JP5069022B2 - Method and system for accurate time delay estimation for use in ultrasound imaging - Google Patents

Description

  The present invention relates generally to imaging systems, and more specifically to a method and system for estimating and correcting time delays in an ultrasound imaging system.

  The ultrasound system comprises an array of transducer elements that are used to transmit a set of waveforms into the object to be imaged and to receive a reflected set of ultrasound signals. Each waveform is delivered with a relative time delay to focus the net transmit waveform to have the desired shape at the desired direction and depth. Similarly, each received signal is individually delayed so that the response of the system to reflected energy having the desired shape for the desired direction and depth is maximized. The delayed received signals are added and processed to create and display an image to be imaged.

  This transmission and reception time delay (collectively referred to as beamforming time delay) is typically calculated on the assumption that sound waves propagate through the body at a known constant velocity. If this assumption is not satisfied, the convergence of transmission and reception will be degraded, and image resolution and contrast will be reduced.

  One way to reduce the degradation in image quality is to adjust the beamforming time delay based on a measure of the relative time delay of the received signal. These relative time delays are conveniently measured after applying the receive beamforming delay. If the known and constant sound speed assumption is correct, the delayed received signals will be better aligned in time, i.e. the arrival time error will be smaller. If this assumption is not correct, the time alignment of the delayed received signals is deteriorated, that is, the arrival time error is increased. Correcting the beamforming delay for the arrival time error improves its focusing and improves image resolution and contrast. The arrival time delay error can be estimated using one of several methods well known in the art.

  In medical ultrasound imaging, estimation of arrival time errors must be fast, accurate and reliable. Furthermore, it is highly desirable to minimize the additional cost required to implement the estimated hardware. As used herein, arrival time error is defined as the difference between two signals. This arrival time error is processed to obtain a time delay correction value which is then applied to correct the beamforming time delay.

  It is desirable that the estimation be fast because the beamforming time delay needs to be updated quickly. This is because the required correction value varies as the transducer moves relative to the imaged object (transducer movement over this patient must be performed by the operator as part of the normal scanning procedure). Or may be due to slight movement of the operator's hand or patient movement or breathing).

  In order to improve image resolution and contrast and to avoid undesired degradation of the image due to beamforming time delay adjustment using incorrect time delay correction values, it is desirable that the estimation be accurate. The estimated arrival time error may be inaccurate for several reasons. For example, if the arrival time error estimate is calculated using the phase of the complex correlation sum, the correlation of the signal contributing to this correlation sum is not good, or an element in the transducer fails and its output signal becomes abnormal It can be noisy. Some transducer elements can generate a noisy signal because they are concealed from the object being imaged by acoustically opaque obstacles such as ribs, which can cause inaccuracies in arrival time error estimates. . It would be undesirable to allow the use of such unreliable or noisy phase estimates to determine time delay correction values. This is because the degradation of beamforming performance due to such inaccurate values can outweigh the benefits of correction using more accurate values.

  In addition, these errors in arrival time error estimates may introduce artifacts in the image, which can lead to incorrect diagnosis and prolonged examination times. In order for the majority of operators to use the time delay correction function routinely, thereby benefiting from improved image resolution and contrast, this artifact rate must be sufficiently small.

In many applications, it is necessary to image between the ribs of the human body (by the ribs), but this imaging (especially so that the desired imaging scan plane points the transducer in a direction perpendicular to the overall direction of the ribs). Transmission and reception of ultrasound from a portion of the transducer (if required) can be difficult because the ribs can be blocked. Furthermore, the sheet-like muscles associated with the ribs are irregular in thickness and orientation, which results in arrival time errors at the transducer location. In order to enable more accurate diagnosis, it is desirable to create a high-quality image while imaging with intercostality.
US Pat. No. 5,388,461 (corresponding Japanese published patent publication H08-507951) US Pat. No. 5,570,691 (corresponding Japanese published patent publication H10-507101) US early publication patent 2006-0004287 (corresponding Japanese published patent publication 2006-015138) Xu et al. Time Delay Estimation Using Wavelet Transform for Pulsed-Wave Ultrasound. Annals of Biomedical Engineering. 23: p. 612-621. 1995.

  Accordingly, there is a need for a method and system for accurately and reliably estimating and compensating for arrival time errors in ultrasound systems while minimizing system cost and size.

  Briefly, in one aspect of the invention, a method for correcting beamforming time delay in an ultrasound system is provided. The method includes transmitting an ultrasonic energy beam within the subject. The ultrasonic energy beam is generated using a transducer element array, and each transducer element is configured to transmit a pulse of ultrasonic energy with a transmission beam forming time delay. The method further includes the steps of configuring each transducer element to receive an ultrasonic energy beam with a receive beamforming time delay and receiving a plurality of echo signals, and an arrival time for each echo signal and each imaging direction. Estimating the error. The method further includes correcting transmission and reception beamforming time delays and creating an ultrasound image of the object using the corrected transmission and reception beamforming time delays.

  In another embodiment, an ultrasound system for estimating beamforming time delay is provided. The ultrasound system is a transducer array having a set of array elements arranged in a pattern, each of which transmits an ultrasound energy beam into the object during a transmit mode and a receive mode. A transducer array operable separately to generate an echo signal in response to vibrational energy striking the transducer. The ultrasound system is coupled to a transducer array and transmits separate transmit signal pulses with respective transmit beamforming time delays during transmit mode so that a directional transmit beam is generated. A transmitter operable to apply to each is included. The receiver is coupled to the transducer array and samples the echo signal generated by each of the array elements during the receive mode so that a corresponding plurality of received signals are generated, and for each echo signal sample And is operable to add a receive beamforming time delay. The system further includes a beamformer system configured to estimate an arrival time error for each echo signal and each imaging direction and correct the transmit and receive beamforming time delays, and to create an ultrasound image. And a configured image processor.

