculator 413 as simultaneous and parallel outputs.

The parallel complex output of the FFT calculator 413 (W
1, W2-Wu) and Fresnel-Kirchhof
input group (Si1, Sq1) from the f-integration circuit 2, (Si
2, Sq2) to (Siu, Squ) are complex dividers (41
41, 4142 to 414u).

The complex divider (4141, 4142 to 414)
u) is an integer k = 1 to u, j is an imaginary unit, Wk = Wik
+ JWqk as {(Sik) + j (Sqk)} / (W
ik + jWqk), and u parallel complex outputs are input to the evaluation function circuit 7.

After the optimization of the boundary model is completed, Fresnel-Kirch is used to realize the corrected phasing.
input group (Si1, Sq1) from the hoff integration circuit 2,
(Si2, Sq2) to (Siu, Squ) are input to the inverse FFT circuit 421 in the impulse response generation circuit 4, and the inverse FFT circuit 421 is a time window optimization circuit 42
2 is input.

The inverse FFT circuit 421 is constituted by an FFT circuit which receives an input obtained by inverting the sign of the imaginary part component, and the output is stored in the storage circuit 423 in the time window optimizing circuit 422.

The stored time-domain impulse response is output via output 8 to the transmission circuit 12 of FIG.

The impulse response before performing the inverse FFT is also stored in the storage circuit 423 by a branch output (not shown) in the inverse FFT circuit 421.

Here, on the basis of the impulse response in the time domain obtained by the inverse FFT, a quantization delay suitable for realizing the finite time length with high accuracy and an impulse response coefficient group determined simultaneously and complementary are separated. Window optimization circuit 42
Perform within 2.

In this processing, the operation of obtaining the sum of squares of the impulse response coefficients within the time window over a predetermined time width (the number of data points) is performed while moving from the start point to the end point of each impulse response. , The quantization delay output 6 is determined based on the time window position giving the maximum value.

Next, the result of performing an operation excluding the phase rotation corresponding to the quantization delay from the original impulse response in the frequency domain is again passed through the inverse FFT circuit 421 to obtain an impulse response coefficient group CI (CI1 to CIc). Back propagation circuit (15
1, 152 to 15c).

FIG. 5 is a block diagram showing a schematic configuration of the frequency-independent coefficient calculating section (311, 312 to 31s) shown in FIG.

In the frequency independent coefficient calculation unit shown in FIG. 5, the input (rxi, ryi, Rxi, Ryi, n
xi, nyi), output (ai), and output (τi), which are not shown in the other drawings and are reciprocals (1 / _cr , 1 / c) of sound speeds assumed for each layer on the model as inputs common to boundary elements.
c _R ) is input.

The coefficients (ai) and (τi) are output based on the above (Equation 5) and (Equation 6).

Input (rxi, ryi, Rxi, Ryi)
Are multipliers (500, 501, 502, 50
3), and the sum of squares of r and R is calculated by an adder (51).
0,511).

The sum of squares is calculated by the square root reciprocal arithmetic unit (512, 51).
3), and 1 / | r | and 1 / | R | are calculated.

The delay circuits (514, 515, 516, 51
7) delays and outputs (rxi, ryi, Rxi, Ryi) to the multipliers (520, 521, 522, 523), respectively, and synchronizes them with the outputs of the square root reciprocal calculators (512, 513).

Multipliers (520, 521, 522, 52)
3) is (rxi / | r |, ryi / | r |, Rxi /
| R |, Ryi / | R |), and outputs an adder (524,
525) is (rxi / | r | + Rxi / | R |), (r
yi / | r | + Ryi / | R |) is output.

The multiplier (532, 533) is connected to the adder (52).
4, 525) and the result obtained by delaying the inputs nxi, nyi so that they are synchronized by the delay circuits (530, 531), and multiplying the result by the output of the adder (534) (rxi / | r
| + Rxi / | R |) nxi + (ryi / | r | + Ry
i / | R |) nyi.

A square root reciprocal calculator (512, 512)
513) are delayed so as to be synchronized by the delay circuits (540, 541), and the results obtained by sequentially multiplying the results obtained through the logarithmic operators (570, 571) by the multipliers (542, 543) are as follows:
The coefficient ai according to (Equation 5) is obtained.

The output of the square root reciprocal operation unit (512, 513) passes through the reciprocal operation unit (560, 561) and the multiplier (55).
(0, 551), the reciprocals 1 / c _r and 1 / c _R of the sound velocity are multiplied. As a result of adding their outputs by the adder 552, (| r | / _cr + | R | / c _R ) get.

