JP4903928B2 - Three-dimensional flow imaging method and apparatus using coded excitation - Google Patents

Description

【００７２】
［Ｓ_H ］を算出するのに用いられる最終的なスケーリング・ファクタは、縦横スケーリング・ファクタと回転スケーリング・ファクタとの積である。
【００７３】
Ｘ_s ＝Ｘ_asＸ_rs
Ｙ_s ＝Ｙ_asＹ_rs
Ｚ_s ＝１．０
これらのスケーリング・ファクタの組合せを用いて、データ空間内でＸ軸、Ｙ軸及びＺ軸に沿って１単位ずつ離隔している各ボクセルが、画像平面上でＸ軸及びＹ軸に沿って最大で１単位ずつ離隔しているピクセルへ確実に写像されるようにする。
【００７４】
超音波イメージング・システムのオペレータが、データ・ボリュームを通る２次元スライスを現在のＸ，Ｙ，Ｚ回転配置方向で見たいという場合には、オペレータは「表示モード」キーを押し（図１０及び図１１の工程１２８）、「ボリューム回転」モードから「切断平面」（フォーマット設定し直し）モードに切り換える。図１５を参照して説明すると、配向ボックス１６２の色が変わり（例えば、緑から白へ）、色付き（例えば、緑）の多角形１７８が配向ボックス内のデータ・ボリュームの中央に現れる（工程１３０）。緑の色は、ボックス内のこの部分を動かすことができることを示している。次いで、システムは、データ・セットの中央を通る２次元スライスの現在の配置方向における初期表現１８０を作成する（工程１３２）。
【００７５】
このスライス１８０の形状は、配向ボックス１６２内の緑の多角形１７８の形状と整合する。次いで、オペレータは、トラックボールを用いて、データ・セットを通してスクロールすることができる。トラックボールが動いているとシステムが判定した場合（工程１３４）、このデータ・セットの相次ぐ２次元スライスが、選択された配置方向及び位置で実時間で表示される。トラックボールが動いていない場合（工程１３４）に「表示モード」キーが押されていれば（工程１３６）、システムは「切断平面」モードを出て、「切断平面回転」モードに入る。
【００７６】
「切断平面回転」モードでは、配向ボックス内の緑の多角形が移動し（工程１３８）、データ・ボリューム内でのスライスの位置を視覚的に指示すると共に、表示パネルが更新されて、スライスの位置を全ボリュームの百分率として示す（即ち、５０％はデータ・ボリュームの中央に相当する）。データ・ボリュームを通るスライスの位置を指示する緑の多角形が白に変わり、配向ボックスが緑に変わって、スライスがこの時点で固定され、且つデータ・ボリュームをスライスに対して回転させ得ることを示す。次いで、システムは、フォーマット設定し直されたスライスを前述の方式で表示する（工程１４０）。使用者はこの時点で、トラックボールを動かすことにより、配向ボックスを新たなＸ，Ｙ，Ｚ配置方向に定めることができる（工程１４２）。その際、使用者がトラックボールを動かすにつれて、配向ボックスはトラックボールの動きに追従し、白の多角形の形状が変化して、スライスの形状を示す。配置方向が設定されたとき、使用者は「計算」キーを押し（工程１４４）、システムが、配向ボックスに対する多角形の角度によって指示される角度で捉えた新たな２次元スライスを表示するようにする。「表示モード」キーが押されていなければ（工程１４６）、オペレータは、データ・ボリュームの配置方向を再設定することができる。「表示モード」キーが押されていれば、システムは「ボリューム回転」モードへ戻る。
【００７７】
２次元データの３次元投影を生成するアルゴリズムは周知であり、データをフォーマット設定し直してデータ・セットを通る任意のスライスを生成する手法も周知である。図７に示されているように、投影後又はフォーマット設定し直し後のデータは、ホスト・コンピュータによってＸＹ表示メモリ７０に出力される。次いで、投影後又はフォーマット設定し直し後のデータの画像フレームは、ビデオ・プロセッサ６２へ送られ、シネ・メモリ６６によって取得される。ビデオ・プロセッサは、配向ボックス及び他の図形記号を投影後のフォーマット設定し直されたデータの画像フレームに重ね合わせて、表示モニタに出力する。
【００７８】
Ｂモード・フィードスルー・パラメータを調節することにより、使用者が二次元及び三次元投影で画像化された流れに対して通常のＢモード背景画像の変化量を見ることができるようにしてもよい。造影剤は表示画像内の流れ情報を強調するために被検体に注入することができる。
【００７９】
以上、本発明の好ましい特徴を例示し説明してきたが、当業者には様々な修正及び変更が考えられよう。例えば、本発明は二相符号を使用することに制限されず、多相符号もまた使用することができる。従って、特許請求の範囲は本発明の真の精神に含まれるこのようなすべての修正及び変更を包含するものであることを理解されたい。
【図面の簡単な説明】
【図１】従来の超音波イメージング・システムの様々なサブシステムを全体的に示すブロック図である。
【図２】本発明の好ましい一実施例による超音波イメージング・システムを示すブロック図である。
【図３】本発明の好ましい実施例の一例による非バーカー型の単一回送信の符号化励起［１，１，１，−１，−１，１，−１，１］と共に用いるのに適した１６タップの不整合フィルタのフィルタ係数を示すグラフである。
【図４】図５に示す１６タップの不整合フィルタのフィルタ係数が符号［１，１，１，−１，−１，１，−１，１］と畳み込みされたときに、該フィルタよって発生される復号後の信号を示すグラフである。
【図５】本発明の好ましい実施例によるＢモード信号のフィードスルーによる「スロー・タイム」のウォール・フィルタ処理を示す流れ図である。
【図６】本発明のもう１つの好ましい実施例による超音波イメージング・システムを示すブロック図である。
【図７】本発明の好ましい実施例によるピクセル強度データの相次ぐ三次元投影を含んでいるフレームを再構成する手段を示すブロック図である。
【図８】超音波プローブの走査平面に垂直な方向に超音波プローブをリニア走査することにより取得されるデータ・ボリュームを示す線図である。
【図９】図８に示すデータ・ボリュームをフォーマット設定し直すことにより得られる任意角度における個々のスライスを示す線図である。
【図１０】フロー・イメージング・データを取得し表示する手順を示す流れ図である。
【図１１】フロー・イメージング・データを取得し表示する手順を示す流れ図である。
【図１２】使用者が画定した関心ボリュームを指示する図形と共にセクタ走査方式の超音波画像の表示を示す線図である。
【図１３】データ・ボリュームの配置方向を表示する図形と共に任意の角度で形成された投影後の超音波画像の表示を示す線図である。
【図１４】本発明の好ましい実施態様で有利に使用される投影変換手順を示す線図である。
【図１５】データ・ボリュームに対するスライスの位置及び配置方向を表す図形と共に２次元スライスの表示を示す線図である。[0001]
BACKGROUND OF THE INVENTION
The present invention relates generally to ultrasound imaging systems. Specifically, the present invention relates to a three-dimensional imaging method and apparatus for blood flow and contrast agents.
[0002]
BACKGROUND OF THE INVENTION
Conventional ultrasound scanners produce a grayscale (B-mode) image that represents a thin slice through a region of anatomy. Also, flow information can be extracted by detecting a Doppler shift between the transmitted ultrasonic pulse and the returned ultrasonic pulse. Conventional ultrasound systems display flow as either average Doppler power (power Doppler imaging) or average velocity (commonly referred to as “color flow imaging”). The flow image can be displayed separately from the B-mode image or as an overlay on the B-mode image.
[0003]
Pulses transmitted with power Doppler and color flow imaging are typically narrower than B-mode pulses in order to obtain Doppler sensitivity. When operating on a packet consisting of as many as 16 transmissions, a high-pass wall filter first blocks echoes from slower moving tissue or vessel walls, resulting in a dynamic range of the signal. Reduce. The number of wall filter output samples per packet is given by (N−W + 1). Here, N is the packet size, and W is the length of the wall filter. Subsequently, the instantaneous Doppler power is calculated as the square of the magnitude of each wall filter quadrature output signal, and the average Doppler power is obtained from the average of all quadrature output signals. Alternatively, the average speed is calculated from the wall filter quadrature output signal based on the Doppler principle (phase change) or time delay. The average flow velocity can be estimated using Kasai's autocorrelation algorithm or time domain cross-correlation algorithm.
[0004]
Traditional color flow imaging has very good flow sensitivity, but the ability to view physical flow is limited in its limited dynamic range (partially dependent on the compression curve) Limited by limited resolution (due to narrowband pulses), limited frame rate (due to large packet size), and axial only flow sensitivity (determined by dependence on Doppler effect) The In addition, conventional color flow imaging produces artifacts such as aliasing and color bleeding.