  For these features, aspects and advantages of the present invention, as well as other features, aspects and advantages, read the following detailed description with reference to the accompanying drawings, wherein like reference numerals represent like parts throughout the drawings. Will deepen your understanding.

FIG. 1 is a block diagram of one embodiment of an ultrasound system 10 implemented in accordance with an aspect of the present invention. The ultrasound system comprises a collection subsystem 12 and a processing subsystem 14. The acquisition subsystem 12 includes a transducer array 18 (consisting of a plurality of transducer array elements 18A-18Z), a transmit / receive switching circuit 20, a transmitter 22, a receiver 24, a beamformer system 26, Is provided. The processing subsystem 14 includes a control processor 28, a demodulator 30, an image mode processor 32, a scan converter 34, and a display processor 36. The display processor is further coupled to a monitor for displaying images. User interface 40 is interfaced with control processor 28 and display monitor 38. This processing subsystem may be further coupled to a remote connection subsystem 42 that includes a web server 44 and a remote connection interface 46. The processing subsystem may further be coupled with a data repository 48 to receive ultrasound image data. This data repository interacts with the image workstation 50.

  As used herein, “operable to”, “configured to” and other expressions shall work together to provide each component with the described effect. Refers to the hardware or software connection between the components to be enabled, and these terms further include analog formulas programmed to perform procedures that provide output in response to a given input signal. It also refers to the operational functions of electrical components such as digital computers, application specific devices (eg application specific integrated circuits (ASICs)).

  This architecture and module may be a dedicated hardware element such as a circuit board having a digital signal processor, or may be software running on a general purpose computer or processor such as a commercially available PC. The various architectures and modules can be combined and separated according to various embodiments of the invention.

  In the collection subsystem 12, the transducer array 18 is in contact with the object 16. The transducer array is coupled to a transmit / receive (T / R) switching circuit 20. T / R switching circuit 20 is coupled to the output of transmitter 22 and the input of receiver 24. The output of receiver 24 is an input to beamformer 26. Beamformer 26 is further coupled to the input of transmitter 22 and the input of demodulator 30.

In the processing subsystem 14, the output of the demodulator 30 is coupled to the input of the image mode processor 32. The control processor interfaces with the image mode processor 32, the scan converter 34 and the display processor 36. The output of image mode processor 32 is coupled to the input of scan converter 34. The output of scan converter 34 is coupled to the input of display processor 36. The output of display processor 36 is coupled to monitor 38.

  The ultrasound system 10 transmits ultrasound energy within a selected region of the object 16 and receives and processes the echo signals backscattered from the object to create and display an image.

  To create a transmit beam of ultrasonic energy, the control processor 28 generates transmit parameters that cause the beamformer 26 to generate a beam of the desired shape at a desired steering angle from a point on the surface of the transducer array 18. The command data to be sent is sent. This transmission parameter is sent from the beamformer 26 to the transmitter 22. The transmitter 22 uses this transmission parameter to properly encode a transmission signal to be sent to the transducer array 18 via the T / R switching circuit 20. This transmitted signal is set to a certain level and time delay with respect to each other and provided to the individual transducer elements of the transducer array 18. This transmitted signal excites the transducer element to emit ultrasonic waves having the same time delay and level relationship. As a result, when the transducer array 18 is acoustically coupled to an object using, for example, an ultrasonic gel, a transmission beam of ultrasonic energy is formed in the object inside the scanning plane along the scanning line. This process is called electronic scanning.

  Transducer array 18 is a bidirectional transducer. When ultrasonic waves are transmitted into the object, the ultrasonic waves are backscattered by the tissue and blood sample inside the object. The transducer array 18 receives this backscattered echo signal at various times depending on the distance to the tissue that returned the sound wave and the angle of the position where the sound wave was returned to the surface of the transducer array 18. The transducer element is responsive to the backscattered echo signal and converts the ultrasonic energy from the backscattered echo signal into an electrical signal.

  This received electrical signal reaches the receiver 24 via the T / R switching circuit 20. Receiver 24 amplifies and digitizes the received signal and provides other functions such as gain compensation. The digitized received signal corresponds to backscattered sound waves received at various times by each transducer element, and retains the amplitude and arrival time information of the backscattered sound waves.

  This digitized received signal is sent to the beamformer system 26. The control processor 28 sends command data to the beamformer system 26. Beamformer system 26 forms a receive beam emanating from a point on the surface of transducer array 18 at a steering angle that typically corresponds to the steering angle of the previous ultrasound beam transmitted along the scan line. This command data is used for

  The beamformer system 26 performs time delay, amplitude weighting and addition in accordance with command data instructions from the control processor 28 to generate a received beam signal corresponding to the sample volume along the scan line within the scan plane within the object. By doing so, an operation for an appropriate received signal is performed. The beamformer system further includes an aberration algorithm that adjusts the time delay and amplitude weights to correct errors introduced by the aberrant tissue layer. It is also possible to correct the aberration by correcting the waveform itself.

