US4817015A — High speed texture discriminator for ultrasonic i…

Claim 1

Independent Claim 520 words

A method for classifying a region of interest (ROI) within a portion of a body using at least two statistical properties of ultrasound echoes from said ROI, and for forming an image of said portion of said body with said ROI, said statistical properties being related to physical scattering properties of said ROI, said method comprising creating said image by emitting at least one burst of ultrasound energy with a piezoelectric transducer into said body, including into said portion of said body and said ROI therein, and receiving with transducer means said ultrasound echoes from respective parts of said ROI and other ultrasound echoes from respective parts of the rest of said portion of said body, as acoustic reflections of each said at least one burst from respective parts within said portion of said body and said ROI, said transducer means outputting a plurality of electrical signals respectively corresponding to said acoustic reflections from said respective parts within said portion of said body, amplifying said electrical signals output from said transducer, and using the respective amplified electrical signals from each said burst to create said image, determining, for said classifying of said ROI with said using of said at least two statistical properties, at least two of: (a) a first order statistical parameter b indicating the square of the mean of intensities of all of said ultrasound echoes from said ROI for each respective burst, each said intensity being proportional to the square of the envelope of the respective ultrasound echo; (b) a first order statistical parameter t, wherein said parameter t is a variance in said intensities of said ultrasound echoes from said ROI plus a sum of a total average of said backscatter intensity in said ROI; (c) a second order statistical parameter d, wherein said parameter d is a measure of average spacing of any periodic and semi-periodic array of specular scatterers in said ROI, said statistical parameter d corresponding to an average spatial frequency location of non-redundant peaks of a noise power spectrum corresponding to the square of the absolute magnitude of a Fourier transformation of said intensities of said ultrasound echoes from said ROI; and (d) a second order statistical parameter p, said statistical parameter p being determined only in the event said statistical parameter b is determined as one of said at least two statistical parameters, wherein said statistical parameter p is the variance of said backscatter intensity due to said periodic and semi-periodic array in said ROI plus the sum of the total average of said backscatter intensity from said ROI, said statistical parameter p being formed by integration of a remainder of said noise power spectrum after subtraction of a Gaussian noise envelope term therefrom, and by then subtracting said statistical parameter b from the integrated remainder; and displaying at least two values, each corresponding to at least one respective one of said at least two statistics parameters; wherein said at least two displayed values allow said classifying based on physical characteristics of said ROI, including whether the material in said ROI in said body is in a normal or abnormal state.

Claim 2

Independent Claim 520 words

A method for classifying a region of interest (ROI) within a portion of a body using at least two statistical properties of ultrasound echoes from said ROI, and for forming an image of said portion of said body with said ROI, said statistical properties being related to physical scattering properties of said ROI, said method comprising creating said image by emitting at least one burst of ultrasound energy with a piezoelectric transducer into said body, including into said portion of said body and said ROI therein, and receiving with transducer means said ultrasound echoes from respective parts of said ROI and other ultrasound echoes from respective parts of the rest of said portion of said body, as acoustic reflections of each said at least one burst from respective parts within said portion of said body and said ROI, said transducer means outputting a plurality of electrical signals respectively corresponding to said acoustic reflections from said respective parts within said portion of said body, amplifying said electrical signals output from said transducer, and using the respective amplified electrical signals from each said burst to create said image, determining, for said classifying of said ROI with said using of said at least two statistical properties, at least two of: (a) a first order statistical parameter b indicating the square of the mean of intensities of all of said ultrasound echoes from said ROI for each respective burst, each said intensity being proportional to the square of the envelope of the respective ultrasound echo; (b) a first order statistical parameter t, wherein said parameter t is a variance in said intensities of said ultrasound echoes from said ROI plus a sum of a total average of said backscatter intensity in said ROI; (c) a second order statistical parameter d, wherein said parameter d is a measure of average spacing of any periodic and semi-periodic array of specular scatterers in said ROI, said statistical parameter d corresponding to an average spatial frequency location of non-redundant peaks of a noise power spectrum corresponding to the square of the absolute magnitude of a Fourier transformation of said intensities of said ultrasound echoes from said ROI; and (d) a second order statistical parameter p, said statistical parameter p being determined only in the event said statistical parameter b is determined as one of said at least two statistical parameters, wherein said statistical parameter p is the variance of said backscatter intensity due to said periodic and semi-periodic array in said ROI plus the sum of the total average of said backscatter intensity from said ROI, said statistical parameter p being formed by integration of a remainder of said noise power spectrum after subtraction of a Gaussian noise envelope term therefrom, and by then subtracting said statistical parameter b from the integrated remainder; and displaying at least two values, each corresponding to at least one respective one of said at least two statistics parameters; wherein said at least two displayed values allow said classifying based on physical characteristics of said ROI, including whether the material in said ROI in said body is in a normal or abnormal state.

