Medical diagnostic ultrasonic imaging system and method

[0002] This application is a continuation-in-part of U.S. patent application Ser. No. 09/213,996, filed December 17, 1998 and now U.S. Pat. No. 6,110,115, the entirety of which is hereby incorporated by reference.

[0003] The present invention relates to medical diagnostic ultrasonic imaging systems, and in particular to improved signal paths that improve resolution and reduce speckle variance in such imaging systems.

[0004] Speckle is generated by interference from waves emanating from multiple scatterers within the sample volume (resolution cell) of an ultrasound imaging system. For log-compressed images, speckle acts as additive noise that interferes with the detectability of small and/or subtle lesions that only have a small echogenicity difference as compared to adjacent tissues. For this reason, speckle variance reduction is critical to enhancing detectability of tissue echogenicity differences.

[0005] Several speckle reduction techniques are widely used in the medical ultrasound field. Most employ some form of spatial, temporal, or frequency compounding. For example, post-detection video filtering, persistence filtering of spatially uncorrelated images, and similar methods are used. These methods cause a loss of spatial or temporal resolution, or both.

[0006] The present invention is defined by the following claims, and nothing in this section should be taken as a limitation on those claims. By way of introduction, the ultrasonic imaging systems described below include a signal path that reduces speckle while preserving or even improving detail resolution, depending on the amount of speckle reduction desired. The signal path includes a whitening filter such as a two-dimensional whitening filter positioned upstream of the detector, and a low pass filter such as a two-dimensional low pass filter positioned downstream of the detector. The disclosed signal path thus provides pre-detection spatial frequency whitening followed by post-detection spatial smoothing. In some of the embodiments described below, the amplitude of the whitened ultrasonic receive signal is adjusted to suppress side lobes that may be associated with spectrum whitening.

[0007] FIG. 1 is a block diagram of an ultrasonic imaging system that incorporates a preferred embodiment of this invention.

[0008] FIGS. 2-5 are block diagrams of filters suitable for use in the embodiment of FIG. 1.

[0009] FIGS. 6 and 7 are graphs showing whitening filter spatial domain characteristics and whitening filter frequency domain characteristics, respectively.

[0010] FIG. 8 is a graph showing the input and output of a pre-detection whitening filter, wherein the input signal is characteristic of a one-dimensional pin target and speckle.

[0011] FIG. 9 is a graph showing the input to a pre-detection whitening filter and the corresponding output from a post-detection low pass filter.

[0012] FIG. 10 is a block diagram of a second preferred embodiment of the ultrasonic imaging system of this invention.

[0013] FIG. 11 is a block diagram of a portion of a receive signal path of an ultrasonic imaging system that incorporates a third preferred embodiment of this invention.

[0014] FIG. 12 is a more detailed block diagram of one embodiment of the suppression factor estimator of FIG. 11.

[0015] FIG. 13 is a block diagram of a variation of the receive signal path of FIG. 11.

[0016] FIG. 14 is a block diagram of one embodiment of the low pass filter of FIG. 11.

[0017] Turning now to the drawings, FIG. 1 is a block diagram of an ultrasonic imaging system 10 that incorporates a preferred embodiment of this invention. The system 10 includes a transmit beamformer 12 that applies transmit waveforms via a multiplexer 14 to a phased array transducer 16. The transducer 16 generates acoustical signals in response to the transmit waveforms, and these acoustical signals are timed and phased to focus ultrasonic energy along one or more desired lines.

[0018] Echoes from the region being imaged are received by the transducer 16 and applied via the multiplexer 14 to a receive signal path 17. The receive signal path 17 includes a receive beamformer 18 that applies beamformed receive signals to a line-to-line phase adjuster 20. The phase adjuster 20 assures adequate coherence between adjacent acoustic lines to allow effective lateral interpolation prior to detection. The output of the phase adjuster 20 is applied to a whitening filter 22, and the whitened receive signal from the filter 22 is amplitude or envelope detected by a detector 24. The detected receive signal is then filtered with a low pass filter 26. The output of the low pass filter 26 is applied to conventional downstream processing such as a conventional image processor.

[0019] As shown in FIG. 1, pre-detection whitening is preferably applied after receive beamformation in this embodiment. Some beamformers generate analytic (complex) output signals that are demodulated down to baseband. The embodiment of FIG. 1 operates with the receive signals applied to the whitening filter 32 at baseband. However, the present invention is not limited to such embodiments. In alternative embodiments, receive signals at frequencies other than baseband (e.g. RF or IF receive signals) can be applied to the whitening filter. In this case, the whitening filter is preferably centered at the center frequency of the receive signal.