[0005]
If the ultrasound probe is swept across a predetermined area of the subject and the slice is stored in memory, it is possible to obtain a three-dimensional data volume that includes both B-mode and flow information. This data volume can be used to project a three-dimensional view of the region of interest. The quality of the 3D projection of color flow and power Doppler data (separately or in combination with the B mode) is affected by the above reasons: reduced sonic frame rate, reduced Doppler sensitivity, flow resolution And "flash" artifacts. Therefore, there is a need for a method for performing three-dimensional flow imaging that is not affected by the drawbacks associated with Doppler flow imaging.
[0006]
A conventional ultrasonic image is formed from a combination of a fundamental signal component and a harmonic signal component, and the latter signal component is generated in a tissue or a nonlinear medium such as a blood flow containing a contrast agent. In some cases, the ultrasound image may be improved by suppressing the fundamental signal component and enhancing the harmonic signal component.
[0007]
Medical ultrasound contrast agents have been developed to aid in the diagnosis of anatomical structures, including the vasculature, which were previously difficult to image. The contrast agent is typically a microbubble having a diameter in the range of 1 μm to 10 μm and is injected into the bloodstream. Since the backscatter signal of the microbubble is much larger than the backscatter signal of the blood cell, the microbubble is used as a marker that enables blood flow imaging. One method for further separating the echoes from these contrast agents is to use the high (or low) harmonic component of the contrast agent echo, which is a harmonic of the surrounding tissue where no contrast agent is present. Much bigger than ingredients. Flow imaging of high (or low) harmonic signals is often done at frequency f₀A narrowband signal of frequency 2f₀(Second harmonic) or frequency f₀/ 2 (subharmonic) is received in the band, and then the conventional color flow processing is performed. This approach has all the limitations of conventional color flow systems: low resolution, low frame rate and axial only flow sensitivity.
[0008]
  In medical diagnostic ultrasound imaging, it is also desirable to optimize the signal-to-noise ratio (SNR). With increased SNR, increased transparency at a given imaging frequency can be obtained, or resolution can be improved by assisting ultrasound imaging at higher frequencies. Coded excitation is a well-known radar technique used to increase the signal-to-noise ratio in situations where the peak power of the transmitted signal cannot be increased, but the average power can be increased. . This situation is often found in medical ultrasound imaging, where system design constraints define the peak amplitude of the signal driving the transducer. In this situation, longer signals can be used to deliver higher average power values, and temporal resolution is restored by correlating the returned signal with a matched filter. It is preferable to use a binary code, i.e. a code that can easily be represented digitally as a series of numbers +1, -1 or 0. Binary codes are also preferred because they contain the maximum energy for a given peak amplitude and pulse duration.
[Patent Document 1]
US-issued patent No. 427685
[Patent Document 2]
US-issued patent 5329929
[Patent Document 3]
Japanese published patent No. 05-228145
[Patent Document 4]
US-issued patent No. 5456257
[Patent Document 5]
US-issued patent 5632277
[Patent Document 6]
US Patent No. 5,706,819
[Patent Document 7]
US Patent No. 5724976
[Patent Document 8]
Japanese Published Patent No. 2008-182680
[Patent Document 9]
US Patent No. 5833613
[Patent Document 10]
US Patent No. 5934288
[Patent Document 11]
US Patent No. 5938611
[Patent Document 12]
U.S. Patent No. 5961463
[Patent Document 13]
US Patent No. 5964706
[Patent Document 14]
Japanese open patent Heisei 11-309145
[Patent Document 15]
US Patent No. 5980459
[Patent Document 16]
Japanese open patent Heisei 11-342129
[Patent Document 17]
US-issued patent No. 5984869
[Patent Document 18]
Japanese published patent 2000-060855
[Patent Document 19]
US Patent No. 6030344
[Patent Document 20]
US Pat. No. 6,042,545
[Patent Document 21]
European Published Patent No. 770352
[Non-Patent Document 1]
Torp et al., "Comparison Between Cross-Correlation and Auto-Correlation Technique in Color Flow Imaging," Proc. 1993 IEEE Ultrason. Symp., Pp. 1039-1042.
[Non-Patent Document 2]
Thomas et al., "An improved Wall Filter for Flow Imaging of Low Velocity Flow," Porc. 1994 IEEE Ultrason. Symp., Pp. 1701-1704.
[Non-Patent Document 3]
Hein et al., "Current Time-Domain Methods for Assessing Tissue Motion by Analysis from Reflected Ultrasound Echoes-A Review," IEEE Trans. Ultrason., Ferroelec., Frequ. Contr., Vol. 40, No. 2, Mar. 1993, pp. 84-102. (19 pages)
[Non-Patent Document 4]
de Jong et al., "Characteristics of Contrast Agents and 2D Imaging," 1996 IEEE Ultrasonics Symposium, pp. 1449-1458.
[Non-Patent Document 5]
Ishihara et al., "Path Lines in Blood Flow Using High Speed Digital Subtraction Echography," Proc. 1992 IEEE Ultrason. Symp., Pp. 1277-1280.
[Non-Patent Document 6]
Ishihara et al., "High-Speed Digital Subtraction Echography: etc.," Proc. 1990 IEEE Ultrason. Symp., Pp. 1473-1476.
[Non-Patent Document 7]
de Jong et al., "Principles and Recent Developments in Ultrasound Contrast Agents," Ultrasonics, vol. 29, pp. 324-330 (1991).
[Non-Patent Document 8]
Newhouse et al., "Second Harmonic Doppler Ultrasonic Blood Perfusion Measurement," Proc. 1992 IEEE Ultrason. Symp., Pp. 1175-1177.
[Non-patent document 9]
Burns et al., "Harmonic Power Mode Doppler Using Microbubble Contrast Agents: etc.," Proc. IEEE 1994 Ultrason. Symp., Pp. 1547-1550.
[Non-Patent Document 10]
Averkiou et al., "A New Imaging Technique Based on the Nonlinear Properties of Tissue," Proc. 1997 IEEE Ultrason. Symp.
[Non-Patent Document 11]
Uhlendrf, "Physics of Ultrasound Contrast Imaging: etc.," IEEE Trans. Ultrason. Ferroele. & Freq. Control, vol. 41, No. 1, pp. 70-79, Jan. (1994).
[Non-Patent Document 12]
Shahnazi et al., "Use of Contrast Agents in Ultrasound," Proc. 1994 Ultrason. Symp., Pp. 85-87.
[Non-Patent Document 13]
Wang et al., "Contrast Medium Assisted Fluid / Flow Measurement," IEEE Trans. Ultrason. Ultrasonic. Ferroele. & Freq. Control, vol. 42, No. 2, pp. 309-315, Mar. (1995). 7 pages)
[0009]
SUMMARY OF THE INVENTION
A method and apparatus for three-dimensional imaging of blood flow and contrast agents with high resolution, high frame rate, high SNR / dynamic range, and flow sensitivity in all directions is provided for complex blood circulation, stenosis. Provides clinical benefits including clear visualization of residual lumens and thrombus movement. High resolution is achieved by using broadband pulses, and high frame rate is achieved by using small packet sizes. A high SNR / dynamic range is maintained by using coded excitation. The flow sensitivity in the range direction is maximized and obtained from the RF decorrelation between pulses, while the flow sensitivity in the range cross direction allows a group of reflectors (eg blood or contrast agent) to cross the beam profile. Therefore, it is obtained from the amplitude uncorrelation between pulses.
[0010]
In the disclosed method, a small packet of encoded wideband pulses is transmitted to a single transmit focal position at a given pulse repetition interval. The backscattered signal from this series of firings is filtered with a slow time to remove echoes from stationary or slow moving reflectors along the transmission path. . Small packet sizes (less than 6 firings) to achieve high frame rates, but with the undesirable side effect of reduced SNR. In the preferred embodiment, the SNR is recovered by using coded excitation. Slow time filtering is preferably performed by a high pass FIR (Finite Impulse Response) or IIR (Infinite Impulse Response) wall filter. The wall filter increases the flow signal to clutter ratio, which can be further increased by administering a contrast agent to the patient prior to imaging. A slice of flow imaging data is acquired by scanning the transmit focal position across a plane.
[0011]
The above scanning process is repeated for each of a number of scanning planes. Image frames are acquired for each of a number of scan planes while sweeping, or moving, the subject that scans the ultrasound probe. Multiple image frames of flow imaging data acquired during the probe sweep form a three-dimensional data volume. The projection image can be reconstructed on various imaging planes using any of a variety of volume rendering techniques (eg, maximum or average pixel projection, synthesis or surface). Specifically, the host computer projects pixel flow imaging data in the data volume onto a plurality of rotated image planes. The projected data resulting from each projection can then be displayed.