Further, an output τi is obtained through a delay circuit 553 for synchronizing with the output ai.

FIG. 6 is a block diagram showing the frequency dependence calculation unit (3) shown in FIG.
21, 322 to 32 s).

An angular frequency signal group (ω1, ω2 to ωu) generated in advance by a control circuit (not shown) becomes a common input to all the frequency-dependent calculation units (321, 322 to 32s), and the multipliers (611, 612 to s). 61u).

These multiplier outputs are output to the sine-cosine circuit group 6
At 3, a sine function value and a cosine function value are generated in parallel.

The input ai is an angular frequency signal group (ω1, ω2-
ωu) and multipliers (621, 622 to 62u)
Is multiplied by a delay circuit (651, 652-65u)
Synchronizes with the output of the sine-cosine circuit group 63 and the multiplier group 64
Is performed for each angular frequency.

As a result of the multiplication, the real part output group (Ii1, Ii2
ＩIu) and an imaginary part output group (Qi1, Qi2 to Qiu).

FIG. 7 is a block diagram showing a schematic configuration of the evaluation function circuit 7 shown in FIG.

First, the evaluation function defined by (Equation 3) and (Equation 4) is used as an expression for a discrete signal value, and Pa
If the expression of rseval is replaced with the operation in the frequency space and the integration operation is performed for each discrete frequency, the evaluation function ρ ′ of the following expression is obtained.
And ε ′.

[0169]

(Equation 8)

[0170]

(Equation 9)

The structure of the evaluation function operation according to these formulas will be described below.

The operation output for each frequency of the inverse filter operation circuit 3 becomes an input group (ECV) of the evaluation function circuit 7.

Each signal of the illustrated input group (ECV) is a pair of parallel complex signals, and is represented by a single signal line in FIG.

The evaluation function circuit 7 includes a post-addition sum-of-square-root operation unit 97, a post-sum-of-square root addition operation unit 98,
9.

The input group (ECV) is a parallel input of a branched and summed square sum square root operation section 97 and a sum of squared square root addition section 98 after addition.

After the addition, the received signals corresponding to the different electroacoustic transducers in the square-sum-of-squares operation unit 97 are sequentially converted by the inverse filter operation circuit 3 into complex adders (911, 912-91).
u).

These are independently accumulated in accumulation registers (9211, 9212 to 921u) for each real part and imaginary part of the signal.

When the accumulation for all the electroacoustic transducers is completed, the square operation circuit (9311, 9312 to 931)
u), the result of the square operation is added by the parallel adder 951 and then stored in the storage circuit 961 via the square root operation circuit 941.

These operations are performed according to a command from a control circuit 99 (not shown).

In the sum-of-squares square root addition section 98, the received signals corresponding to the different electroacoustic transducers are sequentially converted by the inverse filter calculation circuit 3 into the square calculation circuits (9321, 93).
22 to 932u).

These are added by the parallel adder 952 and then added by the adder 923 and the addition register 922 through the square root operation circuit 942.

When the accumulation for all the electroacoustic transducers is completed, the sum is stored in the storage circuit 962.

The storage circuits (961, 9
The result stored in 62) is read out by the control circuit 99, and the similarity evaluation function and the energy evaluation function are calculated.

The results of these operations are stored in the storage circuit (96
1,962).

The calculation of each evaluation function and the storage of the result are repeated until the evaluation functions for all reference points are completed for all the combinations of the boundary shape parameters.

When the repetition is completed, the optimum combination of the boundary shape parameters is selected by optimizing all the reference point evaluation functions.

In the ultrasonic measuring apparatus according to the present embodiment, an optimum combination of boundary shape parameters is selected by optimizing the evaluation function for all reference points as follows.

First, over the entire measurement area in the subject,
When performing optimization with less bias, standard deviations from the average value of the similarity evaluation functions obtained for all reference points are determined for each combination of boundary shape parameters, and a predetermined number is extracted from the smallest one.

Next, the one with the largest energy evaluation function was selected as the optimum combination of boundary shape parameters.

However, when the total number of reference points set in the subject is small and the spatial density is low, optimization was performed only with the minimum standard deviation.

If the correction result of the measurement based on the optimized result is insufficient, the operator uses input means (not shown) to limit the coordinates of the boundary of the specific small area in the measurement area (not shown). The deviation of the similarity evaluation function at a reference point in the region was multiplied by one or more weights.

That is, the influence of the deviation of the similarity in the specific small area on the standard deviation of all reference points is increased, and priority is given to the evaluation of the focal point in the specific small area which is the area of interest.

