Recognition System Using Fusion Normalization Based on Morphological Features of Post-Exercise ECG for Intelligent Biometrics

Figures

• 

  Figure 1. ECG in normal and post-exercise states, after preprocessing.

• 

  Figure 2. User recognition system using normalized ECG.

• 

  Figure 3. (a) Pre-exercise ECG (b) Post-exercise ECG.

• 

  Figure 4. ECG segmented signals based on the R wave peak in pre- and post-exercise cases.

• 

  Figure 5. Block diagram for normalization based on P and T waves.

• 

  Figure 6. Normalization of P and T waves through linear interpolation.

• 

  Figure 7. ECG signal cycle in pre and post-exercise cases as time, frequency, and fusion normalization.

• 

  Figure 8. Conditions for pre- and post-exercise ECG measurement and the morphological features of the measured ECG one cycle. The cycles before and after exercise and compare the user recognition performances before and after normalization.

• 

  Figure 9. (a) ECG cycles before and after exercise (b) ECG cycles before and after exercise: existing normalization (c) ECG cycles before and after exercise: time normalization (d) ECG cycles before and after exercise: fusion normalization.

• 

  Figure 10. Comparative analysis of user recognition performance before and after time normalization, in terms of the number of ECG cycles.

• 

  Figure 11. Comparative analysis of user recognition performance after time normalization, with an increasing number of subjects.

• 

  Figure 12. Comparative analysis of post-exercise ECG normalization recognition performance according to the number of ECG cycles (existing, time, and fusion normalization methods).