Advanced ultrasonic detection of different tissue types

A system for classifying a region of interest as an anatomical feature comprising:

• a processor;
• and
• a non-transitory computer readable medium storing executable instructions executable by the processor comprising:
• an ultrasound interface that receives, from an associated ultrasound system, recorded echoes from the region of interest as a set of radio frequency (RF) signals and constructs an ultrasound image from the RF signals, the RF signals including a set of harmonic signals being extracted from a frequency range associated with second harmonics of the recorded echoes;
• a frequency transform component that transforms the RF signals from a time domain representation to a frequency domain representation;
• a feature extractor that extracts a plurality of features from the frequency domain representation, at least one of the plurality of features representing the set of harmonic signals;
• and
• a classifier that assigns the region of interest to one of a plurality of classes representing anatomical features according to the extracted plurality of features;
• wherein the second program instructions are configured to perform a linear regression of amplitude on frequency from the RF data to provide a regression line and calculate a plurality parameters from the normalized data, comprising:
• a maximum power, a frequency corresponding to the maximum power, a minimum power, a frequency corresponding to the maximum power, a Y-intercept of the regression line, a slope of the regression line, mid-band fit, and an integrated backscatter parameter.

A method for use in an ultrasound system comprising:

• receiving recorded echoes from a region of interest as a set of radio frequency (RF) signals;
• transforming the RF signals from a time domain representation to a frequency domain representation;
• adjusting the frequency domain representation for attenuation as a function of the depth of the region of interest;
• extracting a plurality of features from the frequency domain representation;
• and
• assigning the region of interest to one of a plurality of classes representing anatomical feature according to the extracted plurality of features at a classifier model;
• wherein the RF signals including a set of harmonic signals being extracted from a frequency range associated with second harmonics of the recorded echoes, and extracting a plurality of features from the frequency domain representation comprising extracting at least one feature representing the set of harmonic signals;
• wherein the function of the depth of the region of interest is a first function, associated with a first frequency band, the method further comprising determining a second function associated with a second frequency band, such that adjusting the frequency domain representation for attenuation comprises adjusting the frequency domain representation for attenuation with the first function for signals within the first frequency band and using the second function for signals within the second frequency band.