The input means (not shown) limited by the operator allows a cursor linked to a trackball to designate coordinates of a diagonal point on a measurement image (B mode image or the like) to be corrected, and determines a rectangular area. The mechanism was realized.

All of these operations are realized by a program 991 in the control circuit 99 and a command signal (not shown).

FIG. 8 shows the back propagation circuit (151 to 151) shown in FIG.
It is a block diagram which shows schematic structure of 15c).

In FIG. 8, input v is a digital input of a received signal.

The digital input of the received signal is sequentially stored in the shift register 1001 in chronological order, and the stored signal length is d.

As another input, the time window optimizing circuit 4 shown in FIG.
22, there is one coefficient group (CIf) which is one of the coefficient groups (CI1 to CIc), and one coefficient group (CIf)
f) is also a total number d of parallel digital inputs.

The output of shift register 1001 and one coefficient group (CIF) are determined by multiplier groups (10021, 10022 to 1022).
02d), and are added by the parallel adder 1003.

The result of the addition is subjected to frequency shift processing for quadrature detection in multipliers (10041, 10042) by cosine signal generation circuit 1005 and sine signal generation circuit 1006 having carrier frequency ωs.

The signal after the frequency shift is subjected to a digital low-pass filter (10
071, 10072) and the quadrature output (1008i, 1
008q).

The digital low-pass filter (1
0071, 10072) are back propagation circuits (151 to 15).
Instead of being provided individually in c), a configuration may be provided after the final addition output (18i, 18q) in FIG.

Further, the phase of the output of the cosine signal generation circuit 1005 and the phase of the output of the sine signal generation circuit 1006 may be different for each back propagation circuit (151 to 15c) or may be synchronized.

However, means (not shown) for compensating for the difference from the phase of the carrier at the receiving focal point position when synchronized is provided.

FIG. 9 shows the boundary model generating circuit 1 shown in FIG.
And Fresnel-Kirchhoff integration circuit 2
FIG. 10 is a block diagram showing another circuit configuration of the embodiment.

Boundary model generation circuit 1 and F shown in FIG.
The resnel-Kirchhoff integration circuit 2 is a circuit for reading out and outputting the frequency response of the inverse transfer function calculated in advance for all of the boundary model to be generated and the reference point group and the reception point group from the storage means.

The control circuit 1101 generates an address RAD for reading an impulse response of an inverse transfer function between a reference point and a reception point corresponding to different boundary model parameters according to a built-in control program 11011.

A CD-ROM drive circuit (11021, 1101) which uses a read-only optical disk as a data storage device.
11022 to 1102 u) serially reads data from a built-in CD-ROM in accordance with the address RAD and the read permission REN, and performs serial-parallel conversion (1).
1031, 11032 to 1103u).

The serial-parallel conversion circuit (11031, 110
32 to 1103u) have a built-in data buffer capable of simultaneously reading and writing the read data, and output data of both the real part and the imaginary part in parallel.

These outputs correspond to the inputs (Si
1, Sq1), (Si2, Sq2) to (Siu, Sq
u).

According to this configuration, it is not necessary to configure the coordinate information of an extremely large number of boundary elements generated for each boundary model by a parallel circuit, so that the apparatus configuration can be significantly simplified.

In this embodiment, the transfer function between the focal point of a plurality of ultrasonic waves transmitted or received and the plurality of electroacoustic transducers provided in the subject is determined based on the optimum boundary shape curve. Although the case where the inverse transfer function based on each is obtained, and a predetermined inverse transfer function is selected from the inverse transfer functions and the convolution signal processing is performed with the received signal, the waveform distortion of the received ultrasonic wave is corrected, Based on the optimal boundary shape curve, inverse transfer functions based on transfer functions between the focal points of a plurality of ultrasonic wave transmission or reception waves provided in the subject and the plurality of electroacoustic transducers are obtained. It is also possible to correct the waveform distortion of the transmitted ultrasonic wave by selecting a predetermined inverse transfer function from them and performing the convolution signal processing with the transmitted signal.

Further, in this embodiment, the sound velocity distribution model is constituted by a two-dimensional model, but the actual sound field is three-dimensional.

Therefore, the model may be constructed in three dimensions, in which case the three-dimensional Fresnel-Kirchh
off integral formula

[0215]

(Equation 10)

It goes without saying that estimation can be performed based on the setting of model parameters, the setting of reference points, and the calculation of transfer functions according to the above.

Although the present invention has been described in detail with reference to the embodiments, it is needless to say that the present invention is not limited to the above-described embodiments and can be variously modified without departing from the gist thereof.

