Multi-modal cardiac diagnostic decision support system and method

1. A method for extracting features from cardiac acoustic signals, comprising the steps of:

obtaining a cardiac acoustic signal;

computing a wavelet representation of the acoustic signal; and

extracting physiologically significant features from the cardiac acoustic signal, using a time delay neural network to process the wavelet representation.

2. The method according to claim 1 , wherein the physiologically significant features correspond to at least one of basic heart sounds and components of the basic heart sounds.

3. A method for evaluating cardiac acoustic signals, comprising the steps of:

obtaining a cardiac acoustic signal;

obtaining a respiration signal;

analyzing the cardiac acoustic signal with a wavelet decomposition to extract time-frequency information;

identifying basic heart sounds using time delay neural networks applied to the extracted time-frequency information; and

interpreting the cardiac acoustic signal using the identified heart sounds and the respiration signal.

4. A program storage device readable by a machine, tangibly embodying a program of instructions executable by the machine to perform method steps for determining cardiac event sequences from cardiac acoustic signals, the method steps comprising:

obtaining a cardiac acoustic signal; and

processing a sequence of features extracted from the cardiac acoustic signal by a probabilistic finite-state automaton to determine a most probable sequence of cardiac events given the cardiac acoustic signal.

5. The program storage device according to claim 4 , wherein the probabilistic finite-state automaton is a hidden markov model.

6. The program storage device according to claim 4 , wherein the step of processing a sequence of features extracted from the cardiac acoustic signal further includes:

computing a wavelet representation of the acoustic signal; and

extracting the features from the cardiac acoustic signal, using a neural network to process the wavelet representation.

7. The program storage device according to claim 4 , further comprising instructions for performing the step of

extracting a plurality of clinical findings from the sequence of cardiac events.

8. The program storage device according to claim 7 , further comprising Instructions for performing the step of determining whether to refer the patient for echocardiography, based upon the extracted clinical findings.

9. The program storage device according to claim 7 , further comprising instructions for performing the step of determining whether to refer the patient for further treatment by a cardiac specialist, based upon the extracted clinical findings.

10. The program storage device according to claim 7 , wherein the clinical findings correspond to finds which are typically derived from auscultation of a heart.

11. The program storage device according to claim 7 , wherein the clinical findings correspond to basic heart sounds and murmurs.

12. The program storage device according to claim 11 , wherein the clinical findings comprise at least one of present intensity, intensity profile, duration, time-alignment, and sequence and spectral quality features.

13. A program storage device readable by a machine, tangibly embodying a program of Instructions executable by the machine to perform method steps for determining a status of heart murmurs, the method steps comprising:

obtaining a cardiac acoustic signal;

detecting a murmur, if any, from the cardiac acoustic signal by processing a sequence of features extracted from the cardiac acoustic signal by probabilistic finite-state automaton to determine a most probable sequence of cardiac events given the cardiac acoustic signal; and

determining whether the murmur is one of functional and pathological, based upon expert rules.

14. A system for extracting features from cardiac acoustic signals, comprising:

a sensor adapted to obtain a cardiac acoustic signal;

means for computing a wavelet representation of the acoustic signal; and

a time delay neural network adapted to process the wavelet representation extract physiologically significant features from the cardiac acoustic signal.

15. The system according to claim 14 , wherein the physiologically significant features correspond to at least one of basic heart sounds and components of the basic heart sounds.

16. A system for evaluating cardiac acoustic signals, comprising:

a sensor adapted to obtain a cardiac acoustic signal;

a respiration sensor adapted to obtain a respiration signal;

a wavelet decomposition circuit adapted to analyze the cardiac acoustic signal to extract time-frequency information;

a time delay neural network adapted to identify basic heart sounds from the extracted time-frequency information; and

evaluating means for interpreting the cardiac acoustic signal using the identified heart sounds and the respiration signal.

17. A system for determining a status of heart murmurs, comprising:

a sensor adapted to obtain a cardiac acoustic signal;

an acoustic signal processor adapted to detect a murmur, if any, from the cardiac acoustic signal by processing a sequence of features extracted from the cardiac acoustic signal using probabilistic finite-state automaton to determine a most probable sequence of cardiac events given the cardiac acoustic signal; and

means for determining whether the murmur is one of functional and pathological, based upon expert rules.

18. A program storage device readable by a machine, tangibly embodying a program of instructions executable by the machine to perform method steps for extracting features from cardiac acoustic signals, the method steps comprising:

obtaining a cardiac acoustic signal;

computing a wavelet representation of the acoustic signal; and

extracting physiologically significant features from the cardiac acoustic signal, using a time delay neural network to process the wavelet representation.

19. The program storage device according to claim 18 , wherein the physiologically significant features correspond to at least one of basic heart sounds and components of the basic hear sounds.

20. A program storage device readable by a machine, tangibly embodying a program of instructions executable by the machine to perform method steps for evaluating cardiac acoustic signals, the method steps comprising:

obtaining a cardiac acoustic signal;

obtaining a respiration signal;

analyzing the cardiac acoustic signal with a wavelet decomposition to extract time-frequency information;

identifying basic heart sounds using time delay neural networks applied to the extracted time-frequency information; and

interpreting the cardiac acoustic signal using the identified heart sounds and the respiration signal.

21. A method for extracting features from cardiac acoustic signals, comprising the steps of:

obtaining a cardiac acoustic signal;

computing a wavelet representation of the acoustic signal;

extracting physiologically significant features from the cardiac acoustic signal, using a neural network to process the wavelet representation; and

processing the physiologically significant features extracted from the cardiac acoustic signal using a probabilistic finite-state automaton to determine a most probable sequence of cardiac events represented by the cardiac acoustic signal.