This received beam signal is sent to the processing subsystem 14. The demodulator 30 demodulates the received beam signal, and a pair of I and Q demodulated data values corresponding to the sample volume inside the scanning plane is generated. The demodulated data is transferred to an image mode processor 32 that is configured to create an image . The image mode processor 32 uses the parameter estimation technique to create imaging parameter values from the demodulated data in scan sequence format. The imaging parameters may include parameters corresponding to various imaging modes that are possible, such as B mode, M mode, color velocity mode, spectral Doppler mode, tissue velocity imaging mode, and the like. The imaging parameter value is sent to the scan converter 34. The scan converter 34 processes the parameter data by converting from a scan sequence format to a display format. This conversion includes performing an interpolation operation on the parameter data and creating display pixel data in a display format.

  In addition, the image processor uses image processing algorithms to detect desired features in the image. These features detected by the image processing algorithm are then used to modify the method for calculating or implementing the beamforming time delay. For example, the image processor may mask out correction values for an area, or may select from various estimation techniques based on information derived from the image. In one specific embodiment, an iterative aberration correction algorithm is implemented to quickly calculate the beamforming time delay. The image is divided into several regions, starting with the first region and firing the first beam in the image. During subsequent region launches, data from this first region is collected and processed. These techniques allow time for processing before launching the beam in the next frame.

  The scan converted pixel data is sent to the display processor 36 for performing any arbitrary spatial or temporal filtering on the scan converted pixel data, applying grayscale or hue to the scan converted pixel data, and displaying on the monitor 38. For this purpose, the digital pixel data is converted into analog data. User 40 interacts with beamformer system 26 based on the data displayed on monitor 38.

  As described above, beamformer system 26 performs a time delay operation on the received signal. The manner in which the beamformer system estimates and corrects the beamforming time delay in the received signal is described in further detail with reference to FIG.

FIG. 2 is a block diagram of one embodiment of an adaptive beamformer system 26 . The beamformer system 26 is shown as receiving received signals from the transducer elements 18A-18Z of the transducer array 18 via the multiplexer 27. These transducer elements are also used to transmit ultrasonic energy within a selected area of interest. Each block of the beamformer system 26 is described in further detail below.

  The beamforming delay 62 includes beamformer delay elements 62A-62Z. Each delay element introduces a delay into the received signal received from the corresponding transducer element 18A-18Z. The time-delayed received signal is provided to adder 64 to generate a sum of the time-delayed received signal.

  The sum of the time delayed received signals is provided to a complex filter 68 to generate a complex beam sum signal. This complex beam sum signal is provided to the correlator processor 70 as shown in FIG. Correlator processor 70 includes a plurality of correlator processors 70A-70Z. Each correlator processor receives the beam sum signal and the delayed received signal from the delay elements 62A to 62Z.

  The output of each correlator processor is a complex number called a complex correlation sum. The phase of the complex correlation sum is proportional to the estimated time delay between each received signal and the beam sum signal.

  The correlation sum from each correlator processor corresponds to one beamforming channel and imaging scan line beam. The correlation sum from each correlator processor is provided to the correlation sum processor 74. This correlation sum can be expressed as:

（式１）
上式において、Ｂ^＊（ｒ）はビーム和信号の複素共役を意味しており、またｓ（ｒ）はチャンネル信号である。「Ｂ」と「ｓ」はその両方をベースバンド信号若しくは解析信号とすることがあり、あるいは「Ｂ」をベースバンド若しくは解析信号としかつ「ｓ」を実信号とすることがある。この和は相関レンジサンプル「ｒ１」から「ｒ２」にわたって計算される。

(Formula 1)
In the above equation, B ^* (r) means the complex conjugate of the beam sum signal, and s (r) is the channel signal. “B” and “s” may both be baseband signals or analytic signals, or “B” may be baseband or analytic signals and “s” may be real signals. This sum is calculated over correlation range samples “r1” to “r2”.

  The correlation sum processor may further receive two other signals as inputs and use them to normalize the correlation sum. In one embodiment, the input signal is a square value of the magnitude of the beam sum signal and a square value of the magnitude of the channel signal. The square value of the magnitude of the beam sum signal is added over the correlation range samples, which is expressed as:

（式２）
同様に、チャンネル信号の大きさの２乗値が相関レンジサンプルにわたって加算されており、これは次式で表される。

(Formula 2)
Similarly, the square value of the magnitude of the channel signal is added over the correlation range samples, which is expressed as:

（式３）
代替的な一実施形態では、ビーム和信号の大きさの総和とチャンネル信号の大きさの総和が相関和プロセッサ７４に提供される。ビーム和信号の大きさの総和は次式で表される。

(Formula 3)
In an alternative embodiment, the sum of beam sum signal magnitudes and the sum of channel signal magnitudes are provided to the correlation sum processor 74. The total sum of the beam sum signals is expressed by the following equation.

（式４）
同様に、チャンネル信号の大きさの総和は次式で表される。

(Formula 4)
Similarly, the sum of the magnitudes of channel signals is expressed by the following equation.

（式５）
相関和プロセッサは、各ビーム形成チャンネル及びビームごとに上述の入力信号を用いて１組のビーム形成時間遅延補正値を生成する。この時間遅延補正値は次いで、ビーム形成時間遅延に適用される。

(Formula 5)
The correlation sum processor generates a set of beamforming time delay correction values using the above input signal for each beamforming channel and beam. This time delay correction value is then applied to the beamforming time delay.

  FIG. 3 is a flow diagram illustrating one way in which the correlation sum processor 74 generates beamforming delays. Each step of the method is described in further detail below.

  In step 78, the correlation sum processor calculates a normalized correlation sum for some or all of the beamforming channel and the imaging scanline beam. The normalized correlation sum “C” is expressed as the following (Equation 6).

