Health monitoring device and human electric signal processing method

• US 8,498,696 B2
• Filed: 06/08/2010
• Issued: 07/30/2013
• Est. Priority Date: 06/15/2009
• Status: Active Grant

• Current Assignee: National Taiwan University (Government of The Republic of China)
• Original Assignee: National Taiwan University (Government of The Republic of China)
• Inventors: Yen, Jia-Yush, Lee, Chih-Kung, Lin, Chii-Wann, Wu, Kuang-Chong, Yen, Fan-Che, Tseng, Ching-En
• Primary Examiner(s): Koharski, Christopher D
• Assistant Examiner(s): OROPEZA, FRANCES P
• Application Number:
  US12/796,634
• Publication Number:
  US 20100317934A1
• Time in Patent Office: 1,148 Days
• Field of Search: 600/508, 600/509
• US Class Current: 600/509
• CPC Class Codes: A61B 5/316 Modalities, i.e. specific diagnostic methods
• CPC Class Codes: A61B 5/346 Analysis of electrocardiograms
• CPC Class Codes: A61B 5/35 by template matching
• CPC Class Codes: A61B 5/726 using Wavelet transforms

Abstract

A health monitoring device is provided and includes a detecting unit for detecting a first electric signal from a human body and a processing unit for performing the following steps: receiving the first electric signal; performing wavelet transform on the first electric signal, to filter frequency bands thereof which include undesired noises, and then perform a reverse wavelet transform to obtain a second electric signal; combining frequency bands of the first electric signal which include primary characteristics and then perform a reverse wavelet transform on the combined frequency bands to obtain a fourth electric signal; and comparing the fourth electric signal with a plurality of electric signal patterns of a database to determine whether the fourth electric signal is an electric signal for healthiness.

Claims

1. A health monitoring device, comprising:

a detecting unit for detecting a first electric signal from a human body; and

a processing unit for performing the following steps;

receiving the first electric signal;

performing wavelet transform on the first electric signal, to filter frequency bands thereof which include undesired noises, and then perform reverse wavelet transform to obtain a second electric signal;

combining frequency bands of the first electric signal which include desired primary characteristics and then performing reverse wavelet transform on the combined frequency bands to obtain a third electric signal;

obtaining period information of the second electric signal from the third electric signal, and then re-sampling the second electric signal according to the period information to obtain a fourth electric signal; and

comparing the fourth electric signal with a plurality of electric signal patterns in a database to determine whether the fourth electric signal is an electric signal for healthiness.

2. The health monitoring device as claimed in claim 1, wherein the processing unit further performs the following steps:

performing spectral estimation on the fourth electric signal to obtain a plurality of characteristic coefficients;

comparing the characteristic coefficients with a plurality of characteristic coefficient patterns to determine whether the fourth electric signal is an electric signal for healthiness.

3. The health monitoring device as claimed in claim 1, wherein the processing unit further performs the following steps:

performing normalization on one of the first, second and fourth electric signals so that the magnitude range of the electric signal is consistent with the magnitude range of the electric signal patterns in the database.

4. The health monitoring device as claimed in claim 1, wherein the frequency bands which include undesired noises in the first electric signal are low frequency drift bands and high frequency noise bands.

5. The health monitoring device as claimed in claim 1, wherein the frequency bands of the second bands which include primary characteristics comprise one of a QRS band, P band, T band, or U band.

6. The health monitoring device as claimed in claim 1, wherein the first electric signal is one of an electrocardiogram signal, an electroencephalogram signal, an electromyogram signal or a blood sugar concentration signal.

7. The health monitoring device as claimed in claim 2, wherein spectral estimation is performed by using Autoregressive Model (AR Model).

8. The health monitoring device as claimed in claim 2, wherein step of comparing the fourth electric signal with a plurality of electric signal patterns in a database and step of comparing the characteristic coefficients with a plurality of characteristic coefficient patterns are performed by a Support Vector Machine (SVM).

9. The health monitoring device as claimed in claim 8, wherein the SVM is performed by using a mapping function to map information to a high dimension space, and the mapping function is a Radial-basis function (RBF) Kernel function.

10. The health monitoring device as claimed in claim 1, further comprising a wireless unit, wherein the wireless unit is coupled to the processing unit and transmits the processing unit result by wireless communications, determining whether the fourth electric signal is an electric signal for healthiness.

11. A human electric signal processing method, comprising:

receiving a first electric signal of a human body;

performing wavelet transform on the first electric signal, to filter frequency bands thereof which include undesired noises, and then perform reverse wavelet transform to obtain a second electric signal;

combining frequency bands of the first electric signal which include primary characteristics and then performing reverse wavelet transform on the combined frequency bands to obtain a third electric signal;

obtaining period information of the second electric signal from the third electric signal, and then re-sampling the second electric signal according to the period information to obtain a fourth electric signal.

12. The human electric signal processing method as claimed in claim 11, further comprising:

performing spectral estimation on the fourth electric signal to obtain a plurality of characteristic coefficients.

13. The human electric signal processing method as claimed in claim 12, further comprising:

comparing the fourth electric signal with a plurality of electric signal patterns in a database to determine whether the fourth electric signal is an electric signal for healthiness.

14. The human electric signal processing method as claimed in claim 13, further comprising:

performing normalization on one of the first, second and fourth electric signals so that the magnitude range of the electric signal is consistent with the magnitude range of the electric signal patterns in the database.

15. The human electric signal processing method as claimed in claim 11, wherein the frequency bands which include undesired noises in the first electric signal are low frequency drift bands and high frequency noise bands.

16. The human electric signal processing method as claimed in claim 11, wherein the frequency bands of the second bands which include primary characteristics comprises one of a QRS band, P band, T band, or U band.

17. The human electric signal processing method as claimed in claim 11, wherein the first electric signal is one of an electrocardiogram signal, an electroencephalogram signal, an electromyogram signal or a blood sugar concentration signal.

18. The human electric signal processing method as claimed in claim 12, wherein spectral estimation is performed by using Autoregressive Model (AR Model).

19. The human electric signal processing method as claimed in claim 13, wherein step of comparing the fourth electric signal with a plurality of electric signal patterns in a database and step of comparing the characteristic coefficients with a plurality of characteristic coefficient patterns are performed by a Support Vector Machine (SVM).

20. The human electric signal processing method as claimed in claim 19, wherein the SVM performs a mapping function to map information to a high dimension space, and the mapping function is a Radial-basis function (RBF) Kernel function.