[0020] Successful axial filtering is sensitive to any mismatch between the center frequency of the whitening filter 22 (which is zero for analytic baseband receive signals) and the center frequency of the receive signal applied to the whitening filter. Preferably, the receive signal path 17 includes adaptive demodulators (not shown) to compensate for variations in the signal center frequency which result from frequency-dependent attenuation coefficients that can vary between patients and tissue types. Lateral filtering is generally less sensitive to any mismatch between the center frequency of the whitening filter and the center frequency of the receive signal applied to the whitening filter.

[0021] The whitening filter operates to increase the spatial bandwidth of the receive signal (as measured at the −6 dB points), both axially and laterally. In this way, the whitening filter 22 improves the detail resolution at the cost of increases in range lobe and side lobe levels. The whitening filter 22 also increases the number of independent speckles per unit area.

[0022] The post-detection low pass filter 26 is designed to reduce speckle variance and also to reduce the range and side lobe levels enhanced by the whitening filter 22. The level of smoothing provided by the low pass filter 26 can be adjusted to achieve the desired compromise between speckle reduction and enhanced spatial resolution.

[0023] The pre-detection and post-detection filters 22, 26 can take many forms including the following:

[0024] Three-dimensional filters operating along all three axes:

[0025] azimuth (i.e. lateral), range (i.e. axial), and elevation;

[0026] Two-dimensional filters operating on any two of the three axes;

[0027] One-dimensional filters operating on any one of the three axes.

[0028] The filters 22, 26 can be separable or inseparable, or they can use any desired combination of separability. The filters 22, 26 can also be dependent upon the position in any one, any two, or all of three of the axes. For example, the filters 22, 26 can be range dependent, line or azimuthal position dependent, range and line dependent, and so forth.

[0029] The pre-detection whitening filter and the post-detection low-pass filter can be amplitude-dependent, and/or local statistics (e.g. amplitude variance) dependent.

[0030] The pre-detection whitening filter can be an inverse filter such as a Wiener filter and therefore complex, or it can be a simple, real and symmetric whitening filter. Preferably, the whitening filter is an inseparable inverse filter which is line dependent. However, line and range independent, real, symmetric, and separable two-dimensional filters can be an acceptable compromise for some embodiments, trading hardware/software simplicity against optimum performance.

[0031] FIGS. 2-5 provide schematic illustrations of suitable filters for the whitening filter 22 and the low pass filter 26, including two and three-dimensional filters 30 (FIG. 2), one-dimensional lateral filters 32 (FIG. 3), one-dimensional axial filters 34 (FIG. 4), and inverse filters 36 (FIG. 5). For all of these filters, the filtering region can be shaped as desired, and can be circularly symmetric or elongated, depending upon the imaging system.

[0032] The imaging system 10 operates by whitening the ultrasonic receive signal with the filter 22, detecting the whitened receive signal with the detector 24, and then low pass filtering the detected receive signal with filter 26.

[0033] FIGS. 6 and 7 provide further information regarding a preferred whitening filter 22. FIG. 6 provides a spatial domain graph showing an input signal 50 to the filter 22, the filter impulse response 52 of the filter 22, and the resulting output 54 of the filter 22.

[0034] FIG. 7 is a frequency domain graph that illustrates the spatial frequency spectrum 56 of the input signal to the whitening filter 22, the whitening filter frequency response 58, and the spatial frequency spectrum 60 of the output signal from the whitening filter 22. Note that the spatial frequency output spectrum 60 exhibits a whitened spatial spectrum such that the bandwidth (as measured at the −6 dB points) is maximized and the passband has a relatively flat magnitude response. The maximum bandwidth achievable is ultimately limited by the transducer bandwidth, the center frequency, and the effective aperture size in azimuth and elevation. The signal to noise ratio may also be a limiting factor, because whitening causes a loss in the signal to noise ratio. Note however that the post-detection smoothing described above will help recover most of the loss in the signal to noise ratio, depending on how much of the detail resolution improvement provided by the whitening filter is retained.

[0035] In FIGS. 6 and 7, the input signal to the whitening filter 22 is assumed to have an envelope with a Gaussian shape. Note that the spatial frequency spectrum 60 of FIG. 7 has a RECT-like magnitude, and that the output spectrum 60 is much wider at the −6 dB points than the input spectrum 56.

