Abstract: A noise-adaptive method for processing an ultrasonic image set forms a filtered image set having a selectively enhanced noise component as compared to the original image set. A noise parameter is generated as a function of the image set and the filtered image set, and then the noise parameter is used in ultrasonic image processing. A background noise image is generated from the noise parameter and the original image and is used in ultrasonic image processing.

Abstract: An ultrasonic imaging system includes a processor that varies the receive signal path parameters as a function of the signal to noise ratio of the echo signal. Background noise information is either acquired by imaging with the transmitters turned off, or estimated by using the known differences in bandwidth and correlation lengths of the signal and noise, or computed based on a system noise model using currently prevailing system parameters. Acquired receive signals are then processed as a nonlinear function of the comparison between the acquired receive signal and the background noise values. The family of SNR adaptive processors includes the SNR adaptive filtering in at least one of the azimuth, elevation and time dimensions, SNR adaptive high pass, band pass, and whitening filtering, SNR adaptive synthesis, and SNR adaptive compounding.

Abstract: An ultrasonic imaging system and method provide whitening using a two dimensional pre-detection filter followed by low pass filtering using a two dimensional post-detection filter to reduce speckle variance and enhance spatial resolution of the resulting image. The amplitude of the whitened signal can be adjusted as a function of variance or gradient of the ultrasonic receive signal to reduce undesired side lobes.

Abstract: The preferred embodiments include a diagnostic medical ultrasound system and method for object of interest extraction. Unlike some extraction techniques, the preferred embodiments do not require the user to perform the time consuming task of manually tracing the border of the object of interest. Further, unlike some automatic border detection techniques, the preferred embodiments produce an accurate border of an object of interest even if the object of interest is connected to neighboring anatomy with similar properties.