（式６）
（式６）は「Ｂ」と「ｓ」の両方がベースバンド信号である場合、並びに「Ｂ」と「ｓ」の両方が解析信号である場合に適用される。「Ｃ」の大きさは０と１の間の範囲にある。「Ｃ」の大きさは「Ｂ」が「ｓ」に比例するときに１である。「Ｂ」がベースバンド若しくは解析信号であり、かつ「ｓ」が実信号であるときは、「Ｃ」の大きさはゼロと２の平方根の逆数値との間の範囲にある。別の実施形態では、その正規化相関和「Ｃ」は次の（式７）のように表される。

(Formula 6)
(Equation 6) is applied when both “B” and “s” are baseband signals, and when both “B” and “s” are analysis signals. The magnitude of “C” is in the range between 0 and 1. The size of “C” is 1 when “B” is proportional to “s”. When “B” is a baseband or analytic signal and “s” is a real signal, the magnitude of “C” is in the range between zero and the reciprocal value of the square root of 2. In another embodiment, the normalized correlation sum “C” is expressed as (Equation 7) below.

（式７）
上式において、「Ｎ」は総和を計算する際のレンジサンプル数である。（式７）は、（式６）を用いて計算される正規化相関和に対する近似値であり、この方がディジタルハードウェアでの計算をより簡単にすることができる。（式６）はＮ個のサンプルｘ_１、ｘ_２、．．．、ｘ_Ｎに対する標準偏差の定義である

(Formula 7)
In the above equation, “N” is the number of range samples when calculating the sum. (Expression 7) is an approximate value for the normalized correlation sum calculated using (Expression 6), and this can make the calculation in digital hardware easier. (Equation 6) expresses N samples x ₁ , x ₂ ,. . . , X _N is the definition of standard deviation

を用いて（式７）に変換することが可能であり、これは次の（式８）のように整理することができる。

Can be converted into (Equation 7), which can be organized as in (Equation 8) below.

（式８）
上式において、μはＮ個のサンプルｘ_ｉの平均値

(Formula 8)
In the above equation, μ is the average value of N samples x _i

である。
（式８）内の係数

It is.
Coefficient in (Equation 8)

は所与の統計分布に関する定数である。ｘ＝｜ｓ｜を用いると、（式８）によって（式６）が（式７）に変換される（実信号とスペックル様の複素信号のいずれかの振幅を記述する統計分布に関する次数１の定数係数は無視しており、また以下で記載する相関和処理において関心があるのは相関和の相対的ノルム（ｒｅｌａｔｉｖｅ ｎｏｒｍ）だけであるため棄却することができる）。

Is a constant for a given statistical distribution. Using x = | s |, (Equation 6) is transformed into (Equation 7) by (Equation 8) (degree 1 regarding a statistical distribution describing the amplitude of either a real signal or a speckle-like complex signal. The constant coefficient is ignored, and since only the relative norm of the correlation sum is of interest in the correlation sum processing described below, it can be rejected).

  In step 80, normalized correlation sums are mapped to ranks with respect to beam and transducer elements. In many systems, the number of elements in the transducer array is greater than the number of beamforming channels. For example, a 1D linear and curvilinear transducer array may have 192 elements, but is typically connected to an ultrasound system with 128 beamforming channels. The connection between the channel and the element is realized through a set of programmable multiplexing switches for selecting a subset of transducer elements for each beam direction.

  The elements in a multiple row transducer array with a row of “nRow” and a column of “nCol” have a row index represented by “row = 0, 1, to (nRow−1)” and “col = 0. 1 to (nCol-1) ”. Alternatively, the elements in transducer array 18 are labeled with the element number “el”, where el is equal to “col” + “row” × “nCol”.

  In step 82, the mapped correlation sum is modified to minimize the effect of unreliable time delay estimates. An unreliable time delay estimate is defined as an estimate that can be identified as likely to be incorrect or unusually noisy. The correlation sum processor constructs a mask denoted “mask [el, bm]” that labels whether the correlation sum is reliable or unreliable for each element and beam. By identifying an element whose correlation sum phase is an unreliable arrival time error estimate, the robustness and accuracy of the actually estimated arrival time error is improved. The manner in which the correlation sum is labeled as reliable or unreliable is described in further detail below with reference to FIG.

  Continuing with step 82 of FIG. 3, all items in the mask are first set to zero (this zero is an arbitrary value selected to identify a reliable correlation sum). For each unreliable correlation sum, “mask [el, bm]” is set to 1 which is the second arbitrary value. Furthermore, when the number of unreliable elements for a given beam exceeds a specified threshold, all items in “mask” for that beam are set to 1 (this is illustrated in FIG. 7). More details will be described). Finally, for each item that is 1 in “mask”, its corresponding correlation sum is modified by setting its phase to zero while leaving the amplitude unchanged as shown in (Equation 9). The

（式９）
上式において、Ｃは修正を受ける複素相関和であり、またＣ’は修正後の相関和である。

(Formula 9)
In the above equation, C is a complex correlation sum to be corrected, and C ′ is a corrected correlation sum.

  In step 84, the real and imaginary parts of the modified correlation sum are filtered. In one embodiment, a one-dimensional, real, symmetric low-pass filter applied across the element index and another one-dimensional, real, symmetric low-pass filter applied across the beam index are used. Is done. The lengths of these filters are selected so that the spatial dispersion of the correlation sum phase is not suppressed more than necessary, and the correlation sum phase dispersion can be reduced. In one embodiment, a filter with trigonometric coefficients is used. A triangular filter is an example of a filter whose frequency response does not become negative. For the stable operation of the time delay correction algorithm, the sign of the filter's spatial frequency response across its elements and beam indices must not be changed. This algorithm behaves as one system with negative feedback that corrects the beamforming time delay to force the arrival time error to zero. If the sign of one of the filter's spatial frequency responses changes, the feedback switches from negative to positive as the sign changes, and positive feedback suppresses arrival time errors at several spatial frequencies for the algorithm It will be amplified instead. In one embodiment, the width of the triangular filter is 13 beams for the beam filter and 5 elements for the element filter.