[0036] FIGS. 8 and 9 are graphs that illustrate the effect of pre-detection whitening and post-detection smoothing on detail resolution and speckle variance. For these figures a simulated input signal for the whitening filter 22 was used. This simulated input signal was obtained from a one-dimensional simulated phantom that includes a hypoechoic lesion and a pin target. The solid curves 70, 74 of FIGS. 8 and 9 show this one-dimensional input signal, which may be thought to be taken from any acoustic line in a three-dimensional imaged region, assuming a Gaussian impulse response as shown in FIG. 6.

[0037] The dash-dot curve 72 of FIG. 8 is the output of the pre-detection whitening filter 22, prior to operation of the post-detection smoothing filter 26. The pin target response shows an improvement in detail resolution as well as an associated increase in range/side lobes. The average speckle length of the dash-dot curve 72 is about one-half the average speckle length of the solid curve 70.

[0038] As shown in FIG. 9, the output of the post-detection low pass filter 26 for the same simulation (the dot-dash curve 76) shows that the detail resolution has returned to the original detail resolution of the unfiltered image (down to the −15 dB points). However, speckle variance has been reduced significantly as compared to the original unfiltered image, making the simulated lesion more easily detected.

[0039] It is not essential in all embodiments that the whitening filter be provided in the receive signal path. An alternate embodiment is shown in FIG. 10 in which the whitening filter 40 is included in the transmit/receive signal path 38, downstream of the transmit beamformer 12 and upstream of the multiplexer 14. In these embodiments it is the transmit ultrasonic signal that is designed to whiten the round-trip impulse response. The embodiment of FIG. 10 operates similarly to the embodiment of FIG. 1 discussed above, in that the receive signal applied to the detector 24 has a substantially RECT spectral shape. Other transmit pulse shaping techniques can be used to provide the desired whitening, and whitening techniques can be used both upstream and downstream of the transducer 16 in some embodiments.

[0040] As pointed out above, the pre-detection whitening filter and the post-detection low-pass filter can be amplitude dependent, and/or local statistics (e.g., amplitude variance) dependent. Pre-detection spectrum whitening improves detail resolution in an ultrasound image, but it may provide undesired side effects, as suggested above. One of these undesired side effects is an increase of side lobe levels. FIG. 11 relates to an embodiment that reduces this undesired side effect. Elements 80 through 86 of FIG. 11 can be used in substitution for the whitening filter 22 and the detector 24 of FIG. 1. The system of FIG. 11 is shown as responding to a receive signal A(x,y) from the line-to-line phase adjuster 20, and as supplying a whitened, noise-suppressed receive signal I_ws(x,y) to the low-pass filter 26. The remaining elements of the ultrasonic imaging system may be as described above in conjunction with FIG. 1.

[0041] As shown in FIG. 11, the receive signal A(x,y) is applied to a whitening filter 80, which may be any of the whitening filters described above. The output A_w(x,y) of the whitening filter 80 exhibits a frequency spectrum that is whitened, as described above. The two receive signals A(x,y) and A_w(x,y) are applied in parallel to an intensity detector 82 that generates respective detected receive signals I(x,y) and I_w(x,y). The intensity detector may for example generate the output signals as a function of the product of the respective input signal and its complex conjugate. The receive signal I_w(x,y) is applied to a multiplier 86, and the receive signal I(x,y) is applied to a suppression factor estimator 84. The suppression factor estimator 84 generates a suppression factor S(x,y) as a function of the local statistics of I(x,y). The suppression factor S(x,y) is also applied to the multiplier 86. The resulting product signal I_ws(x,y) generated by the multiplier 86 is then applied to the low-pass filter 26 of FIG. 1.

[0042] FIG. 12 provides further details of the suppression factor estimator 84, which includes a local variance estimator 90 that generates a variance signal V(x,y) as a function of the variance of the receive signal I(x,y). As used herein, the term “variance” is intended broadly to encompass any suitable measure of variability of the input signal to the estimator 90. For example, the local variance estimator 90 can generate the variance signal V(x,y) as a function of local variance, local variance normalized with the square of the local mean, or other statistical parameters that provide an indication of the variability of the input signal. The variance signal V(x,y) is applied as an input to a lookup table 92 that generates an output signal ranging between 1 and 0. Typically, when the variance signal V(x,y) is low, the output signal is set equal to 1. When the variance signal exceeds a certain value, the output signal generated by the lookup table 92 becomes less than 1 and decreases toward 0 or toward a predetermined value between 0 and 1. The output signal of the lookup table 92 can optionally be filtered by a low-pass filter 94 to generate the suppression factor S(x,y) described above.