[0012]
According to a preferred embodiment of the invention, the encoded broadband pulse (centered around the fundamental frequency) sequence is transmitted many times to a specific transmission focal position. Each encoded sequence constitutes one firing. At the time of reception, the received pulse acquired at each firing is compressed and bandpass filtered to isolate, for example, the compressed pulse centered on the fundamental frequency. The compressed and separated signal is then high-pass filtered across each firing using a wall filter. The wall filtered signal is used to image the blood flow without contrast agent injected into the blood.
[0013]
According to another preferred embodiment of the present invention, a contrast agent such as a gas-filled microbubble is injected into the blood and used as a marker for imaging the blood flow. Similar to the embodiment described above, the encoded wideband pulse sequence is transmitted a number of times to a specific transmit focal position. From the interaction between the transmitted ultrasound pulse and the propagation medium, in particular the injected contrast agent, a fundamental signal and a high (or low) harmonic signal are generated. At the time of reception, the received signal is decoded (decoded) and subjected to band pass filter processing to separate the basic signal. The separated fundamental signal is then high pass filtered over multiple firings using a wall filter.
[0014]
As a result of this wall filtering, a fundamental signal reflected from a non-stationary tissue or region of flow located along the transmission path can be extracted. The resulting flow signal can be displayed superimposed on the normal B-mode imaging data vector, ie, added to the B-mode imaging data vector. The received energy at the fundamental frequency that may have contributed to unwanted stationary tissue signals is suppressed by the wall filter.
[0015]
In accordance with yet another preferred embodiment of the present invention, the flow image is either a fundamental signal component or a high (or low) harmonic signal component to provide anatomical landmarks (landmarks). Can be combined with stationary tissue (ie, B-mode) images acquired by detecting by addition or as an overlay. The advantage of the overlay is that it can be realized in color so that the flow area clearly stands out. However, this method requires more complex display hardware and requires more firing (to obtain imaging data representing stationary tissue) than is used for flow imaging. In addition, flash artifacts become noticeable. Introduction by adding background B-mode images (coherent or incoherent) further reduces flash artifacts. By using the wall filter feedthrough method, you can add B-mode images with additional firing (in the case of harmonic B-mode images), or without additional firing. A B-mode image can also be added (in the case of a basic B-mode image). When separate firings are used for flow mode and B mode, the firings may be interleaved to separately threshold the flow signal from the B mode signal prior to display.
[0016]
The method of the present invention differs from conventional power Doppler imaging and color flow velocity imaging in several respects. For example, the method of the present invention does not extract a Doppler signal, thus allowing the use of broadband pulses. Also, the method of the present invention does not require baseband data (baseband data can be used) and does not require speed estimation. Still further, the method of the present invention uses a wall filter with a single output sample to display the quantity directly (without averaging). This is different from conventional power Doppler imaging and color flow velocity imaging. In addition, the summing method that provides the B-mode background is simpler (requires less display hardware) and more efficient than the traditional overlay method (the feedthrough method is the basic B-mode). It does not require additional transmissions to acquire images), is more effective (flash artifacts are lessened).
[0017]
Preferred Embodiment of the Invention
FIG. 1 shows a conventional ultrasonic imaging system. The system includes a transducer array 10 having a plurality of transducer elements 12 that are driven separately. Each transducer element generates a burst of ultrasonic energy when energized by a pulse waveform generated by transmitter 14. The ultrasonic energy reflected from the subject and returned to the transducer array 10 is converted into an analog electric signal by each transducer element 12 for reception, and is sent through a pair of transmission / reception (T / R) switches 18. Applied separately to the receiver 16. The transmitter 14 and receiver 16 operate in response to commands supplied by an operator via an operator interface (not shown) under the control of a host computer or master controller 20. A complete scan is performed by acquiring a series of echoes, at which time the transmitter 14 is momentarily turned on to energize each transducer element 12 and each transducer element 12 The echo signal subsequently generated by 12 is applied to the receiver 16. The receiver 16 converts the analog echo signal to a digital signal and sums each digital signal derived from each transducer element to produce a single beam summed signal, which is used to One line of the image displayed by the display subsystem 22 is formed.
[0018]
Under the direction of the host computer 20, the transmitter 14 drives the transducer array 10 so that the ultrasonic energy is transmitted as a focused beam with directionality. In order to achieve this, the transmission beamformer 26 provides respective time delays to the multiple pulsars 24. The host computer 20 determines the conditions under which the sonic pulse is transmitted. With this information, the transmit beamformer 26 determines the timing and amplitude for each transmit pulse to be generated by the pulser 24. The amplitude of each transmission pulse is generated by an apodization generation circuit 36 such as a high voltage controller that sets a power supply voltage to each pulser. The pulser 24 then sends a transmit pulse through the T / R switch 18 to each of the elements 12 of the transducer array 10. T / R switch 18 protects time gain compensation (TGC) amplifier 28 from high voltages that may be present in the transducer array. The apodization weight (weighting) is selected to achieve an optimal tradeoff between transmit power and sidelobe level. The weight is generated by an apodization generation circuit 36. The apodization generation circuit 36 may include a set of digital-to-analog converters that obtain weighted data from the transmit beamformer 26 and apply this data to the pulser 24. A directional and focused transmit beam combining multiple ultrasound waves transmitted by individual transducer elements by adjusting transmit focus time delay appropriately in a conventional manner and adjusting transmit apodization weights. Can be formed. The apodization weights and transmit focus time delay can be set by the host computer based on system programming and operator input.
[0019]
Each burst of ultrasonic energy is reflected from objects located in successive ranges (distances) along each transmitted beam. The resulting echo signal is sensed separately by each transducer element 12, and a sample of the magnitude of the echo signal at a particular time point represents the amount of reflection that occurred in a particular range. Since there is a difference in the propagation path between the reflection point and each transducer element 12, the echo signals are not detected simultaneously and the amplitudes of the echo signals are not equal. The receiver 16 amplifies a separate echo signal via a respective TGC amplifier 28 provided in each reception channel. TGC is performed by increasing or decreasing the gain as a function of depth. The amount of amplification provided by the TGC amplifier is controlled by a TGC circuit (not shown), which is set by manual operation of the host computer and potentiometer. The amplified echo signal is then supplied to the receive beamformer 30.
[0020]
Under the command of the host computer 20, the reception beamformer 30 tracks the direction of the transmitted beam. The receive beamformer 30 applies an appropriate time delay and receive apodization weight to each amplified echo signal, adds these signals, and reflects from a point located in a specific range in a specific transmit direction. An echo signal is generated that accurately represents the total ultrasonic energy. The receive focusing time delay is calculated in real time using dedicated hardware or read from a lookup table. The receive channel also has circuit components for filtering received pulses. The receive apodization weight and receive focus time delay may be set by the host computer based on system programming and operator input.
[0021]
The time delayed received signals are then summed and provided to a signal processor or detector 32, which converts the summed received signal into display data. In a typical grayscale display, the display data is an envelope of a signal that has undergone some additional processing such as edge enhancement and logarithmic compression. In the case of RF data, the envelope can be detected using a low-pass filter, and in the case of baseband data, (I²+ Q²)^1/2The envelope can be detected using an envelope detector that generates a signal representing. Here, I is an in-phase signal component of baseband data, and Q is a quadrature signal component.
[0022]
A scan converter 34 receives display data from the detector 32 and converts this data into an image desired for display. Specifically, the scan converter 34 converts the sonic image data from a polar coordinate (R × θ) sector format or a Cartesian coordinate linear array format to a suitably scaled Cartesian video rate. Convert coordinates to display pixel data. Next, the scan-converted sound wave data is supplied to be displayed by the display subsystem 22, and the display subsystem 22 images the amplitude of the signal envelope as time-converted as a gray scale. Each scan line is displayed for each transmit beam.
[0023]
FIG. 2 illustrates an ultrasonic flow imaging system according to a preferred embodiment of the present invention. Each transducer element in the transmit aperture is fed N times (where N is preferably 6 or more) using the same encoded waveform by feeding the same transmit sequence N times from the memory 38 to each pulsar. Is also provided). The pulsar 24 drives the elements 12 of the transducer array 10 so that the generated ultrasonic energy is directional or steered as a beam for each transmit firing. To accomplish this, a transmission focus time delay 36 is provided for each pulse waveform generated by the pulser in response to a transmission sequence from memory 38. By appropriately adjusting the transmission focusing time delay in a conventional manner, the ultrasonic beam can be focused at a desired transmission focal position. N pulses are transmitted to this transmit focal position at a predetermined pulse repetition interval (PRI). The transmission sequence is supplied by the host computer based on system programming and operator input.