  In step 86, the phase of the filtered correlation sum is calculated and then converted to a time delay correction value. This time delay correction value is obtained by dividing the phase of the correlation sum by a factor 2πf (where “f” is the nominal center frequency of the received ultrasound signal).

  The advantage of filtering the complex correlation sum rather than the arrival time error is that the arrival time error estimate corresponds to a phase change greater than ± π (ie, the arrival time error is greater than ± 1 / (2f). The accuracy of the arrival time error estimate is greatly improved. In such a case, the phase of the correlation sum in the range from −π to + π is “wrapped” and jumps from a value close to + π to −π and in the opposite direction as shown in FIG.

  4, 5 and 6 are graphs displaying each transducer element (on the “x-axis”) of the transducer array along with its respective correlation sum phase (on the “y-axis”). The solid line represents the ideal correlation sum corresponding to a more smoothly varying arrival time error compared to 1 / (2f) for some elements as seen in the region near 18E and the region near 18M-18Q. It is a phase. The solid circle in FIG. 4 is the phase of this ideal correlation sum after adding a small amount of noise. The solid circle in FIG. 5 is the result of the low-pass filter for the phase of the noisy correlation sum. As can be seen from FIG. 5, the approximation obtained for the desired phase (solid line) in the vicinity of the region where the phase is folded back is not good. FIG. 6 shows the result of performing a low-pass filter on the noisy correlation sum and subsequently calculating the phase of the correlation sum. As can be seen from FIG. 6, the agreement with the true phase (solid line) is significantly improved.

  Continuing with FIG. 3, in step 88, beam steering terms for the up and down and azimuth directions are estimated and removed from the time delay correction values. If the steering term is removed to minimize the geometric distortion of the image, this can lead to misdiagnosis (eg, when the size of some objects in the image is important). Furthermore, for the constraint that the average time delay correction value is zero to minimize the shift of the correction beam within the range, and for the receive beam that is dynamically focused with the shift to the focus depth of the transmit beam minimized. Another constraint may be imposed on the time delay, such as the absence of parabolic terms in the time delay correction to minimize the added focusing shift.

  In step 90, time delay correction values are mapped from element order to channel order. This time delay correction value is provided to the beam forming delays 62A-62Z as shown in FIG. In one embodiment, the time delay correction value is added to each acoustic frame.

  As described in step 82 of FIG. 3, the correlation sum processor is configured to label each correlation sum as reliable or unreliable. FIG. 7 is a flowchart showing a method of determining the reliability of each correlation sum by the correlation sum processor.

  It is assumed that the source of arrival time error (that is, the aberration introducing layer) fluctuates smoothly in space. Thus, a true time delay correction value (proportional to the phase of the observed correlation sum) is less variable across the transducer for a given beam, and less in the beam direction for a given element in the transducer. Can be considered small. Conversely, an unreliable element is an element whose phase variation is not small throughout the transducer or in the beam direction for a given transducer element. In a two-dimensional transducer that is subdivided into elements that span the entire vertical and azimuth dimensions of the transducer, phase smoothness will generally be evaluated across both dimensions of the transducer in addition to the beam direction. In some situations, it may be appropriate to evaluate the phase smoothness only in one direction of the transducer dimensions. This has the advantage that the complexity of the hardware or software for calculating the derivative and the filter described below is reduced. In the following, the interpretation of the term “element direction” may mean a single transducer dimension in the simplified embodiment, or both dimension directions in the general embodiment.

  In step 92, the phase of the correlation sum is calculated. An approximation is calculated for the phase derivative in the element direction and the derivative in the beam direction. In one embodiment, this approximation is a discrete derivative using the nearest neighbor difference. In step 94, a discrete derivative of the phase is calculated across the element index. In step 96, a discrete derivative of the phase is calculated across the beam index. In step 98, the absolute value of the discrete derivative of the phase in the element direction and the absolute value of the discrete derivative of the phase in the beam direction are added for each element and beam.

In step 100, the sum of the absolute values of the discrete derivatives is smoothed by low pass filtering in the element direction. In step 102, the sum of the absolute values of the derivatives is smoothed by low pass filtering in the beam direction. This filtering process is performed in order to reduce the fluctuations introduced by the process of taking the derivative that tends to increase the noise. In one embodiment, neighboring value pairs are added. The output of the two point low pass filter can be arranged to compensate for the half-sample shift produced by the two point discrete derivative.

  In step 104, the filtered sum of the phase derivatives is compared to a user-specified first threshold value. In one embodiment, the first threshold is about 5 radians. For each item of the filtered sum that is greater than the first threshold, the corresponding item in the mask is set to 1 and the correlation sum is marked as unreliable for the element and beam. .

  Similarly, if the correlation sum marked in the previous step as unreliable for a given beam direction exceeds a second threshold, the correlation sum for all elements in that beam direction is unreliable. Marked. This second threshold is similarly specified by the user. In one embodiment, the second threshold is half the number of beamforming channels. Identifying beams where a significant number of their estimated time delay correction values are unreliable, even if the image contains acoustic shadows due to poor contact with disturbing ribs or transducer objects The introduction of artifacts is avoided. In such a situation, the transmit beam and the receive beam are considerably degraded. In these beams, since the reference signal becomes a deteriorated or distorted beam sum signal, there is a high possibility that the correlation sum for all of the elements is not reliable. If most of the elements in the effective aperture are identified as having an unreliable correlation sum, using the remaining correlation sum as a correction value often leads to image degradation rather than image improvement.

