Method and apparatus for detecting and identifying a condition

• US 5,267,151 A
• Filed: 09/07/1990
• Issued: 11/30/1993
• Est. Priority Date: 09/07/1990
• Status: Expired due to Fees

• Current Assignee: Frederic M. Ham, Glenn M. Cohen, Samuel P. Kozaitis
• Original Assignee: Frederic M. Ham, Glenn M. Cohen, Samuel P. Kozaitis
• Inventors: Cohen, Glenn M., Ham, Frederic M., Kozaitis, Samuel P.
• Primary Examiner(s): NOT, DEFINED
• Assistant Examiner(s): Tran, Khai
• Application Number:
  US07/578,793
• Time in Patent Office: 1,180 Days
• Field of Search: 364/413.09, 128/633, 128/677, 356/434, 395/22, 395/23
• US Class Current: 600/322
• CPC Class Codes: G01N 21/3577 for analysing liquids, e.g. polluted water
• CPC Class Codes: G01N 2201/1296 using neural networks
• CPC Class Codes: G06F 2218/02 Preprocessing
• CPC Class Codes: G06F 2218/12 Classification; Matching
• CPC Class Codes: G06N 3/045 Combinations of networks
• CPC Class Codes: Y10S 128/922 including image analysis

Abstract

A method and apparatus for sensing and classifying a condition of interest in a system from background noise in which a parameter representative of the condition of interest is sensed and an electrical signal representative of the sensed parameter is produced. The electrical signal is converted into a digital signal, this digital signal containing a signal of interest representative of the condition of interest and background noise. The digital signal is received by an artificial neural network which filters out the background noise to produce a filtered signal from the digital signal, and classifies the signal of interest from the filtered signal to produce an output representative of the classified signal.

Claims

1. A method of filtering and classifying a signal of interest from background noise, comprising:

receiving in an artificial neural network a digital signal containing a signal of interest and background noise;

filtering out the background noise to produce a filtered signal from the digital signal in a feedforward perceptron of the artificial neural network; and

classifying the signal of interest from the filtered signal in a counterpropagation network of the artificial neural network.

2. The method of claim 1, wherein the counterpropagation network is performance invariant to threshold decision settings.

3. The method of claim 1, wherein the signal of interest is representative of a biological condition.

4. The method of claim 3, wherein the biological condition is glucose concentration in blood.

5. The method of claim 1, wherein the feedforward perceptron is a feedforward perceptron trained by backpropagation.

6. A method of sensing and classifying a condition of interest in a system from background noise, comprising:

sensing a parameter representative of the condition of interest and producing an electrical signal representative of the sensed parameter;

converting the electrical signal into a digital signal, said digital signal containing a signal of interest representative of the condition of interest and background noise;

receiving in an artificial neural network the digital signal containing the signal of interest and the background noise;

filtering out the background noise in the artificial neural network to produce a filtered signal from the digital signal; and

classifying the signal of interest from the filtered signal in the artificial neural network to produce an output representative of the classified signal.

7. The method of claim 6, wherein the step of sensing includes performing infrared absorption and evanescent field spectroscopy.

8. The method of claim 6, wherein the step of filtering is performed by a feedforward perceptron.

9. The method of claim 8 wherein the step of classifying is performed by a counterpropagation network.

10. A device for filtering and classifying a signal of interest from background noise, comprising:

means for receiving in an artificial neural network a digital signal containing a signal of interest and background noise;

means for filtering out the background noise from the digital signal to produce a filtered signal; and

means for classifying the signal of interest from the filtered signal.

11. The device of claim 10, wherein the means for filtering is a feedforward perceptron.

12. The device of claim 11, wherein the means for classifying is a counterpropagation network.

13. The device of claim 10, wherein the means for classifying is a counterpropagation network.

14. A device for sensing and classifying a condition of interest in a system from background noise, comprising:

means for sensing a parameter representative of the condition of interest and producing an electrical signal representative of the sensed parameter;

means for converting the electrical signal into a digital signal, said digital signal containing a signal of interest representative of the condition of interest and background noise;

means for receiving in an artificial neural network the digital signal containing the signal of interest and the background noise;

means for filtering out the background noise in the artificial neural network to produce a filtered signal from the digital signal; and

means for classifying the signal of interest from the filtered signal in the artificial neural network to produce an output representative of the classified signal.

15. The device of claim 14, wherein the means for filtering is a feedforward perceptron.

16. The device of claim 15, wherein the means for classifying is a counterpropagation network.

17. The device of claim 14, wherein the means for classifying is a counterpropagation network.

18. A device for measuring a condition of a biological system, comprising:

a sensor that senses a parameter representative of the condition of interest in the biological system, said sensor producing a signal representative of the sensed parameter and containing background noise;

a converter coupled to the sensor to convert the signal produced by the sensor to a digital signal, said digital signal containing a signal of interest and background noise; and

an artificial neural network coupled to the converter, said artificial neural network filtering the background noise from the digital signal to produce a filtered signal and classifying the signal of interest from the filtered signal in the artificial neural network to produce an output representative of the classified signal.

19. The device of claim 18, wherein the sensor includes means for performing infrared absorption and evanescent field spectroscopy.

20. The device of claim 19, wherein the sensor includes an infrared waveguide, an infrared source, and a detector, said infrared source acting on the biological system and providing an infrared signal via said infrared waveguide to the detector, said detector converting said infrared signal into an electronic signal.

21. The device of claim 20, wherein the artificial neural network includes a feedforward perceptron coupled to the converter and which performs the filtering by the artificial neural network, and a counterpropagation network coupled to the feedforward perceptron and which performs the classifying by the artificial neural network.

22. The device of claim 21, wherein the condition of the biological system to be measured is concentration of specific types of molecules in a fluid.

23. The device of claim 21, wherein the condition of the biological system to be measured is concentration of glucose in blood.

24. The device of claim 23, wherein the device is implantable in vivo.

25. The device of claim 18, wherein the device is implantable in vivo.