[0024]
The basic concept of single excitation coding excitation involves modulating a specially designed code sequence based on a transmission burst (base sequence) of length P. An encoded pulse sequence consisting of n bursts is often referred to as an n-chip code. The encoded pulse sequence has a length of n × P, allowing a larger sonic dose or a shorter basic sequence to be used to interrogate the flowing blood. The output signal of the decoding (ie compression) filter is a compressed signal pulse having a length equal to or close to the original transmission burst length P, but the amplitude of this signal pulse is n times longer. Generated by the generated pulse sequence. This encoding technique does not adversely affect the frame rate, and there are far fewer opportunities for inappropriate decoding due to high speed or adaptive rotation.
[0025]
A single encoded waveform is transmitted and the received waveform is convolved with the impulse response of the decoding filter to compress the waveform energy into small time intervals. The decoding filter may be a matched filter (the filter coefficient here is the same as the transmission code), or may be a mismatched filter. The mismatch filter is designed to minimize the square sum error between the filter output signal (with the code sequence as the input signal) and the Kronecker delta function. Preferred single-transmission codes include Barker codes and non-Barker codes such as length 8 codes [1,1,1, -1,1, -1, -1,1]. . When this code is convolved with the mismatched filter of length 16 shown in FIG. 3, the output signal is as shown in FIG.
[0026]
For each transmission, the echo signal from the transducer element 12 is supplied to the respective receive channel 40 of the receive beamformer. Under the command of the host computer 20 (FIG. 1), the receive beamformer tracks the direction of the transmitted beam. The receive beamformer provides an appropriate receive focus time delay 42 to the received echo signal and sums these echo signals to accurately determine the total ultrasonic energy reflected from a particular location along the transmit beam. A composite echo signal represented by The time-delayed received signal is added in the reception adder 44 for each of the N transmission firings focused at a specific transmission focal position.
[0027]
The received signals added for successive transmission firings are supplied to a filter 46, which performs decoding, bandpass filtering, and wall filtering. This is accomplished by proper selection of filter coefficients by the host computer based on system programming and operator input. The filter 46 performs filtering processing over N transmission firings, and supplies the filtered signal to the detector 32. The detector 32 envelopes the filtered signal from the firing to the firing. Form a line. After post-processing (including edge enhancement and logarithmic compression) and scan conversion, a single scan line is displayed by the display subsystem 22 (FIG. 1). This procedure can be used for each transmit focus position (if there is one transmit focus position for each beam angle) or for each vector (if there are multiple transmit focus positions for each beam angle). And so on, so that each scanning line is displayed.
[0028]
In accordance with the preferred embodiment of the present invention, filter 46 includes an FIR filter 48 having an input coupled to the output of receive summer 44, an input coupled to FIR filter 48, and an output coupled to detector 32. And a vector adder 50. The FIR filter has M filter taps for receiving a respective filter coefficient set of M filter coefficients for each transmit firing. The filter coefficient for the nth transmission firing is a_nc₁, A_nc₂,. . . , A_nc_MWhere a_nIs the scalar weight for the nth transmission firing, where n = 1, 2,. . . , N and c₁, C₂,. . . , C_MIs a set of filter coefficients selected such that the FIR filter 48 compresses the received pulses and passes most of the desired fundamental frequency band. Specifically, the filter coefficient c₁, C₂,. . . , C_MIs a first set of filter coefficients b that is a function of the frequency band to be passed₁, B₂,. . . , B_PIs a second set of filter coefficients d, which is a coefficient of either a matched filter or a mismatched filter₁, D₂,. . . , D_QAnd is obtained by convolution. Here, M = (P + Q−1). Scalar weight a₁, A₂,. . . , A_NForms a slow time “wall” filter that selectively passes signals from reflectors moving at a speed greater than a predetermined threshold, ie, the slow time filter coefficient is It has been chosen to block low frequencies corresponding to slower motion. The successive FIR filter output signals for N transmit firings are coherently accumulated in the vector adder 50. This is equivalent to a wall filter with a single output signal sample. The output signal of the vector adder then undergoes envelope detection, post-processing, scan conversion and display in a known manner.
[0029]
Filter coefficient a_nc₁, A_nc₂,. . . , A_nc_MAre supplied from the filter coefficient memory 52 to the filter 48 by the host computer for each transmit firing. For example, for the first transmission firing, a set of filter coefficients a₁c₁, A₁c₂,. . . , A₁c_MIs supplied to the FIR filter 48, and for the second transmission firing, a set of filter coefficients a₂c₁, A₂c₂,. . . , A₂c_MIs fed to the FIR filter and so on. The filter coefficients are programmable according to the diagnostic application. Different sets of filter coefficients can be stored in a look-up table in the host computer's memory, allowing the system operator to select the desired set of coefficients. For applications where the number of transmit firings N = 2, pairs of filter coefficient pairs are stored in memory, and one set of filter coefficients of the selected pair is stored in the first transmit firing. Previously transferred to the FIR filter, the other set of filter coefficients of this selected pair is transferred to the FIR filter after the first transmission firing and before the second transmission firing. Similarly, for applications where the number of transmit firings N = 3, two or three sets are used to filter the received signal resulting from the first to third firings. Are stored in the memory. The same procedure is followed for applications where the number of transmission firings N> 3.
[0030]
The time interval between each of N transmissions per focus position is controllable by the user and determines the “slow time” filter cutoff frequency. A longer interval between each of the N transmissions to a particular focal position results in a lower cutoff frequency with higher sensitivity to slow flow.
[0031]
The flow image can be displayed by itself for maximum flow contrast or can be summed with a B-mode background image. When the flow image is superimposed on the conventional B-mode image, the diagnostician can observe the blood flow with respect to a known anatomical landmark at the time of medical diagnosis. The flash artifacts seen in this addition of the flow image and the B-mode image are less than that seen in conventional color flow imaging.
[0032]
In a preferred embodiment, a B-mode image derived from the fundamental frequency band is added to the flow image. This is accomplished by perturbing one of the “slow time” filter weights so that the fundamental B-mode signal component passes or feeds through the wall filter. For example, the weight a for the first transmission firing (or for any other transmission firing)₁In addition, a perturbation can be applied by an amount α as shown in FIG. B-mode feedthrough allows a flow image to be displayed superimposed on a conventional B-mode image. Alternatively, the flow image can be displayed superimposed on the conventional B-mode image in color. This feedthrough method is one method for obtaining a B-mode background image without performing additional firing. Other methods for obtaining B-mode background images include those that send pulses specifically for B-mode images (this method is required for ROI), and wall filter even if all objects are stationary. Some send slightly different pulses in one packet so that the output signal never goes to zero.
[0033]
In an alternative preferred embodiment, the background B-mode image may also be a high (or low) harmonic image, which causes reverberation and out-of-plane acoustic “noise” that may mask flow imaging. Some image artifacts such as This can be achieved using any one of several methods.
[0034]
In a preferred embodiment for forming a harmonic B-mode background image, additional uncoded pulses are denoted f₀(Usually at the edge of the lower frequency band of the transducer)₀(Second harmonic) or frequency f₀Filtering is performed by a band-pass filter (incorporated in the FIR filter 48 shown in FIG. 2) centering on / 2 (subharmonic). This unencoded transmission is followed by at least two such encoded transmissions as described above. The received signal for all of these transmissions is filtered in slow time by a wall filter. In the case of two encoded transmissions followed by one uncoded transmission, the wall filter scalar weight is [a₀, A₁, A₂] = [1,1, -1]. The order in which encoded and unencoded transmissions are transmitted can be changed so that the unencoded transmissions are located in the middle or end. Such a configuration has the effect of passing the harmonic B-mode background image vector while subtracting one flow image vector from the other flow image vector.
[0035]
In another preferred embodiment for forming a harmonic B-mode background image, an even number of unencoded pulses are transmitted sequentially followed by at least two encoded pulses. Every other uncoded pulse in the transmitted packet can be negative, so that echoes from negative uncoded pulses are added to echoes from positive uncoded pulses. And cancels the fundamental signal and leaves the harmonic signal. Subsequently, the harmonic signal is added to the flow signal either before detection (coherent) or after detection (incoherent). For example, a transmission packet may be unencoded in two such that the first unencoded pulse has a positive polarity and the second unencoded pulse has a negative polarity. It can consist of a pulse and at least two encoded pulses. The received signal is then filtered by the wall filter in slow time for all of these transmissions. If two unencoded transmissions with opposite polarities are followed by 2 unencoded transmissions, the wall filter's scalar weight is [a₀, A₁, A₂, A_Three] = [1,1,1, −1]. The order in which encoded transmissions and non-encoded transmissions are transmitted can be permutated in a permutation manner. This configuration has the effect of passing the harmonic B-mode background image vector through the wall filter while subtracting one flow image vector from the other flow image vector.