  The beamforming arrival time error can also be labeled as unreliable using an image processing algorithm. In one embodiment, the image is processed to detect a “signature” associated with the defect correction estimate. As used herein, the term “signature” means any identifiable feature associated with the reliability or accuracy of the associated beamforming time delay estimate within the image. These “signatures” may take various forms, such as local statistical parameters, tissue characteristics, and position relative to the anatomy.

  FIG. 8 is a flow diagram illustrating a method for identifying unreliable beamforming arrival time errors using image data. The illustrated method estimates the reliability of the time delay correction value using local statistics. In step 105, statistical parameters for the region of interest in the image are estimated using the beam sum or image data. Statistical parameters include mean brightness, standard deviation, higher moments, parameters related to the shape of the distribution, and a goodness index that quantifies how well a particular statistical distribution describes actual data (But not limited to).

  In step 106, statistical parameters are filtered over the entire image or time period to improve statistical parameter estimation. Alternatively, statistical parameters can be estimated over a larger area. The size of the estimation region or the amount of filtering is calculated based on the need to locate the parameter and the resulting quality of the estimate. In step 107, the filtered statistical parameters are used to label regions that do not have a set of statistical properties that help to estimate the time delay correction value. In step 108, the labeled region is excluded from use in the time delay estimate. In step 109, a check is performed to ensure that there are no small isolated regions for which the estimate is rejected or retained. If such regions exist, these regions are eliminated by local interpolation to avoid introducing new artifacts. The labeling for these areas can be incorporated into the mask described in step 104 of FIG.

  It is known to those skilled in the art that certain time delay estimation algorithms work best in the region of fully grown speckle. The amplitude of the complex signal from the speckle region has a statistical distribution called a Rayleigh distribution. It is possible to determine whether a set of signal amplitudes is reasonably well described by a Rayleigh distribution. As used herein, a region of interest (ROI) refers to a set of beams and range samples for which the correlation sum is calculated. If the signal amplitude in the ROI is not properly described by the Rayleigh distribution, the ROI can be moved or the ROI size can be changed until the statistical distribution is approximately the Rayleigh distribution. If no region containing an approximate Rayleigh distribution of signal amplitudes can be found for a small set of adjacent beams, its arrival time error should be interpolated with respect to the value set using the surrounding arrival time error estimates. Is also possible. In addition to the size and location of the ROI, it is possible to adjust other parameters used in arrival time estimation, such as the phase derivative threshold.

  As mentioned above, tissue types can be identified using various tissue characterization techniques. If a tissue type has more useful properties for estimating arrival time errors, the region of interest for estimating arrival time errors can be moved to this region based on statistical information from the image. . An example of such a tissue type is the liver because it is common for the liver to include a large area of speckle-like scatterers. Similarly, if a particular tissue type is known to be unsuitable for estimating the arrival time error, it can be detected in the image, but the corresponding incorrect arrival time error estimate need not be applied. . One example of a tissue type that is not speckle-like and therefore not suitable for many estimation methods for arrival time errors is the diaphragm.

A blood vessel or other echo-free region is an example in which a significant signature exists in an image. FIG. 9 is an image of a region representing the region of interest 120 and the blood vessel 122 . The size of the blood vessel 122 is such that all or most of the ROI 120 is inside the blood vessel. The echo from the blood is much smaller than the echo for the surrounding tissue 124. As a result, the sonic energy from the ROI 120 that provides an estimate of the arrival time error is quite small. The surrounding tissue 124 reflects energy in the form of transmitted beam side lobes, and the resulting signal is larger than the signal reflected from echoless blood. Accordingly, the surrounding tissue 124 serves as the off-axis of the “high brightness” target, and the time delay correction estimate is used to steer the corrected beam toward the edge of the blood vessel or to the side of the transmit and receive beams. There is a tendency to grow robes. In such cases, the image is used to determine whether the ROI 120 is in a blood vessel or an echo-free region. Using this information from the image, it is possible to create an algorithm that avoids artifacts due to sidelobe growth and steering toward the edge of the blood vessel.

FIG. 10 is a flowchart describing an algorithm for detecting whether a region of interest is located near or within a blood vessel. In step 110, the average signal amplitude “A _i ” is calculated for each beam over its region of interest. This average amplitude can be an average value of logarithmically compressed amplitude data or an average value of linear amplitude data. This average value can be calculated from the scan-converted data or can be calculated from the raw data before the scan conversion. In step 111, a region average “L _i ” over “M” beams can be calculated for each of “N” beams in the image. The area average value can be a simple average of the “A _i ” values, or it can be a space weighted average value. The length “M” of the averaged area may be a function of the image data, and may be fixed or selected by the user.

In step 112, a region of interest for a beam is considered to be inside the vessel if “A _i ” is less than L _i by at least some specified threshold. Such a beam is labeled “dark” (since the echoless area is typically displayed using black pixels). This threshold is set so that the beam located at the center of the blood vessel is labeled “dark”. However, in this way, the edge of the blood vessel is not a sharp transition from light to dark in the image (especially the uncorrected image), so the beam is inside the blood vessel but is Often, beams in the vicinity are not labeled “dark”. To avoid artifacts in these situations, neighboring beams around the labeled area are also labeled. In step 113, the front “n” beams and the rear “n” beams of the beam identified as “dark” are also labeled “dark”. “N” is a compromise between reducing the effectiveness of corrections in regions that do not have an incorrect arrival time estimate and avoiding artifacts caused by an incorrect estimated arrival time error. Selected. In step 114, the beam labeled "dark" is incorporated into the mask described in step 104 of FIG. 7 which labels the correlation sum as reliable or unreliable.

