METHOD AND APPARATUS FOR RECOGNIZING MOVING ANATOMICAL STRUCTURES USING ULTRASOUND

• US 20120083699A1
• Filed: 06/07/2010
• Published: 04/05/2012
• Est. Priority Date: 06/09/2009
• Status: Active Grant

• Current Assignee: Koninklijke Philips Electronics N.V. (Koninklijke Philips N.V.)
• Original Assignee: Koninklijke Philips Electronics N.V. (Koninklijke Philips N.V.)
• Inventors: Wohlschlager, Markus, Reuter, Stefan, Dubielczyk, Alexander
• Granted Patent: US 9,636,081 B2
• Time in Patent Office: Days
• US Class Current: 600/453
• CPC Class Codes: A61B 5/02411 of foetuses
• CPC Class Codes: A61B 5/4362 Assessing foetal parameters
• CPC Class Codes: A61B 5/7257 using Fourier transforms
• CPC Class Codes: A61B 5/7264 Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems neural networks per se G06N3/00; expert systems per se G06N5/00
• CPC Class Codes: A61B 5/7267 involving training the classification device
• CPC Class Codes: A61B 8/02 Measuring pulse or heart rate
• CPC Class Codes: A61B 8/0866 involving foetal diagnosis; pre-natal or peri-natal diagnosis of the baby
• CPC Class Codes: A61B 8/488 involving Doppler signals
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems

Abstract

A method of recognizing at least one moving anatomical structure using ultrasound data that operates by receiving ultrasound data 100 (100). The ultrasound data comprises Doppler shift information which provides information descriptive of the velocity of at least one anatomical structure. The ultrasound data is first divided into a series of time frames (102). A classification is then assigned to each of the time frames using the Doppler shift information (104). The at least one anatomical structure is then recognized by using the classification of each time frame (106). This is possible, because different anatomical structures produce different patterns in the Doppler shift information.

Claims

1. A method of recognizing at least one moving anatomical structure using ultrasound data, the method comprising the steps of:

receiving ultrasound data (100, 200, 400), wherein the ultrasound data comprises Doppler shift information descriptive of the velocity of the at least one anatomical structure,

dividing the ultrasound data (102, 404) into a series of time frames,

assigning each of the time frames a classification (104) using the Doppler shift information, and

recognizing the at least one anatomical structure (106) using the classification of each time frame.

2. The method of claim 1, wherein the step of assigning each of the time frames a classification using the Doppler shift information comprises the steps of:

identifying fetal heart valve motion data using the Doppler shift information,

identifying fetal heart wall motion data using the Doppler shift information, and

wherein a fetal heart is recognized as the anatomical structure using the classification of each of the time frames.

3. The method of claim 1, wherein the step of assigning each of the time frames a classification using the Doppler shift information comprises the steps of identifying fetal body motion using the Doppler shift information, and wherein the anatomical structure is identified to be a fetal body using the classification of each of the time frames.

4. The method of claim 1, wherein the method further comprises the step of constructing a feature vector (406) for each of the time frames, wherein the classification is assigned using a pattern recognition module (408) that recognizes a classification for each feature vector, and wherein the at least one anatomical structure is determined using the pattern recognition module.

5. The method of claim 4, wherein the pattern recognition module is further adapted for recognizing different types of fetal body motion using the feature vector.

6. The method claim 4, wherein the method further comprises the steps of receiving a measurement from a labor contraction sensor (310) and constructing the feature vector using the measurement from the labor contraction sensor and/or wherein the method further comprises the steps of receiving a phono cardiography measurement from a microphone and constructing the feature vector using the phono cardiography measurement from the microphone and/or wherein the method further comprises receiving a measurement from a electrocardiography system and constructing the feature vector using the measurement from the electrocardiography system and/or wherein the method further comprises receiving a measurement from a pulse oximetry system and constructing the feature vector using the measurement from the pulse oximetry system and/or wherein the method further comprises receiving a measurement from a saturation of peripheral oxygen system and constructing the feature vector using the measurement from the saturation of peripheral oxygen system and/or wherein the method further comprises receiving a measurement from a non-invasive blood pressure system and constructing the feature vector using the measurement from the saturation of non-invasive blood pressure system.

7. The method of claim 4, wherein the pattern recognition module is an implementation of a hidden Markov model.

8. A computer program product comprising machine executable instructions for execution by a control system (312), the machine executable instructions comprising the steps of:

receiving ultrasound data (100, 400), wherein the ultrasound data comprises Doppler shift information descriptive of the velocity of the at least one anatomical structure,

dividing the ultrasound data (102, 404) into a series of time frames, assigning each of the time frames a classification (104) using the Doppler shift information, and

recognizing the at least one anatomical structure (106) using the classification of each time frame.

9. A control system comprising a microprocessor, wherein the microprocessor is programmed to perform the steps of:

receiving ultrasound data (100, 400), wherein the ultrasound data comprises Doppler shift information descriptive of the velocity of the at least one anatomical structure,

dividing the ultrasound data (102, 404) in to a series of time frames,

assigning each of the time frames a classification (104) using the Doppler shift information, and

recognizing the at least one anatomical structure (106) using the classification of each time frame.

10. The control system of claim 9, wherein the step of assigning each sequential time frame a classification using the Doppler shift information comprises the steps of:

identifying fetal heart valve motion data using the Doppler shift information,

identifying fetal heart wall motion data using the Doppler shift information, and

wherein a fetal heart is recognized as the anatomical structure using the classification of each of the time frames.

11. The control system of claim 10, wherein the fetal heart valve motion data is identified using a high pass filter (322) on the ultrasound data and wherein the fetal heart wall motion data is identified using a low pass filter (324), wherein the high pass filter uses a cutoff frequency between 200 Hz and 400 Hz, preferably between 250 and 350 Hz and/or wherein the low pass filter uses a cutoff frequency between 200 Hz and 400 Hz, preferably between 250 and 350 Hz.

12. The control system of claim 10, wherein the step of assigning each time frame a classification using the Doppler shift information comprises the steps of identifying fetal body motion data using the Doppler shift information, and wherein the anatomical structure is identified to be a fetal body using the classification of each of the time frames, and wherein the fetal body motion data is identified using a low pass filter with a cutoff frequency between 1 Hz and 15 Hz, preferably with a cutoff frequency between 8 and 12 Hz.

13. A fetal monitor, comprising:

an ultrasound system adapted for measuring Doppler shifted ultrasound signals using an ultrasonic transducer (308), wherein the ultrasound system is adapted for generating ultrasound data using the Doppler shifted ultrasound signals, and

a control system (312) according to claim 9.

14. The fetal monitor of claim 14, wherein the fetal monitor further comprises a labor contraction sensor (310), wherein the control system is further programmed for performing the steps of:

constructing a feature vector (406) for each of the time frames and the measurement from the labor contraction sensor,

wherein the classification is assigned using a pattern recognition module that recognizes a classification for each feature vector, and wherein the anatomical structure is determined using the pattern recognition module.

15. The fetal monitor of claim 15, wherein the fetal monitor is adapted for indicating the at least one anatomical structure on an indicator and/or wherein the fetal monitor is adapted for warning an operator with an alarm if the at least one anatomical structure does not include a fetal heart and/or wherein the fetal monitor is adapted for displaying a fetal heart rate and a measure of the confidence that the fetal heart rate is from a fetus. Method and apparatus for recognizing moving anatomical structures using ultrasound.