[0036]
In an alternative preferred embodiment of the present invention, the SNR is recovered using a double transmission coded excitation, eg, a Golay code pair. Specifically, the SNR is increased by transmitting a pair of basic sequences encoded in a Golay manner sequentially to the same focal position on each beam, and then decoding the beam-added data. A pair of basic sequences encoded by the Golay equation is formed by convolving the basic sequence with Golay code pairs after oversampling. A Golay code pair is a pair of binary (+1, −1) sequences having the property that the sum of the autocorrelations of the two sequences is a Kronecker delta function. An oversampled Golay sequence is a Golay sequence with zeros between each +1 and −1, and the number of zeros is greater than or equal to the length of the basic sequence minus one. equal. Golay codes do not have range sidelobes. For each firing, decoding is performed using an oversampled Golay sequence corresponding to the Golay-type encoded base sequence used at the time of transmission. By transmitting a sequence of two pulses encoded in polarity according to a Golay pair, the correlation of each of the received beam summed signals with the corresponding oversampled Golay sequence, and the addition of these correlations, results in an SNR of Increases are possible and there is virtually no degradation in image resolution or contrast. Practically, range sidelobes occur due to code distortion, but tend to be lower than the noise floor and do not adversely affect image quality. Tissue movement that occurs during the transmission of two sequences of Golay pairs also introduces code distortion and increases range sidelobes. By transmitting the second sequence as soon as the echo from the first sequence is completely received, the time interval between the two transmissions can be minimized. Minimizing the interval between each transmission minimizes motion-induced code distortion.
[0037]
The ultrasound imaging system shown in FIG. 2 employs complementary code processing. Thus, instead of transmitting a packet of N transmit pulses (eg, each pulse being a tone burst), a packet of 2N transmit pulses is transmitted, and successive transmit pulses are complementary (eg, Are encoded alternately by the respective codes of the (Golley) code pair so as to form the alternately encoded pulse sequences A and B. For example, encoding a sequence of M transmission pulses with a first M-digit transmission code and encoding another sequence of M transmission pulses with a second M-digit transmission code. Thus, an encoded pulse sequence A can be formed, where the first and second transmission codes are complementary. For each transmission firing, autocorrelation is performed by loading a received code equal to the transmitted code for this firing into the decoding filter and filtering the received signal. Then, using a wall filter, for example, add each sequence after autocorrelation so that the scalar weight of each odd-numbered pair transmission is negative with respect to the scalar weight of each even-numbered pair transmission. . According to the preferred embodiment shown in FIG. 6, a transmitter (not shown) provided in beamformer 54 processes packets consisting of complementary encoded pulse sequences A and B in alternating order. That is, firing is performed in the order of ABAB. The received signal obtained from this packet can be expressed as follows.
[0038]
A₁B₁A₂B₂A_ThreeB_Three... A_NB_N
Where A_iIs a received signal for the i-th firing of code A, and B_iIs a received signal for the i-th firing of code B, which is a complementary code of code A, and 2N is the number of firings in the packet. These received signals are compressed by the filter 48 and bandpass filtered. In the case of an RF (radio frequency) beamformer output signal, the demodulator 56 converts the compressed RF signal into its I and Q components, and these I and Q components are converted into a corner turner. Load into memory 58. A wall filter 60 of (1, 1, −1, −1) is applied to each down range position across the firing, ie “slow time”, and each range point. To generate the respective difference signals as follows.
[0039]
(A₁+ B₁)-(A₂+ B₂)
(B₁+ A₂)-(B₂+ A_Three)
(A₂+ B₂)-(A_Three+ B_Three)
(B₂+ A_Three)-(B_Three+ A_Four)
...
(A_N-1+ B_N-1)-(A_N+ B_N)
(The parentheses are inserted to make the mathematical expression clearer.) Thus, adjacent complementary code sequences are added together and the resulting sum is wall filtered. The wall filter output signal is envelope detected by detector 32, scan converted by scan converter 34, color mapped by video processor 62, and displayed on display monitor 64. The wall filter 69 typically allows (N−W + 1) signal samples to be generated, where N is the packet size and W is the length of the wall filter. . The wall filter in the implementation shown in FIG. 2 corresponds to N = W. In addition to using wideband pulses and small packet sizes, the compression curve and display map are selected to provide a greater number of effective display bits. Increase vector density to increase resolution. The flow image is displayed as a color overlay (by a selectable color map loaded into the video processor 62) on the B-mode image.
[0040]
In the embodiment of FIG. 6, even if two encoded pulse sequences are required for each transmission firing, the encoded pulse sequences are staggered to create a fire in the packet. The total number of rings need only be increased by the number of additional wall filter taps since the number of points fed to the flow detector is the same, not by a factor of two. In this way, the frame rate can be kept sufficiently high by using a relatively short wall filter. This implementation also maintains the original velocity dynamic range of ± PRF / 2. Here, the PRF is the frequency at which the individual encoded pulse sequences A or B are fired, not the frequency at which the pair of complementary encoded pulse sequences is fired.
[0041]
When using a single transmission code, the present invention requires at least two transmissions of the code, and the resulting echo is filtered by the wall filter in “slow time”. In contrast, using Golay codes, the acquisition of two data points at different times requires four transmissions, twice for each Golay code of the Golay code pair. Thus, using a two-transmission code requires at least four encoded transmissions. During wall filtering, the same set of scalar weights is applied to the echoes of both Golay codes of the Golay code pair. Additional unencoded transmissions can be included in the packet to obtain a harmonic B-mode background image as before.
[0042]
In the embodiment according to FIG. 2, the decoding filter, the bandpass filter and the wall filter are combined as one FIR filter and vector adder which is dynamically supplied with the appropriate filter coefficients. Alternatively, a separate filter can be used. In the embodiment shown in FIG. 6, the decoding filter and the band pass filter are combined as one FIR filter, but the wall filter is separated. It should also be understood that both the embodiment shown in FIG. 2 and the embodiment shown in FIG. 6 can be used with either single transmission or multiple transmission coded excitation.
[0043]
For both the implementations of FIGS. 2 and 6, the frame rate can be increased by using parallel receive hardware, and more than one from a single transmit vector by using parallel receive hardware. Receive vectors can be processed simultaneously. The packet size, pulse repetition interval (PRI) and region of interest (ROI) can be controlled by the user. The PRI determines the cutoff frequency of the wall filter, and a larger PRI results in a lower cutoff frequency with higher sensitivity to slow flow. When a background B-mode image is formed from a separate firing from that used to form the flow image (ie, without feedthrough), the separate B-mode firing is interleaved with the flow firing. The frame rate can be increased and the flow image can be thresholded separately from the B-mode image.
[0044]
In a preferred embodiment, the scanning direction of the ultrasound beam is opposite to the direction of blood flow, which can increase the apparent flow sensitivity (more uniform fill to the vessel edge). -out)), which can increase resolution (smaller scatterer size). The scanning direction can be changed electronically under user control or can be changed by manually turning the probe.
[0045]
One way to significantly increase flow contrast (ie, sharpness from the background) is to use contrast agents in the fundamental mode. The contrast agent is typically a microbubble encapsulating a gas having a diameter between 0.1 and 10 microns. When introduced into the body through injection, the contrast agent acts as a highly reflective marker of blood flow and perfusion. Since the ultrasonic energy incident on the microbubbles is strongly reflected at the incident (fundamental) frequency and the resonance (harmonic and subharmonic) frequencies, the contrast agent is applied using both the fundamental wave imaging method and the harmonic wave imaging method. Imaging. The preferred method uses the fundamental frequency in the contrast agent echo and not the harmonic frequency. Multiple firings of the same pulse are sequentially transmitted to a specific transmission focal position. More specifically, the fundamental frequency f is transmitted to each transmission focal position.₀N encoded pulses centered around are transmitted. At reception, the FIR filter centered on the fundamental frequency compresses the received pulse and substantially separates the desired fundamental component. Subsequently, the wall filter extracts the basic flow signal over N transmissions. This process is repeated for multiple transmit focal positions in the scan plane and for each of the multiple scan planes. Each scan plane is acquired in succession as the ultrasound probe is swept across the object (object volume) that is scanning (by hand or by a probe positioning device), ie, moving relative to the object. . Data from each scan plane is processed to display an image frame.
