Surgery – Diagnostic testing – Detecting nuclear – electromagnetic – or ultrasonic radiation
Reexamination Certificate
2001-12-26
2004-03-02
Imam, Ali M. (Department: 3737)
Surgery
Diagnostic testing
Detecting nuclear, electromagnetic, or ultrasonic radiation
Reexamination Certificate
active
06699189
ABSTRACT:
FIELD OF THE INVENTION
The present invention is directed to ultrasound imaging and more particularly to a method and apparatus for performing such imaging in which aberrations are estimated and corrected.
DESCRIPTION OF RELATED ART
The use of ultrasound to produce images for medical diagnosis has become common as a result of its nonionizing nature, the ability to produce images resulting from the inherent differences in mechanical properties of various soft tissues, and advances in technology. Current applications include examination of the heart, abdomen, and fetus. In most areas, diagnosis is now generally based on the size, position, contour, and motion of structures as well as on their relative transmission and reflection properties. Although clinical activity continues to expand, present instrumentation using simple geometric focusing is limited by tissue inhomogeneities that produce wavefront distortion. This distortion can degrade geometric focusing to yield images unsatisfactory for accurate diagnosis.
Transmission measurements of wavefront distortion produced by the abdominal wall, breast, and chest wall have been made, as have pulse-echo measurements of ultrasonic distortion produced by the breast and liver. These measurements all indicate that both point resolution, which measures the ability to distinguish between isolated scatterers, and contrast resolution, which measures the distinguishability of speckle patterns and can be more important clinically than point resolution, are severely reduced in many clinical ultrasonic imaging situations.
Wavefront distortion becomes a more serious problem as ultrasonic imaging systems are designed for operation at higher frequencies. This is because excursions in arrival time are larger relative to the wavelength at higher frequencies. These time shifts can significantly reduce the resolution improvement ordinarily associated with focusing at higher ultrasonic frequencies. Therefore, the need for estimation and correction of aberration becomes greater with increasing frequency.
Wavefront distortion can result from arrival time shifts, amplitude variations, and waveform shape changes, among which arrival time shifts may be the most important. Several efforts have been reported to compensate the arrival time shifts. It is known to use cross correlation of signals to estimate arrival time fluctuations and to maximize brightness by adjustment of focus delays. A least-mean-square-error method has been used to estimate the arrival time. In all the above studies, the underlying aberration model is a phase screen in the receiving aperture and only time shifts are taken into account. However, an improved compensation method has been described in which a phase screen is placed at some distance away from the aperture and the estimation of arrival time is made after the wavefronts are backpropagated to the position of the phase screen. This approach is capable of compensating more than time shifts.
Methods based on a single phase screen model are not able to remove aberration completely. First, the aberration is assumed to be introduced by a thin layer, while the thickness of the medium that produces the aberration is generally not negligible in practice. Second, attenuation in tissue is not considered, even though muscle and fat have a frequency dependent attenuation that is, for example, 4.0 dB/cm and 1.7 dB/cm, respectively, at a frequency of 2.6 MHz. Third, propagation through the tissue may produce strong waveform distortion or even multiple pulses that can cause time-shift estimation to fail or cause time shifts to be an inappropriate description of aberration. Simulations of ultrasonic pulse propagation through the abdominal wall and chest wall have shown that weak fluctuation models do not fully describe wavefront distortion.
SUMMARY OF THE INVENTION
In light of the foregoing, it is apparent that a need exists in the art to improve the removal of aberrations. It is therefore an object of the invention to provide an ultrasound imaging technique which does so. It is another object of the invention to improve on the above-noted techniques of the prior art, particularly those based on a single phase screen model.
To achieve the above and other objects, the present invention is directed to a technique for estimation and correction of aberration based on a common random input filter model. In the model, instead of using a phase screen to describe the aberration, the propagation paths are more generally described by a bank of linear filters. The model is capable of including not only time shifts but also amplitude and waveform variations that may also be important components of the distortion. The aberration is described with an accuracy that depends on the number of parameters in each filter rather than any artificial assumption about the nature of the propagation path. After the filter parameters are known, the aberration can be compensated by corresponding inverse filters in beam formation on reception and appropriate time reversal of waveforms for focusing on transmit.
However, the estimation of the linear filter parameters is a challenge because, although the output of each path is known, the signal from the common source is unknown. This challenge has attracted the attention of investigators in other fields of endeavor, particularly in the area of communications, where the basic problem often arises and is known as blind system identification (BSI).
In BSI, in the general case where the channels have nonminimum phase, higher-order statistics (HOS) were first used to identify the system because HOS are rich in information. However, inherent difficulties limit the application of HOS. A relatively large number of data samples is required, and the convergence rate is slow. Also, HOS methods may yield a local minimum. Methods based on second-order statistics (SOS) are, thus, preferable. Second-order statistics were found to contain phase information that can be use to recover channel information in the case of cyclostationary signals. However, in this approach, statistical assumptions about the input signal must be made and this is inappropriate particularly when signals are short. Breakthroughs in BSI have recently been made. The cross-relation method provides an algorithm based on SOS for the identification of channels driven by an arbitrary unknown input. That method uses an equality between the output of one channel convolved with the response of another channel and the output of the latter channel convolved with the response of the former channel, a relation for which that approach is named, and necessary and sufficient identifiability conditions are known in terms of the channel characteristics and the input signal. A second step is known that refines the channel estimation by using the channel outputs and initial estimates of filter parameters. This approach is called the two-step maximum likelihood method. An alternative approach is known, called the subspace method. The subspace method is based on the orthogonality of “signal” and “noise” subspaces.
A common random input filter model is used for estimation and correction of wavefront aberration in ultrasonic b-scan imaging. In the model, aberration between the focus and the transducer elements is represented by the response of a linear filter bank to a common random signal. The response of each filter in the bank is found using a two-level extension of an existing subspace method for blind system identification. The received waveforms are compensated using an inverse filter, and the transmit waveforms are predistorted using time reversal.
To test the model, experiments were conducted using a two-dimensional array system to obtain echoes from a point reflector and from a random medium in each case through an aberrator. The aberrator is a phantom that mimics the wavefront distortion produced by a human abdominal wall and the random medium is made to mimic ultrasonic characteristics of a human liver. The results indicate that the method can improve both the transmit and the recei
Lin Feng
Waag Robert C.
Blank Rome LLP
Imam Ali M.
University of Rochester
LandOfFree
Ultrasound distortion compensation using blind system... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Ultrasound distortion compensation using blind system..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Ultrasound distortion compensation using blind system... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3247256