[0043] The suppression factor estimator 84 described above can be taken as an example of a first circuit operative to generate a first signal indicative of transitional regions that are susceptible to side lobe artifacts, i.e., regions near bright targets. Many variations are possible for the suppression factor estimator. For example, the suppression factor estimator can use a parameter related to gradient instead of variance. One specific alternative is the absolute value of normalized gradient of the receive signal I(x,y): $\frac{⟨{\left|\frac{\partial l}{\partial x}\right|}^2+{\left|\frac{\partial l}{\partial y}\right|}^2⟩}{{⟨l⟩}^2},$

[0044] where the symbol<>denotes mean.

[0045] The multiplier 86 can be taken as an example of a second circuit that is operative to vary gain of the signal path as a function of the suppression factor. Of course, this second circuit is also not limited to the multiplier shown and also can take many forms. For example, the second circuit can be constructed using analog or digital techniques to perform multiplication, addition, or lookup functions. Thus, the second circuit can include multipliers, adders and lookup tables as well as other functional elements. The first and second circuits can be implemented with a single digital processor, executing appropriate routines at respective times.

[0046] Alternative embodiments employ multiple whitening filters. The example of FIG. 13 is similar to that of FIG. 11 described above, except that multiple whitening filters 80′ provide respective receive signals A_w1(x,y) . . . A_wm(x,y) to the intensity detector, and the intensity detector provides multiple detected receive signals I_w1(x,y) . . . I_wm(x,y) to the multiplier 86 ′.

[0047] Alternative embodiments of the low-pass filter of FIG. 1 use a single low-pass filter or multiple low-pass filters. The example of FIG. 14 shows a low-pass filter 26′ that includes multiple low-pass filters 27′ that differ in filter characteristics and provide respective filtered receive signals I_LPF1(x,y) . . . I_LPFl(X,Y) to a weighted summer 29′ that is also responsive to the suppression factor S(x,y) of FIG. 11.

[0048] It will be understood that the present invention can be adapted for use with the widest variety of ultrasonic imaging systems, including the widest variety of beamformers, detectors and filters. For example, the transmit and receive beamformers 12, 18 can be analog, digital, or hybrid beamformers. In some cases, the receive beamformer 18 may provide receive signals that are phased suitably such that the line to line phase adjuster 20 may be eliminated. The beamformers disclosed in U.S. Pat. Nos. 5,685,308 and 5,675,554 are suitable for use with this invention.

[0049] Whitening filters are well-known to those skilled in the art and can take analog or digital forms. Any suitable form can be used with this invention. Similarly, low pass filters such as the low pass filter 26 are well-known to those skilled in the art. Any suitable form including both analog and digital filters can be used with this invention. For example, a whitening filter may be an inverse, pseudo-inverse, or edge enhancement filter that fully or partially compensates for the transmitter, transducer, and receiver filter responses as well as for propagation effects such that the total round-trip impulse response of the imaging system has a substantially flat pass-band with a bandwidth at the −6 dB points that is as wide as possible.

[0050] As used herein the term “receive signal” is intended broadly to encompass any signal responsive to ultrasonic echoes along the receive signal path 17, which may in some embodiments include the image processor. Thus, the term “receive signal” may be applied to signals before or after beam-formation.

[0051] The terms “upstream” and “downstream” are intended broadly to cover the flow of information as it moves from the transmit beamformer 12 to the image processor. Thus, the whitening filters 22, 40 are upstream of the detector 24, and the low pass filter 26 is downstream of the detector 24.

[0052] The term “function of” is intended broadly. For example, a parameter that is said to be a function of another parameter may optionally also be a function of additional parameters. Thus, a first parameter is said to be a function of the second parameter, whether or not the first parameter is additionally a function of third, fourth or other parameters. The function may be linear, nonlinear, stepwise linear or any other desired shape.

[0053] The pre-detection whitening filter described here can also be applied to the input images of a multi-image technique, where multiple images are combined pre- or post-detection, e.g. compounding. For example images to be compounded can be whitened before detection and compounding, followed by low-pass filtering.

[0054] Finally, it will be understood that for ease of understanding the block diagram of FIG. 1 has been simplified. In many embodiments there will be intermediate processing steps at various stages along the signal path. Such intermediate processing may in some embodiments be found between the whitening filter 22 and the detector 24, as well as between the detector 24 and the low pass filter 26.

[0055] The foregoing detailed description has discussed only a few of the any forms of this invention can take. For this reason, it is intended that this detailed description be regarded as illustrative rather than limiting. It is only the following claims, including all equivalents, that are intended to define the scope of this invention.