[0046]
As shown in FIG. 7, control of the system is centralized on the host computer 20. The host computer 20 receives operator input via an operator interface 90 (eg, keyboard and trackball) and controls the various subsystems. The scan converter 34 includes an acoustic line memory 68 and an XY display memory 70. The B-mode flow imaging data stored in the polar line (R-θ) sector format in the acoustic line memory 68 is converted into appropriately scaled Cartesian coordinate intensity data, which is converted to the XY display memory 70. Is remembered. Each image frame from the XY display memory 70 is sent to the video processor 62. Prior to gray mapping, frames of B-mode flow imaging data in video processor 62 are stored in cine memory 66 in a first-in first-out manner. The storage may be continuous or may occur as a result of an external trigger event. The cine memory 66 is like a circular image buffer operating in the background while constantly acquiring image data that is displayed to the user in real time. When the user freezes the system (via a suitable device on the operator interface 90), the user can see the image data previously acquired in the cine memory.
[0047]
When imaging, a long series of latest images is stored in the cine memory 66 and is continuously updated. Some systems are designed to store R-θ acoustic images and other systems are designed to store XY video images. The image loop stored in the cine memory 66 can be played on the display monitor via the trackball (interface 90), and a portion of the image loop can be selected for hard disk storage.
[0048]
The selected image sequence stored in the cine memory 66 is transferred to the host computer 20 for 3D reconstruction. Multiple frames of flow imaging data acquired during the probe sweep form a three-dimensional data volume. Host computer 20 retrieves the data volume from cine memory 66 and uses any of a variety of volume rendering techniques (eg, maximum or average pixel projection, synthesis or surface) on various imaging planes. Reconstruct the projected image. The projected data resulting from each projection is then written back to another portion of the cine memory or to the memory of the scan converter and from there through the video processor 62 to the display monitor 64.
[0049]
The host computer 20 includes a central processing unit (CPU) 72 and a random access memory 74. The CPU 72 has a memory for storing routines used to convert the acquired flow imaging data volume into a number of three-dimensional projection images taken at different angles. The CPU 72 controls the flow of data among the XY display memory 70, the video processor 62, the cine memory 66, and the CPU itself via the system control bus 76. Each frame of imaging data represents one of a number of scans or slices for the subject and is stored sequentially in the acoustic line memory 68, the XY display memory 70, and the video processor 62. Prior to gray mapping, a frame of B-mode flow imaging data is sent from video processor 62 to cine memory 66. Overlapping frames representing the object volume to be scanned are stored in cine memory 66 to form the source data volume. After the source data volume is acquired, the CPU 72 can configure a three-dimensional projection of the data with any slice that passes through the source data volume.
[0050]
Conventional systems have the ability to superimpose (superimpose) graphic symbols on any ultrasound image. The superimposition of graphics on the image frame is performed in the video processor 62, which receives the ultrasound image frame from the XY display memory 70 and the graphic data from the graphic display memory 78. The graphic data is processed by the graphic processor 80 and input to the graphic display memory 78. The graphics processor 80 is synchronized with other subsystems by the host computer 20.
[0051]
To form a 3D image, the CPU 72 can convert the source data volume retrieved from the cine memory 66 into an imaging plane data set. Such successive transformations are, for example, maximum projections, minimum projections, composite projections, surface projections or average projections which are carried out in angular increments ranging from + 90 ° to -90 °, for example with an angular increment of 10 °. Various projection techniques may be included. Each pixel in the projected image contains transformed data derived by projection onto a given image plane.
[0052]
In freehand 3D ultrasound scanning, the transducer array (1D to 1.5D) is translated in elevation to produce a set of substantially parallel image planes through the anatomy of interest. get. These images are stored in cine memory and can be retrieved later for 3D reconstruction by the system computer. If the spacing between image frames is known, the three-dimensional volume can be reconstructed with the out-of-plane dimension and the scan plane dimension set to an accurate aspect ratio. However, significant estimates of the spacing between slices can cause considerable distortion in the geometry of the three-dimensional object.
[0053]
Conventional ultrasound scanners continuously collect B-mode flow imaging data in cine memory. Three-dimensional data volume is stored in cine memory when the entire anatomical area is swept with the probe, using either freehand scanning or some type of mechanical probe mover . The distance traveled by the probe can be determined by several techniques. A user can provide an estimate of the swept distance. If the probe is moved at a constant speed by the probe mover, the distance can be easily determined. Alternatively, a position sensor can be attached to the probe to determine the position of each slice. Markers applied on the anatomy or in the data could also provide the required positional information. As another method, there is a method in which the displacement of the scanning plane is directly estimated from the speckle decorrelation degree between successive image frames.
[0054]
Referring to FIG. 8, the ultrasound probe 10 sweeps an area of the subject (either manually or by a probe mover) (arrow 82 indicates a linear sweep) and the spacing between slices is known. If there is a slice 84 stored in memory, a three-dimensional data volume 86 can be acquired. By processing this data volume (eg, using projection onto the imaging plane), a three-dimensional view (image) of the area of interest can be formed. In addition, by reformatting the data, individual slices 84 at any angle can be generated (see FIG. 9), which allows the user to view the anatomy being examined. Whatever is, it is possible to obtain the desired accurate view. Algorithms that generate 3D projections of 2D data are well known, and techniques for reformatting data and generating arbitrary slices through a data set are also well known. The problem that arises here is how to display information so that an observer can easily associate a two-dimensional slice with a three-dimensional anatomical structure.
[0055]
10 and 11 show a flowchart of the acquisition and display procedures. The user begins by sweeping the area of interest with an ultrasonic probe (step 100). For example, the sweep can be obtained by a freehand sweep that performs linear motion or rocking. Once the data has been acquired, the user “freezes” the cine memory by pressing the freeze key on the operator interface 90 (FIG. 7) (step 102) and moves the trackball to A range of cine memory frames (slices) to be included in the Z dimension of the data volume is selected (step 104). As shown in FIG. 12, when the trackball is moved, a Z dimension selection gauge 152 appears on the display screen 150. The position of indicator 154 relative to gauge 152 is then controlled using a trackball. The indicator can be moved to the desired leftmost point, and then this leftmost point is locked by pressing a predetermined key on the operator interface. The indicator can then be moved to the desired rightmost point, and then this rightmost point is locked by pressing the same predetermined key. Thereby, slices to be included in the data volume are set. The operator then enters a particular desired “3D mode” by pressing the appropriate key on the interface (step 106).
[0056]
Upon entering 3D mode, the operator must first select the XY dimensions and position of the ROI within the data volume (step 108). This step is performed by manipulating a region of interest box 156 (shown in FIG. 12) that appears at a predetermined position on the display screen 150 in response to a 3D mode key being pressed. The region of interest box 156 can be sized and translated in the X and Y directions to encompass the imaged structure 160 appearing on the sector scan image 158. The region of interest box 156 is translated by moving the trackball and sized by operating a four-way rocker switch built into the operator interface. For example, for a swing switch, moving the switch down increases the size of the region of interest for the Y dimension, moving the switch up decreases the size for the Y dimension, and moving the switch to the right increases the size for the X dimension. However, when the switch is moved to the left, the size is programmed to decrease with respect to the X dimension.
[0057]
After the ROI is defined, the operator selects the desired 3D projection type (minimum, maximum or average pixel projection, surface, synthesis, etc.) and the desired display mode (step 110 in FIGS. 10 and 11), Press the render key on the operator interface. The defined ROI is then retrieved by the host computer 20 from the cine memory 66 (see FIG. 7). The host computer scans the retrieved data for duplicate frames and discards these data (step 112). The host computer then calculates the interval between slices for this data set (step 114). (The spacing between slices is assumed to be constant throughout the length of the data volume.) For example, the spacing between slices is US patent application Ser. No. 09/045 filed Mar. 20, 1998. , 780, and can be calculated using the adaptive speckle correlation method.
[0058]
After the interval between slices is calculated, the system enters a “volume rotation” mode. This mode is one of the three lower modes included in the 3D mode. Referring to FIG. 7, in the “volume rotation” mode, after a signal representing a colored (eg, green) orientation box is generated by the graphics processor 80 and formatted in the XY format in the graphics display memory 78, Sent to the video processor 62. The video processor causes a green orientation box to be displayed on the display screen (step 116). At the same time, the host computer performs a selected maximum resolution pixel projection of the defined data volume based on the calculated spacing between slices (step 118). The projected three-dimensional image is sent to the XY display memory 70 and then sent to the video processor 62. The projected three-dimensional image is also acquired by the cine memory 66. The video processor 62 causes the projected three-dimensional image to be displayed on the display screen along with the orientation box. Both the orientation box and the initial projection are oriented so that the Z-axis is toward the back of the screen, the Y-axis is vertically oriented, and the X-axis is horizontally oriented. That is, the orientation box appears as a rectangle with X and Y dimensions proportional to the X and Y dimensions of the selected region of interest. Data slices are acquired along the Z axis. This is defined as a zero angle projection.