Claim 3

Dependent Claim 43 words

The method of claim 1, comprising use of a plurality of said bursts for providing respective ones of said ultrasound echoes for said determining of said statistical parameters, and wherein each said determined parameter d is obtained from an average noise power spectrum.

Claim 4

Dependent Claim 41 words

The method of claim 1, wherein each said statistical parameter is determined with respect to a range direction of a scanning beam for providing each said burst and said ultrasound echoes, said ROI having a corresponding depth in said range direction.

Claim 5

Dependent Claim 35 words

The method of claim 1, comprising use of a plurality of said bursts, wherein each said determined statistical parameter is with respect to a lateral direction transverse to a direction of propagation of said bursts.

Claim 6

Dependent Claim 25 words

The method of claim 1, wherein each said parameter d and said parameter p that is determined corresponds to an autocorrelation function of said intensities.

Claim 7

Dependent Claim 35 words

The method of claim 1, comprising eliminating non-Gaussian interference due to blood vessels in said ROI to increase accuracy of said determining of each said at least two statistical parameters b, t, d and p.

Claim 8

Dependent Claim 38 words

The method of claim 1, comprising freezing said image of said ROI for determining at a slower rate each of said at least two statistical parameters for said ROI, and for analysis of material moving in said ROI.

Claim 9

Dependent Claim 22 words

The method of claim 1, wherein all of said statistical parameters b, t, d, and p are determined for each said ROI.

Claim 10

Dependent Claim 73 words

A device for performing the method of claim 1, comprising: memory means for storing data including pixel data of said image, each said pixel data being the magnitude of the envelope of a respective one of said ultrasonic echoes and means for squaring said pixel data, wherein said squared data is also stored in said memory means, and two of said statistics parameters are determined by said processing means from said squared data.

Claim 11

Dependent Claim 50 words

The method of claim 1, comprising using an average of a plurality of noise power spectra including said noise power spectrum, for said determining of said at least two statistics parameters, when said at least two determined statistical parameters include at least one of said statistical parameters d and p.

Claim 12

Dependent Claim 61 words

The method of claim 1, comprising performing on-line in real time said determining of said at least two statistical parameters and said displaying of said at least two respective values, for successive displaying of ultrasound images including said ROI with successive displays of said at least two respective values, for said classifying of the body under investigation within each said ROI.

Claim 13

Dependent Claim 338 words

A texture discriminator device for classification of texture of an ultrasound image in a body under investigation for any selected region of interest (ROI), comprising: a master clock for determining pace of operation of the texture discriminator, an ROI memory for storing pixel data of said image from within said region of interest, and means for updating said ROI memory under control of said master clock, an ROI address counter to determine the memory address of said ROI memory for retrieval and renewal of contents of said ROI memory, means for squaring said pixel data stored by said ROI memory and means for transmission of said squared data for further processing, an intensity memory wherein the squared data are stored for further processing, a signal processor for determining at least two of four statistical parameters b, t, d and p from said data stored in said intensity memory for said ROI, and means for displaying at least two respective values for analysis of normal and abnormal states of constituent matter in said ROI, each said respective value corresponding to at least one respective one of said at least two statistical parameters, wherein; said parameter b corresponds to a squared mean of intensity of said image and represents a measure of mean ultrasound backscatter intensity; said parameter t corresponds to a variance in the intensity image plus a sum of a total average of said backscatter intensity; said parameter d corresponds to a measure of average spacing of any periodic and semi-periodic array of tissue specular scatterers and is obtained from an average spatial frequency location of non-redundantpeaks of a noise power spectrum of said ROI; and said parameter p corresponds to the variance of said backscatter intensity due to said periodic and semi-periodic tissue array plus the sum of the total average of said backscatter intensity, and is obtained by integrating the remainder of said noise power spectrum after subtraction of a Gaussian noise envelope term therefrom, and by then subtracting the parameter b from the integrated remainder.