[0218]

The effects obtained by typical ones of the inventions disclosed in the present application will be briefly described as follows.

(1) According to the present invention, based on the optimum boundary shape curve obtained from a plurality of assumed boundary shape curves in advance, the focal point of the transmission or reception of a plurality of ultrasonic waves provided in the subject. Calculate inverse transfer functions based on transfer functions between the transmitter and a plurality of electro-acoustic transducers, select a predetermined inverse transfer function from the transfer functions, and perform convolution signal processing with a transmitted signal or a received signal. Accordingly, it is possible to correct the waveform distortion of the transmitted ultrasonic wave or the received ultrasonic wave caused by the non-uniform sound speed in the subject.

(2) By applying the ultrasonic measuring apparatus of the present invention to a medical ultrasonic diagnostic apparatus, the contrast resolution can be improved mainly at the abdomen and the like of a human body. Accuracy can be improved.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a schematic configuration of an ultrasonic measurement device according to an embodiment of the present invention.

FIG. 2 is a block diagram showing a schematic configuration of a boundary model generation circuit 1 shown in FIG.

FIG. 3 shows Fresnel-Kirchhof shown in FIG.
FIG. 3 is a block diagram illustrating a schematic configuration of an f-integration circuit.

FIG. 4 is a block diagram showing a schematic configuration of an inverse filter operation circuit and an impulse response generation circuit shown in FIG. 1 and a connection relationship with an evaluation function circuit;

FIG. 5 is a block diagram illustrating a schematic configuration of a frequency-independent coefficient calculation unit illustrated in FIG. 3;

FIG. 6 is a block diagram illustrating a schematic configuration of a frequency dependent calculation unit illustrated in FIG. 3;

FIG. 7 is a block diagram illustrating a schematic configuration of an evaluation function circuit illustrated in FIG. 1;

FIG. 8 is a block diagram showing a schematic configuration of a back propagation circuit shown in FIG. 1;

9 shows a boundary model generation circuit and Fres shown in FIG.
FIG. 14 is a block diagram illustrating another circuit configuration of a nel-Kirchhoff integration circuit.

FIG. 10 is a flowchart illustrating a processing procedure of the ultrasonic measurement device according to the present embodiment.

FIG. 11 is a diagram for explaining transfer function calculation based on a Fresnel-Kirchhoff diffraction model.

[Explanation of symbols]

1. Boundary model generation circuit 2. Fresnel-Kir
chhoff integration circuit, 3 ... inverse filter operation circuit, 4 ...
Impulse response generation circuit, 5 delay control circuit, 6 quantization delay output, 7 evaluation function circuit, 8 impulse response data output, 12 transmission circuit, 18i real component addition output, 18q imaginary component Addition output, 20,99,1101
... Control circuit, 21,11011 ... Control program, 22
... data bus selector, 24 ... latch circuit group, 63 ... sine-cosine circuit group, 64 ... multiplier group, 81, T ... reference point,
82 ... layer boundary, 97 ... sum-of-squares-square operation unit after addition, 98
... Sum-of-squares square root addition operation units, 101 to 10 c... Electro-acoustic transducers, 111, 112 to 11 c.
2 to 14c: analog-digital converter, 151, 1
52 to 15c: Back propagation circuit, 161i, 162i to 16
ci, 161q, 162q-16cq, 423,96
1,962... Storage circuit, 171,172,331i, 3
32i to 33ui, 331q, 332q to 33uq, 5
10,511,524,525,534,552,92
3: adder, 231, 232 to 23s: storage circuit group, 3
11, 312 to 31s: frequency independent coefficient calculator, 32
1,322 to 32s: frequency-dependent calculator, 412: waveform buffer, 413: FFT calculator, 4141, 4142
414 u: complex divider, 421: inverse FFT circuit, 42
2. Time window optimization circuit, 500 to 503, 520 to 52
3,532,533,542,543,550,55
1,611,612-61u, 621,622-62
u, 10021, 10022 to 1002d, 1004
1,10042: multiplier, 512, 513: square root reciprocal calculator, 514 to 517, 530, 531, 540, 5
41,553,651,652-65u delay circuit, 5
70,571 ... logarithmic calculator, 560,561 ... reciprocal calculator, 911, 912-91u ... complex adder, 922, 9
211, 9212 to 921u... Cumulative register, 931
1,9312-931u, 9321,9322-932
u: square operation circuit, 951, 952, 1003: parallel adder, 941, 942: square root operation circuit, 1001: shift register, 1005: cosine signal generation circuit, 1006
... Sine signal generation circuit, 10071, 10072 ... Digital low frequency pass filter, 11021, 11022-1
102u ... CD-ROM drive circuit, 11031, 110
32-1103u... Serial-parallel conversion circuit.

