Skip to Main content Skip to Navigation
Theses

Biométrie par signaux physiologiques

Abstract : In general, biometrics aims to identify individuals from their biological characteristics. This practice tends to replace the traditional methods of identity verification of individuals, among others, passwords and security codes. Nowadays, biometrics found wide application and research of new biometric methods is topical. The objective of this thesis is to develop and evaluate new biometric methods based on tamper-proof characteristics that can not be changed voluntarily. In this context, the physiological signals are considered. Thus, we proposed three methods of biometric identification. The first method uses the electrocardiogram (ECG) as individual signature, while the second is based on the use of surface electromyography signals (EMG) in response to a force of fixed intensity. The final technique explored, uses the motor responses obtained after electrical stimulation. These methods consist first to acquire the physiological signals in healthy people. These signals are modeled by wavelets networks to extract relevant features. The identification phase is performed automatically by neural networks. According to the results obtained from experiments performed, the proposed methods lead to interesting performance identification. The first method, using the electro-cardiographic signal, achieves a recognition rate of 92%, while the identification by EMG signals, in response to a force of a fixed intensity, allows a correct identification of 80 %. Finally, a performance of 95% is obtained by identification by motor response. For these three techniques explored, the robustness to noise ratio was studied
Document type :
Theses
Complete list of metadatas

Cited literature [56 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-00778089
Contributor : Abes Star :  Contact
Submitted on : Friday, January 18, 2013 - 4:37:11 PM
Last modification on : Wednesday, September 4, 2019 - 1:52:09 PM
Long-term archiving on: : Friday, April 19, 2013 - 4:04:28 AM

File

TH2011PEST1009_complete.pdf
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-00778089, version 1

Collections

Citation

Samer Chantaf. Biométrie par signaux physiologiques. Autre. Université Paris-Est, 2011. Français. ⟨NNT : 2011PEST1009⟩. ⟨tel-00778089⟩

Share

Metrics

Record views

1099

Files downloads

6601