Surgery – Diagnostic testing – Detecting nuclear – electromagnetic – or ultrasonic radiation
Reexamination Certificate
1999-07-09
2001-10-02
Jaworski, Francis J. (Department: 3737)
Surgery
Diagnostic testing
Detecting nuclear, electromagnetic, or ultrasonic radiation
Reexamination Certificate
active
06296612
ABSTRACT:
FIELD OF THE INVENTION
This invention relates to ultrasonic diagnostic systems which measure the velocity of blood flow using spectral Doppler techniques. In particular, the invention relates to the continuous display of such information, including maximum and mean blood flow velocities.
BACKGROUND OF THE INVENTION
Ultrasonic scanners for detecting blood flow based on the Doppler effect are well known. Such systems operate by actuating an ultrasonic transducer array to transmit ultrasonic waves into the object and receiving ultrasonic echoes backscattered from the object. For blood flow measurements, returning ultrasonic waves are compared to a frequency reference to determine the frequency shifts imparted to the returning waves by moving objects including the vessel walls and the red blood cells inside the vessel. These frequency shifts translate into velocities of motion.
In state-of-the-art ultrasonic scanners, the pulsed or continuous wave Doppler waveform is com- puted and displayed in real-time as a gray-scale spectrogram of velocity versus time with the gray-scale intensity (or color) modulated by the spectral power. The data for each spectral line comprises a multiplicity of frequency data bins for different frequency intervals, the spectral power data in each bin for a respective spectral line being displayed in a respective pixel of a respective column of pixels on the display monitor. Each spectral line represents an instantaneous measurement of blood flow.
In the conventional spectral Doppler mode, an ultrasound transducer array is activated to transmit by a transmit ultrasound burst which is fired repeatedly at a pulse repetition frequency (PRF). The PRF is typically in the kilohertz range. The return radiofrequency (RF) signals are detected by the transducer elements and then formed into a receive beam by a beamformer. For a digital system, the summed RF signal from each firing is demodulated by a demodulator into its in-phase and quadrature (I/Q) components. The I/Q components are integrated (summed) over a specific time interval and then sampled. The summing interval and transmit burst length together define the length of the sample volume as specified by the user. This so-called “sum and dump” operation effectively yields the Doppler signal backscattered from the sample volume. The Doppler signal is passed through a wall filter, which is a high pass filter that rejects any clutter in the signal corresponding to stationary or very slow-moving tissue, including a portion of the vessel wall(s) that might be lying within the sample volume. The filtered output is then fed into a spectrum analyzer, which typically takes the complex Fast Fourier Transform (FFT) over a moving time window of 64 to 256 samples. The data samples within an FFT analysis time window will be referred to hereinafter as an FFT packet. Each FFT power spectrum is compressed and then displayed via a gray map on the monitor as a single spectral line at a particular time point in the Doppler velocity (frequency) versus time spectrogram.
Typically the I and Q components of the Doppler signal are filtered separately by identical wall filters, which can be implemented as either an FIR or IIR filter. For s harp rejection of low-frequency clutter, a narrow transition band in the frequency response of the filter is required. Typically, the wall filter cutoff frequency is manually selected via a front-panel control key. Usually the wall filter cutoff frequency is increased when bright, low-frequency clutter is seen in the spectral image. Each time the wall filter cutoff setting is changed, a corresponding set of filter coefficient values are read out of a lookup table (LUT) and loaded into the wall filters.
The main limitation with the manually selected filter approach in the prior art is that once the cutoff frequency is set, the wall filter does not change even though the clutter frequency and bandwidth may vary with time, due to radial and/or lateral motion of the vessel walls over the cardiac cycle. As a result, the selected filter cutoff is often optimal only for a small portion of the cardiac cycle.
SUMMARY OF THE INVENTION
The present invention is a method and an apparatus for adaptive wall (high-pass) filtering which overcomes the limitation of the prior art manually selected filter approach. In accordance with the preferred embodiment of the invention, the wall filter cutoff frequency is selected automatically, thereby improving ease of use and productivity. A further advantage is that a wall filter cutoff frequency tailored to each new FFT packet can be used.
In accordance with the preferred embodiment of the invention, low-frequency clutter is removed in the Doppler I/Q data prior to FFT processing. The I/Q data is passed through a low-pass filter whose cutoff frequency is set at the highest anticipated clutter frequency (e.g. 40% of PRF) for the current Doppler application. The low-pass filter rejects the flow frequency components above the clutter frequency range. The total power of the low-pass filter output, i.e., the sum of (I
n
2
+Q
n
2
) over the FFT (or a fraction of the FFT) packet size M, is then computed.
In accordance with the preferred embodiments, a system noise model is used to predict the mean system noise power in the low-pass filter output. The mean system noise power predicted by the system noise model provides a noise threshold to gage how much clutter power is present in the current FFT packet. If no significant clutter is present, then wall filter selection logic will automatically select the lowest wall filter cutoff frequency in a filter coefficient LUT. If significant clutter power is present in the FFT packet, the algorithm proceeds to compute the mean and variance of the clutter frequency over the FFT packet. The estimated mean and variance of the clutter frequency are then fed into the filter selection logic, which selects the most suitable filter cutoff for the current clutter signal. To avoid having the wall filter cutoff frequency fluctuate too much from one FFT packet to the next, some persistence function may be applied to the prescribed wall filter cutoff frequency. Once the new optimal filter cutoff is selected, the rest of the processing is the same as in conventional Doppler wall filtering.
It should be clear to those skilled in the art that the method of the invention can be implemented in hardware (e.g., a digital signal processing chip) and/or software.
REFERENCES:
patent: 5269308 (1993-12-01), Hagiwara et al.
patent: 5299174 (1994-03-01), Forestieri et al.
patent: 5383464 (1995-01-01), Shiba
patent: 5443071 (1995-08-01), Banjanin et al.
patent: 5544659 (1996-08-01), Banjanin
patent: 5910118 (1999-06-01), Kanda et al.
patent: 6146331 (2000-11-01), Wong
Kulakowski Richard M.
Mo Larry Y. L.
Cabou Christian G.
Flaherty Dennis M.
General Electric Company
Jaworski Francis J.
Vogel Peter J.
LandOfFree
Method and apparatus for adaptive wall filtering in spectral... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for adaptive wall filtering in spectral..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for adaptive wall filtering in spectral... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2608516