──────────────────────────────────────────────────続 き Continuation of the front page (56) References JP-A-63-51846 (JP, A) JP-A-5-42140 (JP, A) JP-A-5-95946 (JP, A) 116162 (JP, A) W. Flax et al, IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, November 1988, vol. 35, no. 6, pp. 758-767 Levin Nock et al, The Journal of the Acoustic Society of America, May 1989, vol. 85, no. 5, p.p. 1819-1833 Dong-Lai Liu et al, The Journal of the Acoustic Society of America, January 1994, vol. 95, no. 1, pp. 542-555 Dong-Lai Liu, The Journal of the Acoustic Society of America, August 1994, vol. 96, no. 2, pt. 1, pp. 649-660 Kazuhiko Hamamoto et al., IEICE Technical Report, October 1994, vol. 94, no. 270 pp. 1-8 (58) Field surveyed (Int. Cl. ⁷ , DB name) A61B 8/00

Claims (7)

(57) [Claims] A plurality of electro-acoustic transducers for transmitting ultrasonic waves into a subject, receiving reflected echoes with the plurality of electro-acoustic transducers, subjecting the received signals to signal processing, In an ultrasonic measurement apparatus that obtains information in a specimen, a plurality of boundary shape curves are assumed when the subject is regarded as a two-layer medium structure having different sound speeds, and a plurality of reference points provided in the subject and the plurality of electroacoustics. First means for obtaining a transfer function between the converter and each of the plurality of boundary shape curves, and further obtaining an inverse transfer function from the obtained transfer function; and The inverse transfer function and the actually obtained received signal in a specific time section are subjected to a convolution operation in the time domain or the frequency domain to obtain a reflection signal for each reference point in relation to the plurality of electroacoustic transducers. Estimating means for estimating at Selecting means for selecting an optimum curve from the plurality of boundary shape curves based on the reflection signal at each reference point estimated by the step, and based on the optimum boundary shape curve selected by the selection means Determine the inverse transfer function between the focal point of the ultrasonic wave transmitted or received and the plurality of electroacoustic transducers provided in the subject respectively, a predetermined one of the determined inverse transfer function An ultrasonic measuring apparatus comprising: signal processing means for performing convolution signal processing of an inverse transfer function and a transmission signal or a reception signal. 2. The method according to claim 1, wherein the first means is Fresnel-K.
2. The ultrasonic measurement apparatus according to claim 1, further comprising means for calculating a transfer function between a plurality of reference points provided in the subject and the plurality of electro-acoustic transducers by irchhoff integration. . 3. The energy evaluation function defined by a square root of a time integration value of a square after adding all of the reflection signal groups estimated at the same reference point, the selection means includes: The similarity evaluation function determined by dividing the energy evaluation function by the sum of the square roots of the values is obtained at all the reference points, and the optimization is performed on all the reference point evaluation functions using the evaluation function value group as a variable. 3. An ultrasonic measuring apparatus according to claim 1, wherein said means is a means for selecting an optimum curve from the boundary shape curves. 4. The method according to claim 3, wherein the total reference point evaluation function is a standard deviation from an average value of the similarity evaluation function values, and optimization is performed with a minimum of the standard deviation. Sound wave measurement device. 5. A method according to claim 1, further comprising the step of changing a spatial density of a reference point, which is a variable of the reference point evaluation function, and a weight coefficient by which a deviation of each reference point is multiplied. The ultrasonic measurement device according to claim 3 or claim 4. 6. The method according to claim 1, wherein the signal processing unit is configured to determine, based on the selected boundary shape curve, a focal point between a plurality of ultrasonic wave transmission or reception waves provided in the subject and the plurality of electroacoustic transducers. Of the transfer function of Fresnel-Kirchhoff
Each is calculated by integration, and a quantization delay operation and an impulse response coefficient group determined for each inverse transfer function obtained from the calculated transfer function and a convolution operation in the time domain of the transmission signal or the reception signal are performed. The ultrasonic measuring apparatus according to any one of claims 1 to 5, wherein the ultrasonic measuring apparatus is a unit. 7. The method according to claim 1, wherein the first means calculates an inverse transfer function between a plurality of reference points provided in the subject and the plurality of electroacoustic transducers, for each of the plurality of boundary shape curves. The ultrasonic measurement according to any one of claims 1 to 6, wherein the calculation result is stored in a storage unit, and the inverse transfer function is sequentially read from the storage unit. apparatus.