Claim 14

Dependent Claim 338 words

A texture discriminator device for classification of texture of an ultrasound image in a body under investigation for any selected region of interest (ROI), comprising: a master clock for determining pace of operation of the texture discriminator, an ROI memory for storing pixel data of said image from within said region of interest, and means for updating said ROI memory under control of said master clock, an ROI address counter to determine the memory address of said ROI memory for retrieval and renewal of contents of said ROI memory, means for squaring said pixel data stored by said ROI memory and means for transmission of said squared data for further processing, an intensity memory wherein the squared data are stored for further processing, a signal processor for determining at least two of four statistical parameters b, t, d and p from said data stored in said intensity memory for said ROI, and means for displaying at least two respective values for analysis of normal and abnormal states of constituent matter in said ROI, each said respective value corresponding to at least one respective one of said at least two statistical parameters, wherein; said parameter b corresponds to a squared mean of intensity of said image and represents a measure of mean ultrasound backscatter intensity; said parameter t corresponds to a variance in the intensity image plus a sum of a total average of said backscatter intensity; said parameter d corresponds to a measure of average spacing of any periodic and semi-periodic array of tissue specular scatterers and is obtained from an average spatial frequency location of non-redundantpeaks of a noise power spectrum of said ROI; and said parameter p corresponds to the variance of said backscatter intensity due to said periodic and semi-periodic tissue array plus the sum of the total average of said backscatter intensity, and is obtained by integrating the remainder of said noise power spectrum after subtraction of a Gaussian noise envelope term therefrom, and by then subtracting the parameter b from the integrated remainder.

Claim 15

Dependent Claim 83 words

The texture discriminator of claim 12, wherein said signal processor comprises: means for high speed implementation of a Fourier transform, to obtain said noise power spectrum for determining said parameter d and said parameter p, for determining said Gaussian noise envelope to be subtracted from said noise power spectrum for determining said parameter p from said noise power spectrum and said parameter b; and a memory means for storing interstage measurements and a predetermined Gaussian noise function corresponding to said Gaussian noise envelope.

Claim 16

Dependent Claim 20 words

The texture discriminator of claim 13, comprising means for retrofittedly connecting said texture discriminator to a medical ultrasound B-scan system.

Claim 17

Dependent Claim 25 words

The texture discriminator of claim 12, comprising means for eliminating non-Gaussian interference for high speed analysis for said classifying of said matter in said ROI.

Claim 18

Dependent Claim 33 words

The texture discriminator of claim 12, comprising means for selective freeze frame operation of said image to allow said analysis of matter that moves in and out of said ROI in real time.

Claim 19

Dependent Claim 18 words

The texture discriminator of claim 12, comprising means for selectively varying the size and shape of said ROI.

Claim 20

Independent Claim 409 words

A method for classifying a region of interest (ROI) within a portion of a body using at least two statistical properties of ultrasound echoes from said ROI, and for forming an image of said portion of said body with said ROI, said statistical properties being related to physical scattering properties of said ROI, said method comprising creating said image by emitting at least one burst of ultrasound energy with a piezoelectric transducer into said body, including into said portion of said body and said ROI therein, and receiving with transducer means said ultrasound echoes from respective parts of said ROI and other ultrasound echoes from respective parts of the rest of said portion of said body, as acoustic reflections of each said at least one burst from respective parts within said portion of said body and said ROI, said transducer means outputting a plurality of electrical signals respectively corresponding to said acoustic reflections from said respective parts within said portion of said body, amplifying said electrical signals output from said transducer, and using the respective amplified electrical signals from each said burst to create said image, determining, for said classifying of said ROI with said using of said at least two statistical properties, at least two of: (a) a first order statistical parameter b indicating the square of the mean of intensities of all of said ultrasound echoes from said ROI for each respective burst, each said intensity corresponding to the square of the envelope of the respective ultrasound echo; (b) a first order statistical parameter t indicating the mean of the squares of all of said intensities of said ultrasound echoes from said ROI; (c) a second order statistical parameter d indicating an average spacing of any periodic and semi-periodic array of specular scatterers in said ROI, said parameter d being obtained from non-redundant peaks of a noise power spectrum corresponding to a squared Fourier transformation of said intensities of said ROI; and (d) a second order statistical parameter p corresponding to the integral of a remainder of a subtraction of a Gaussian noise envelope from said noise power spectrum, less said statistical parameter b; and displaying at least two values, each corresponding to at least one respective one of said at least two statistical parameters; wherein said at least two respective displayed values allow said classifying of the tissue of said region of interest in said portion of said body as being in normal and abnormal states.

