"A biometric recognition system through photoplethysmogram (PPG) signal" by Lea Monica B. Alonzo

A biometric recognition system through photoplethysmogram (PPG) signals

Date of Publication

4-2019

Document Type

Master's Thesis

Degree Name

Master of Science in Manufacturing Engineering

Subject Categories

Manufacturing

College

Gokongwei College of Engineering

Department/Unit

Manufacturing Engineering and Management

Thesis Adviser

Homer S. Co

Defense Panel Chair

Nilo T. Bugtai

Defense Panel Member

Renann G. Baldovino
Luis Miguel F. Bañuelos

Abstract/Summary

The development of a biometric recognition system using photoplethysmogram (PPG) signal is presented in this study. Empirical mode decomposition (EMD) and power spectral density (PSD) of the PPG signals were tested for performance as the biometric traits. K-nearest neighbors algorithm (KNN), support vector machine (SVM), and random forest (RF) were the primary classifiers tested. An algorithm was made to train, test, and k-fold cross-validate data both from public and local database. Trained data was also used for live testing. The system was able to acquire PPG data of a user using Contec CMS 50D+ pulse oximeter and store the data to a desktop using Python. A graphical user interface was made to allow two main functions, which are enrollment and recognition.
Results from data using public database, local database, and live testing showed varying performances The system is less accurate in recognizing live data. However, it produced positive performance when tested on previously stored data from public and local database. It can then be concluded that PPG can be used for biometric recognition system and the weaknesses of the produced system may be addressed through gathering and training with of larger sets of data.

Abstract Format

html

Language

English

Format

Electronic

Accession Number

CDTG008116

Keywords

Biometric identification; Plethysmography; Machine learning; Hilbert-Huang transform

Upload Full Text

wf_no

Embargo Period

1-7-2025

This document is currently not available here.

Share

COinS