  Another example of a feature in an image that tends to generate an incorrect arrival time error estimate is a very bright target, ie a strong reflector of sound waves. FIG. 11 shows the high-intensity scatterer 126 located just outside the ROI 128. An example of a high-intensity scatterer is a diaphragm that is typically much brighter than the surrounding tissue. If the high-intensity scatterer is located just outside the ROI, the sound waves associated with the side lobes of the transmitted beam will be reflected by this high-intensity scatterer and make a significant contribution to the received signal from the ROI. The sidelobe acoustic energy 132 reflected from the high intensity target 126 is almost as large as the sound wave 130 reflected from the tissue 128 in the main beam path. In some cases, the sidelobe energy is greater than the sound wave reflected from the region of interest.

  In one embodiment, if a high-intensity target is detected in the image, the estimated value of the aberrator near the high-intensity scatterer is ignored, and the correction value for that region is unaffected around Interpolated from the estimated value. In another embodiment, the ROI size range is increased to reduce the effects of high intensity scatterers.

  Image processing algorithms may be used to detect discontinuities in images that appear after applying correction values. Similarly, image processing algorithms can detect areas of the image that have improved brightness, steeper boundaries, and a more appropriate statistical distribution, and label these corrections as more likely to be correct. Is possible. By quantifying the reliability of the estimates within each region, it is possible to use a weighting filter for the correlation sum across each region, which contributes to the contribution from the unreliable correlation sum compared to the reliable correlation sum. Will be reduced.

  Images may be used to determine and quantify tissue movement relative to the transducer. As the transducer or tissue moves, the arrival time error pattern shifts across the transducer. If there is a significant time interval between the measurement of the arrival time error pattern and the correction of the beamforming time delay, this shift prevents the beamforming time delay correction value from being applied correctly for that beamforming channel, This can result in sub-optimal image improvements. By estimating the shift of the arrival time error pattern, the shift can be compensated.

  The above-described invention has several advantages, including improved accuracy obtained by normalizing the correlation sum before filtering. The amplitude of the complex correlation sum is a measure of the similarity between the two signals being correlated. Normalizing the correlation sum prior to the filtering process places a greater weight on the reliable correlation sum than the unreliable correlation sum. Without normalization, for example, the contribution to the element filtered correlation sum by an element that is very large but noisy, but the signal is small but the time delay estimate for the element is reliable It becomes larger compared to neighboring elements. Similarly, the contribution to the filtered correlation sum by a beam that is unusually high in luminance but noisy is also greater than in neighboring beams where the time delay estimate is reliable.

  This method minimizes the contribution of “dead” or extremely noisy transducer elements by normalizing the correlation sum, which reduces the correlation between the noisy channel signal and the beam sum signal. Derived from. Thus, additional processing or hardware required for detection and compensation for insensitive and noisy transducer elements or for insensitive and noisy system channels is minimized.

  The method described above further uses a simple estimator for the noisy phase to strictly ignore the noisy phase region, thereby introducing image artifacts without requiring complex and expensive hardware to compute. Is minimized. FIG. 12 shows the noisy phase estimation region and data mask as an example, where the data mask is the noise between elements 18I and 18Q that are zero to indicate that the estimate is reliable. Two of the elements within the active region have not been properly identified. FIG. 13 shows the result of substituting the correlation sum for the elements marked as unreliable (ie, the data mask is 1) with the amplitude of the correlation sum and subsequently filtering the correlation sum. Two of the noisy correlation sums are not modified, but the phase of the filtered and modified correlation sum is close to zero over the entire noisy region as desired.

  The present invention performs a smooth interpolation between a beam for which a reliable arrival time estimate is available and a beam whose arrival estimate is unreliable, thereby obstructing the image between these regions. The introduction of boundaries and discontinuities is avoided. FIG. 14 is a diagram showing the correlation sum phase regarding a certain element as a function of the beam number. The beams in the vicinity of 18I to 18Q are marked as unreliable in this example, and the phase of the correlation sum is set to zero to maintain the amplitude. The phase after the filter processing is shown in FIG. The phase drops smoothly to zero near the left boundary of the noisy region and increases smoothly from zero near the left boundary of the noisy region. The same smooth interpolation is performed on all elements of these beams, which smooths the transition on the image between the corrected and uncorrected beams.

Although only certain features of the invention have been illustrated and described herein, many modifications and changes will occur to those skilled in the art. Accordingly, it is to be understood that the appended claims are intended to cover all such modifications and changes as fall within the true spirit of the invention.