[0059]
In the “volume rotation” mode, the operator can use the trackball to rotate the orientation box around the X and Y axes. Rotation about the Z axis is done using a separate rotation knob on the operator interface. The orientation box follows the movement of the trackball and rotary knob in “real time”, which allows the user to determine the orientation of the box as desired. The rotational position for each axis is shown on the display panel. To achieve real-time rotation of the projected image, the system determines whether the trackball is moving (step 120). If the trackball is moving, the orientation box follows the trackball and decimated rendering is performed in real time (step 122). If the trackball motion has stopped (step 124), the full resolution pixel projection is again created and displayed.
[0060]
FIG. 13 shows an exemplary projection 164 at any angle indicated by the orientation box 162. The orientation box has a marker 166 in the lower left front corner, and the rear corner of the box is broken to help the user identify the front and back of the box. As an aid to visualization, box 168 is overlaid on projection 164, which is aligned with orientation box 162 and is deep enough to appear darker as the box goes “back” of the data volume. Is shaded. The user can reposition and reproject the orientation box 162 as many times as desired using the trackball and rotary knob. In addition, a plus 90 ° or minus 90 ° rotation can be performed using a special key on the operator interface.
[0061]
The rendering of the flow imaging data volume is accomplished using an object-order volume visualization technique in which each voxel is mapped from data space to image space according to a projection matrix. . The goal of this rendering method is to efficiently generate a 2D image from 3D data while minimizing artifacts.
[0062]
Data volume is N along the X, Y and Z axes respectively._x × N_y × N_z Each data sample is known as a voxel. In this data coordinate system, it is assumed that the voxels are equally spaced by one unit along each axis with the center of the data volume as the origin. Since the voxels in the actual data volume may not be equally spaced, a scaling matrix [S that handles the aspect ratio of the data samples_A ] Is used. The data sample can then be rotated around the X, Y and Z axes by the desired angle and projected onto the image plane using orthographic transformation. This process can be written as:
[0063]
[X_d , Y_d , Z_d ] [S_A ] [R] = [x_i , Y_i , Z_i ]
Where [x_d , Y_d , Z_d ] Are the coordinates of the voxels in the data volume, [R] is the desired rotation matrix, and [x_i , Y_i , Z_i ] Represents an image space pixel. Projection transformation [S_A ] [R] is Z^Three R^Three (Here, the coordinates in R space are real numbers, and the coordinates in Z space are integers.) The z component is discarded and R is applied by applying a truncation operation (floor operation, ie rounding a real number to an integer).² Z² Convert to With this projection transformation, the voxel data is processed according to the selected projection method. For example, if maximum pixel projection is used, each image space pixel stores only the maximum value of the voxel data set that has been mapped to each image space pixel by the projection transformation described above. This process is shown in FIG. In the figure, voxel data 170 is initially shown. After multiplication by vertical and horizontal scaling, it becomes voxel data 172, and further rotated by a desired angle, it becomes voxel data 174, which is projected onto the image plane. It is shown that the data becomes voxel data 176.
[0064]
Using the process shown in FIG. 14, adjacent data samples are not necessarily projected onto adjacent pixels on the image plane, so there are blank spots or “holes” in the projected image. Free. In order to remove these holes, introducing two additional scaling steps into the projection yields the following equation:
[0065]
[X_d , Y_d , Z_d ] [S_A ] [R] [S_H ] [S_H ^-1 ] = [X_i , Y_i , Z_i ]
This projection process is then broken down into two steps.
[0066]
[X_d , Y_d , Z_d ] [S_A ] [R] [S_H ] = [X_t , Y_t , Z_t ]
[X_t ′, Y_t '] [S_H ^-1 ] = [X_i , Y_i ]
In the first step, Z^Three Data coordinates [x_d , Y_d , Z_d ] Is R^Three Intermediate coordinates [x_t , Y_t , Z_t ] Is mapped. The z component is discarded and Z is² [X_t ′, Y_t '] Is obtained. In the second step, an inverse scaling operation is performed in the two-dimensional image space.
[0067]
To remove holes from the image, the scaling matrix [S_H ] Guarantees that adjacent data samples (up to one sample apart in the data space along the X, Y and Z axes) are mapped to adjacent pixels. There must be. There are two factors that contribute to image holes. The first factor is due to the scaling of the sample points that handle the aspect ratio, and the second factor is due to the rotation of the data samples. Scaling matrix [S_H ] Removes holes by scaling along the X and Y axes of the image.
[0068]
The scaling factor that handles the aspect ratio is calculated using the following equation:
[0069]
X_as= 1.0 / ‖ [1, 0, 0] [R^-1] [S_A ^-1 ] ‖
Y_as= 1.0 / ‖ [0,1,0] [R^-1] [S_A ^-1 ] ‖
Z_as= 1.0
Using these scaling factors, the intermediate scaling matrix [S_H ′] And this intermediate scaling matrix [S_H ′] Is used to calculate the rotational scaling factor.
[0070]
The scaling factor for correcting the rotation is the four diagonal vectors V = {[1,1,1], [-1,1,1], [-1,1, -1,], [1, 1, −1]} is projected onto the image plane, and the maximum separation along the X and Y dimensions is independently calculated.
[0071]
[Expression 1]

[0072]
[S_H ] Is the product of the vertical and horizontal scaling factor and the rotational scaling factor.
[0073]
X_s = X_asX_rs
  Y_s = Y_asY_rs
  Z_s = 1.0
Using a combination of these scaling factors, each voxel that is separated by one unit along the X, Y, and Z axes in the data space is maximized along the X and Y axes on the image plane. To ensure mapping to pixels that are separated by one unit.
[0074]
If the operator of the ultrasound imaging system wants to view a two-dimensional slice through the data volume in the current X, Y, Z rotational orientation, the operator presses the “Display Mode” key (FIGS. 10 and 10). 11, step 128), switching from “volume rotation” mode to “cut plane” (reformat) mode. Referring to FIG. 15, the color of the orientation box 162 changes (eg, from green to white) and a colored (eg, green) polygon 178 appears in the center of the data volume in the orientation box (step 130). ). The green color indicates that this part of the box can be moved. The system then creates an initial representation 180 in the current orientation of the two-dimensional slice that passes through the center of the data set (step 132).
[0075]
The shape of this slice 180 matches the shape of the green polygon 178 in the orientation box 162. The operator can then use the trackball to scroll through the data set. If the system determines that the trackball is moving (step 134), successive two-dimensional slices of this data set are displayed in real time in the selected placement direction and position. If the “display mode” key is pressed (step 136) when the trackball is not moving (step 134), the system exits the “cut plane” mode and enters the “cut plane rotation” mode.
[0076]
In the “cut plane rotation” mode, the green polygon in the orientation box is moved (step 138), visually indicating the position of the slice in the data volume, and the display panel is updated to show the slice's position. The position is shown as a percentage of the total volume (ie 50% corresponds to the middle of the data volume). Notice that the green polygon indicating the position of the slice through the data volume turns white, the orientation box turns green, the slice is fixed at this point, and the data volume can be rotated relative to the slice. Show. The system then displays the reformatted slice in the manner previously described (step 140). At this point, the user can move the trackball to define the orientation box in a new X, Y, Z orientation (step 142). In doing so, as the user moves the trackball, the orientation box follows the movement of the trackball and the shape of the white polygon changes to show the shape of the slice. When the orientation is set, the user presses the “Calculate” key (step 144) so that the system displays a new two-dimensional slice captured at the angle indicated by the polygon angle relative to the orientation box. To do. If the “display mode” key has not been pressed (step 146), the operator can reset the data volume placement direction. If the “display mode” key is pressed, the system returns to the “volume rotation” mode.
[0077]
Algorithms for generating 3D projections of 2D data are well known, and techniques for reformatting data and generating arbitrary slices through a data set are also well known. As shown in FIG. 7, the data after projection or reformatting is output to the XY display memory 70 by the host computer. The image frame of the data after projection or reformatting is then sent to the video processor 62 and obtained by the cine memory 66. The video processor superimposes the orientation box and other graphic symbols on the image frame of the reformatted data after projection and outputs it to the display monitor.
[0078]
By adjusting the B-mode feedthrough parameters, the user may be able to see the amount of change in the normal B-mode background image for flows imaged in 2D and 3D projections. . The contrast agent can be injected into the subject to enhance the flow information in the display image.
[0079]
While the preferred features of the invention have been illustrated and described, various modifications and changes will occur to those skilled in the art. For example, the present invention is not limited to using a two-phase code, and a multiphase code can also be used. 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 this invention.
[Brief description of the drawings]
FIG. 1 is a block diagram generally illustrating various subsystems of a conventional ultrasound imaging system.
FIG. 2 is a block diagram illustrating an ultrasound imaging system according to a preferred embodiment of the present invention.