Claim 21

Independent Claim 409 words

A method for classifying a region of interest (ROI) within a portion of a body using at least two statistical properties of ultrasound echoes from said ROI, and for forming an image of said portion of said body with said ROI, said statistical properties being related to physical scattering properties of said ROI, said method comprising creating said image by emitting at least one burst of ultrasound energy with a piezoelectric transducer into said body, including into said portion of said body and said ROI therein, and receiving with transducer means said ultrasound echoes from respective parts of said ROI and other ultrasound echoes from respective parts of the rest of said portion of said body, as acoustic reflections of each said at least one burst from respective parts within said portion of said body and said ROI, said transducer means outputting a plurality of electrical signals respectively corresponding to said acoustic reflections from said respective parts within said portion of said body, amplifying said electrical signals output from said transducer, and using the respective amplified electrical signals from each said burst to create said image, determining, for said classifying of said ROI with said using of said at least two statistical properties, at least two of: (a) a first order statistical parameter b indicating the square of the mean of intensities of all of said ultrasound echoes from said ROI for each respective burst, each said intensity corresponding to the square of the envelope of the respective ultrasound echo; (b) a first order statistical parameter t indicating the mean of the squares of all of said intensities of said ultrasound echoes from said ROI; (c) a second order statistical parameter d indicating an average spacing of any periodic and semi-periodic array of specular scatterers in said ROI, said parameter d being obtained from non-redundant peaks of a noise power spectrum corresponding to a squared Fourier transformation of said intensities of said ROI; and (d) a second order statistical parameter p corresponding to the integral of a remainder of a subtraction of a Gaussian noise envelope from said noise power spectrum, less said statistical parameter b; and displaying at least two values, each corresponding to at least one respective one of said at least two statistical parameters; wherein said at least two respective displayed values allow said classifying of the tissue of said region of interest in said portion of said body as being in normal and abnormal states.

Claim 22

Dependent Claim 132 words

The method of claim 18, wherein a plurality of said bursts of ultrasound are provided for said classifying of said ROI within said portion of said body under investigation, to produce respective ones of said ultrasound echoes with each said ultrasound echo corresponding to a respective position within said body including said ROI, said image is formed by displaying a data corresponding to each said ultrasound echo at a respective position of the image, and said determining of said second order statistical parameter d is based on an average of noise power spectra including said noise power spectrum, each said noise power spectrum being provided by a respective square of a Fourier transformation of respective ultrasound echoes from respective positions within said body for each respective one of said plurality of bursts.

Claim 23

Dependent Claim 139 words

The method of claim 18, wherein a plurality of said bursts of ultrasound are provided for said classifying of said ROI within said proportion of said body under investigation, to produce respective ones of said ultrasound echoes with each said ultrasonic echo corresponding to a respective position within said body including said ROI, said image is formed by displaying a data corresponding to each said ultrasonic echo at a respective position of the image, and said Fourier transform, the square of which is used to provide said statistical parameter d, is of a plurality of respective intensities from different ones of said bursts, corresponding to respective ultrasound echoes from respective points in said portion of said body with said ROI which are aligned largely transversely to the respective ones of said points corresponding to any one of said bursts.

Annuity Stage	Window Opens	Payment Deadline	Surcharge Period Until	Estimated Fee	Status
3.5 Years (Year 4)	1992-03-27	1992-09-26	1993-03-28	$2000.00	Window Closed / Lapsed
7.5 Years (Year 8)	1996-03-27	1996-09-26	1997-03-28	$3760.00	Window Closed / Lapsed
11.5 Years (Year 12)	2000-03-27	2000-09-26	2001-03-28	$7700.00	Window Closed / Lapsed

High speed texture discriminator for ultrasonic imaging

Abstract

Inventors

Bibliographic Details

Claims Analysis

References Cited (Backward Citations)

USPTO Annuity & Maintenance Windows

Assigned Categories & Tags

Assign Category Tag