1 is a block diagram of one embodiment of an ultrasound system implemented in accordance with an aspect of the present invention. 1 is a block diagram of one embodiment of a beamformer system according to one aspect of the invention. FIG. 6 is a flowchart illustrating a method for estimating an arrival time error and generating a time delay correction value. It is a graph showing a comparison between a transducer element and the phase of its respective complex correlation sum. It is a graph showing a comparison between a transducer element and the phase of its respective complex correlation sum. It is a graph showing a comparison between a transducer element and the phase of its respective complex correlation sum. 6 is a flow diagram illustrating one method for labeling complex correlation sums. 6 is a flow diagram illustrating one method for labeling correlation sums using image data. It is an image of the tissue showing the existence of a blood vessel in the vicinity of the region of interest. 6 is a flowchart illustrating a method for detecting the presence of a blood vessel in the vicinity of a region of interest. It is an image of the tissue showing the presence of a high-intensity scatterer in the vicinity of the region of interest. FIG. 5 is a graph representing a complex correlation sum phase and corresponding data mask for transducer elements in a transducer array. FIG. 5 is a graph representing a complex correlation sum phase and corresponding data mask for transducer elements in a transducer array. FIG. 5 is a graph representing a complex correlation sum phase and corresponding data mask for transducer elements in a transducer array. FIG. 5 is a graph representing a complex correlation sum phase and corresponding data mask for transducer elements in a transducer array.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Ultrasonic system 12 Acquisition subsystem 14 Processing subsystem 16 Imaging object 18 Transducer array 18A-18Z Transducer element 20 T / R switching circuit 22 Transmitter 24 Receiver 26 Beam former system 28 Control processor 30 Demodulator 32 Image mode Processor 34 Scan Converter 36 Display Processor 38 Monitor 40 User Interface 42 Remote Connection Subsystem 44 Web Server 46 Interface 48 Data Repository 50 Imaging Workstation 62 Beamformer Delay 62A-62Z Beamformer Delay Element 64 Adder 68 Complex Filter 70 Correlator processor 70A-70Z Correlator processor 74 Correlation sum processor 120 Region of interest 122 Blood vessel 124 Tissue 126 High intensity Runtime 128 region of interest 132 side lobe energy

Claims (8)

A method for correcting a beamforming time delay of an ultrasound system (10) comprising:
Transmitting an ultrasonic energy beam through an object using an array of transducer elements (18A-18Z), wherein each of the transducer elements transmits an ultrasonic energy beam with a transmit beam time delay. A configured transmission process;
Configuring each transducer element to receive an ultrasonic energy beam with a receive beamforming time delay and receiving a plurality of echo signals;
Estimating a beam forming time delay error for each echo signal and each imaging direction;
Correcting transmit and receive beamforming time delays by generating a complex correlation sum representing a time delay between each received echo signal and a sum of one or more received echo signals;
Creating an ultrasound image of the object using the corrected transmit and receive beamforming time delays;
Including
The correction step further includes
Calculating a normalized correlation sum;
Mapping the normalized correlation sum to an imaging direction and a transducer element direction;
Correcting the correlation sum;
Filtering the corrected correlation sum across the imaging direction and the entire transmitting element to create a filtered correlation sum;
Converting the phase of the filtered correlation sum into a corresponding time delay correction;
Correcting beam forming time delay by the time delay correction;
including,
Method. The correction step includes
Labeling each mapped correlation sum as a reliable or unreliable correlation sum;
Setting each correlation sum labeled untrusted to an absolute value;
The method of claim 1 comprising: The labeling step includes
Calculating the phase of each mapped correlation sum corresponding to each transducer element and imaging direction;
Determining a continuously varying pattern within the calculated phase for each transducer element and imaging direction, wherein the correlation sum is labeled as reliable when the continuously varying pattern is present. A determination process;
The method of claim 2 comprising: The labeling step further includes
Comparing the sum of absolute values of phase differentiation in the element direction and the sum of absolute values of phase differentiation in the imaging direction with a first threshold;
Labeling the correlation sum as unreliable when the calculated sum of absolute values of the phase derivatives exceeds a first threshold;
Labeling all of the mapped correlation sums for the corresponding imaging direction as unreliable when the amount of unreliable correlation sum in each imaging direction exceeds a second threshold;
The method of claim 2 comprising:   The labeling step includes identifying a signature in the image using an image processing algorithm and labeling whether the beamformer time delay error is reliable or unreliable based on the identified signature. The method of claim 2 comprising. The image processing algorithm is:
Identifying multiple regions of interest in the image,
Calculating statistical parameters for each identified region of interest;
Applying the statistical parameter and labeling whether the estimated beamformer time delay error is reliable or unreliable;
The method of claim 5, wherein the method is configured to perform. An ultrasound system (10) for estimating beamforming time delay comprising:
A transducer array (18) having a set of array elements arranged in a pattern, each element transmitting an ultrasonic energy beam into the object during transmit mode and between receive modes. A transducer array (18) operable separately to generate an echo signal in response to vibrational energy striking the transducer;
A separate transmit signal pulse is applied to each of the array elements during transmission mode, coupled to the transducer array and having a respective transmit beamforming time delay so that a directional transmit beam is generated. A transmitter (22) operable to:
A sample of echo signals generated by each of the array elements during a reception mode is coupled to the transducer array and corresponding to each echo signal sample so as to generate a corresponding plurality of received signals. A receiver (24) operable to add a receive beamforming time delay;
A beamformer system (26) configured to estimate an arrival time error for each echo signal and each imaging direction to correct the transmit and receive beamforming time delays;
An image processor (32) configured to create an ultrasound image;
With
The beamformer system (26) further includes
Generating a correlation sum for each system beamforming channel and imaging direction;
Calculating a normalized correlation sum;
Mapping the normalized correlation sum to imaging direction and transducer element order;
Modifying the mapped correlation sum;
Filtering the modified correlation sum to generate a filtered correlation sum;
Calculating the phase of the filtered correlation sum;
Converting the phase of the filtered correlation sum into a correction time delay;
Correcting the transmit and receive beamforming time delays using the correction time delays;
Is configured to run,
Ultrasound system (10). 8. The system of claim 7, wherein the beamformer system is configured to label each mapped correlation sum as a reliable or unreliable correlation sum.

Images