FIG. 3 is suitable for use with a non-Barker-type single transmission coded excitation [1,1,1, −1, −1,1, −1,1] according to an example of the preferred embodiment of the present invention; 6 is a graph showing filter coefficients of a 16-tap mismatch filter.
4 is generated by the filter when the filter coefficients of the 16-tap mismatch filter shown in FIG. 5 are convolved with the code [1, 1, 1, −1, −1, 1, −1, 1]. It is a graph which shows the signal after decoding performed.
FIG. 5 is a flow diagram illustrating “slow time” wall filtering with B-mode signal feedthrough according to a preferred embodiment of the present invention.
FIG. 6 is a block diagram illustrating an ultrasound imaging system according to another preferred embodiment of the present invention.
FIG. 7 is a block diagram illustrating means for reconstructing a frame containing successive three-dimensional projections of pixel intensity data according to a preferred embodiment of the present invention.
FIG. 8 is a diagram showing a data volume acquired by linearly scanning an ultrasonic probe in a direction perpendicular to a scanning plane of the ultrasonic probe.
FIG. 9 is a diagram showing individual slices at an arbitrary angle obtained by reformatting the data volume shown in FIG.
FIG. 10 is a flowchart showing a procedure for acquiring and displaying flow imaging data.
FIG. 11 is a flowchart showing a procedure for acquiring and displaying flow imaging data.
FIG. 12 is a diagram showing a display of an ultrasonic image of a sector scanning method together with a figure indicating a volume of interest defined by a user.
FIG. 13 is a diagram showing a display of a projected ultrasonic image formed at an arbitrary angle together with a graphic for displaying the arrangement direction of the data volume.
FIG. 14 is a diagram showing a projection transformation procedure advantageously used in a preferred embodiment of the present invention.
FIG. 15 is a diagram showing a display of a two-dimensional slice together with a graphic representing the position and arrangement direction of the slice with respect to the data volume.

Claims (10)

A transducer array having a number of transducer elements, transmitting wave energy centered at the fundamental frequency in response to electrical activation, and converting the returned wave energy into an electrical signal;
A transmitter coupled to the transducer array, wherein the first and second are configured to activate a plurality of the transducer elements and focus on each of a number of transmit focal positions in each of a number of scan planes. A transmitter programmed to transmit wave energy encoded according to a transmission code during a transmission event;
For each of a plurality of scan lines in each of the plurality of scan planes, following the first and second transmission events, first and second received signals are derived from electrical signals generated by the plurality of transducer elements, respectively. A receiver programmed to form each;
Having first and second sets of filter coefficients, the first and second basic signal components of the first and second received signals are compressed and passed through the band, and exceed a predetermined threshold A wall filter programmed to selectively pass a frequency component corresponding to a wave energy reflector moving at a velocity, wherein the compression is applied to each of a plurality of scan lines in each of the plurality of scan planes. A wall filter formed by deriving and forming a flow signal at least partially from the first and second fundamental signal components passed through the band and the selectively passed frequency components;
A memory for storing and forming a respective set of flow signals as a data volume for each of the plurality of scan planes;
A computer programmed to create imaging data representing a three-dimensional projection of the data volume to have a controllable orientation on the imaging plane by applying a projection transformation to the data volume;
A subsystem for displaying an image that is a function of the three-dimensional projection;
Equipped with a,
All coefficients of the second set of filter coefficients other than at least one coefficient are negative values of the first set of filter coefficients, and the at least one coefficient of the second set of filter coefficients is B An imaging system that is perturbed to allow mode-based signal feedthrough .   The system according to claim 1, wherein the second set of filter coefficients is a negative value of the filter coefficients forming the first set. The transmitter is further programmed to activate a plurality of the transducer elements to transmit wave energy that is uncoded but focused for each transmit focal position during a third transmit event. ,
The receiver is further programmed to form a third received signal from electrical signals generated by the plurality of transducer elements following the third transmit event for each of the scan lines;
The wall filter includes the compressed and bandpassed first and second fundamental signal components and the selectively passed frequency component for each of a number of scan lines in each of the number of scan planes. The system of claim 1, further programmed with a third set of filter coefficients that band pass the harmonic signal components of the third received signal for summing. The transmitter, during the third and fourth transmission events, transmits the respective wave energy that is not encoded but focused for each transmission focal position into the encoded during the third transmission event. Activating a plurality of the transducer elements such that unfocused wave energy is transmitted with an opposite polarity to the uncoded focused wave energy during the fourth transmission event Is programmed to do more,
The receiver is further configured to form third and fourth received signals, respectively, from electrical signals generated by the plurality of transducer elements following the third and fourth transmit events for each of the scan lines. Programmed,
The wall filter substantially cancels the respective fundamental signal components of the third and fourth received signals by the filter coefficients forming the third and fourth sets, and performs a large number in each of the multiple scan planes. For each of the scanning lines, the respective harmonic signal components of the third and fourth received signals are band-passed, and the compressed and band-passed first and second basic signal components are passed through it. The system of claim 1, further programmed to add the selectively passed frequency components.   The system of claim 1, wherein the transducer element includes a piezoelectric element that transmits an ultrasonic wave in response to an electrical activation and converts the returned ultrasonic wave into an electrical signal. A transducer array having a number of transducer elements, transmitting wave energy centered at the fundamental frequency in response to electrical activation, and converting the returned wave energy into an electrical signal;
A transmitter coupled to the transducer array, focused for each of a number of transmit focal positions in each of a number of scan planes, and a first Golay of a Golay code pair during a first and third transmit event Transmitting a focused wave energy encoded with a code and transmitting a focused wave energy encoded with a second Golay code of the Golay code pair during a second and fourth transmission event. A transmitter programmed to activate the transducer element;
For each of a number of scan lines in each of the number of scan planes, first to fourth received signals are derived from electrical signals respectively generated by the plurality of transducer elements following the first to fourth transmission events. A receiver programmed to form each;
The first to fourth basic signal components of the first to fourth received signals are decoded and band-passed, respectively, and a frequency component corresponding to a wave energy reflector that moves at a speed exceeding a predetermined threshold is selectively selected. A wall filter programmed with a first through a fourth set of filter coefficients, wherein each of a plurality of scan lines in each of the plurality of scan planes is subjected to the decoded and band pass first. A wall filter forming a flow signal derived at least in part from the fourth through fourth fundamental signal component and the selectively passed frequency;
A memory for storing and forming a respective set of flow signals as a data volume for each of the plurality of scan planes;
A computer programmed to create imaging data representing a three-dimensional projection of the data volume so as to have a controllable orientation on the imaging plane by applying a projection transformation to the data volume;
A subsystem for displaying an image that is a function of the three-dimensional projection;
Equipped with a,
The third set of filter coefficients is a negative value of the first set of filter coefficients, and the fourth set of filter coefficients is a negative value of the second set of filter coefficients, and is at least one pair of filter coefficients. An imaging system in which the filter coefficients are perturbed to allow B-mode fundamental signal feedthrough . 7. The system according to claim 6 , wherein the third set of filter coefficients is a negative value of the first set of filter coefficients, and the fourth set of filter coefficients is a negative value of the second set of filter coefficients. . The transmitter is further programmed to activate a plurality of the transducer elements to transmit wave energy that is unencoded but focused for each of the transmit focal positions during a fifth transmission event. ,
The receiver is further programmed to form, for each of the scan lines, a fifth received signal from electrical signals generated by the plurality of transducer elements following the fifth transmission event;
The wall filter band-passes the harmonic signal component of the fifth received signal for each of a plurality of scanning lines in each of the plurality of scanning planes, and the first to th 7. The system of claim 6 , further programmed with a fifth set of filter coefficients to add a fourth fundamental signal component and the selectively passed frequency. The transmitter is further programmed to activate a plurality of the transducer elements to transmit unencoded focused wave energy during the fifth and sixth transmission events for each of the transmit focal positions. ,
The uncoded focused wave energy of the fifth transmission event has an opposite polarity to the uncoded focused wave energy of the sixth transmission event, and the reception The device is further programmed to form fifth and sixth received signals, respectively, from electrical signals generated by the plurality of transducer elements following the fifth and sixth transmit events for each of the scan lines. And
The wall filter substantially cancels the respective fundamental signal components of the fifth and sixth received signals, and for each of the multiple scan lines in each of the multiple scan planes, In order to pass the respective harmonic signal components of the received signal in a band, and to add the first to fourth fundamental signal components decoded and band-passed and the selectively passed frequency, 7. The system of claim 6 , further programmed with a fifth and sixth set of filter coefficients. The system according to claim 6 , wherein the transducer element includes a piezoelectric element that transmits an ultrasonic wave in response to an electrical activation and converts the returned ultrasonic wave into an electrical signal.

Images