personal authentication by SINGLE- CHANNEL ecg

Bio- signal A material carrier of the information about the state of the analyzed biological systems.. Find featuresP, Q, R, S, T peaks Statistical data... ECG Acquisition Kardia mobile

Hanoi University of Science and Technology

Telecommunication and Electronics Department

Purposes

•

Ba ck gro un

d i nfo rm ati on

2

•

Blo ck dia gra m

3

•

EC

G a cqu isi tio

n

4

•

Pre -pr oce ssi ng

5

•

Fea ture ext ra cti on

6

•

Cla ssi fca tio n

7

•

Res ult s

8

•

Co ncl usi

on

outline

Biometric authentication:

"Are you indeed

Mr or Mrs A?"

Biometric identifcation

"Who are you?"

Bio- signal

A material carrier of the information about the state of the analyzed biological systems

Give more detailed characteristics about the system

Non- electric bio- signals Electrical bio-signals

1 Background information

ECG

Find features

P, Q, R, S, T peaks

Statistical data

3 ECG Acquisition

 Kardia mobile

o To converts electrical impulses from fngertips into ultrasound signals

transmitted to the mobile device’s microphone

o Specifcations

Frequency Response 0.5Hz - 40 Hz

Battery life 12 months typical use

3 ECG Acquisition

 Kardia mobile

3 ECG Acquisition

 Web plot digitizer

 To digitize the signals into numeric format

 Sample rate: 350 Hz

3 ECG Acquisition

 Web plot digitizer

Experiment set up

Object sit on the chair, put two hands on the table, the

device is in front of the object and next to the phone

4 steps:

• Step 1: Press on “Record your EKG” in Kadia app

• Step 2: Put your fngers on device as Figure 15 and adjust

posture until having continuous signal to start to run

There are 2 seconds to stabilize device and relax

• Step 3: After that, keep posture in 1 minute recording

• Step 4: When recording fnishes, fll up your individual

information as instructor in app

=> Divide into 2 group: the authenticated person and the others.

Filter confguration

Filter: Band-pass flter

• High-pass flter: 0.05 and 0.5 Hz (low-frequency cutoff )

• Low-pass flters : 40, 100, and 150 Hz (high-frequency cutoff)

Cutting segments: from 20s to 50s

Choose

4 Pre- processing

 Pre-processing : Chebychev band-pass flter 0.5- 40 Hz

5 Feature extraction

Scaling function of Daubechies Wavelet:

A progression {αk; kϵZ} satisfying the following four conditions for all integer N≥2:

The expression relating the mother wavelet to the scaling function is:

5 Feature extraction

5 Feature extraction

Decomposition level 4

Find peak >=

60% max value

Let frst peak be R peak

Find the other peaks based on the minimum and maximum

Find peaks inversely on original signal

From R peaks, fnd the other peaks based

on duration of them

Mean values of

P, Q, R, S, T peaks

5 Feature extraction

• Find R peak in decomposition level 4

5 Feature extraction

o Mean

o Median absolute deviation (MAD)

o Standard deviation (SD) o Skewness: a measure for the degree of symmetry in the variable distribution

o Kurtosis: a measure for the degree of tailedness in the variable distribution

Unsupervised

classifcation

Supervised classifcation

Support vector machine (SVM) K- Nearest Neighbor

6 Classif cation

7 Results

o Total data: 150 samples / 60 samples of authenticated

person

120 training data : 60 data of authenticated person, 6

data/ each other ( total 60)

32 testing data: the ratio 16 /16data

o Feature extraction: 11 features

 MeanP, MeanQ, MeanR, MeanS, MeanT

 Mean, Median, SD, MAD, Skewness, Kurtosis

7 Results

Accuracy of train and different validate

⇒ Test/ train ration: 30/ 70

⇒ Weighted KNN, Medium Gausisan SVM

(Test/ train ratio)

KNN

7 Results

Trained model: Medium Gaussian SVM

Trained model: Weighted KNN

8 Conclusion

Classifcation accuracy of SVM is higher than one of KNN

Base on biometric criteria: KNN is better

Some causes affect to results:

T h a n k y o u

References

o Joseph D B.( n.d ) Biomedical engineering fundamental (3rd edition) Trinity college, Hartfort, Conecticuit ,U.S.A.

o https://www.gemalto.com/govt/inspired/biometrics

o http://cimss.ssec.wisc.edu/wxwise/class/aos340/spr00/whatismatlab.htm

o Ruben J.M.D , Neson E.V.P, Erika U.,Wavelet Daubechies (db4) Transform Assessment for WorldView-2 Images Fusion, Distrital

University Francisco José de Caldas, Carrera 7 No 40B – 53, Bogotá D.C., Colombia

o Englehart, K.; Hudgins, B.; Parker, P.A.; Stevenson, M Classifcation of the myoelectric signal using time-frequency based

representations Med Eng Phys 1999, 21, 431–438

o Phinyomark, A.; Phukpattaranont, P.; Limsakul, C Feature reduction and selection for EMG signal classifcation Expert Syst Appl

2012, 39, 7420–7431