| KEYWORD |
Deep learning for Biometric Identification using PPG signals
keywords ARTIFICIAL INTELLIGENCE, AUTOENCODERS, BIOMETRIC IDENTIFICATION, BIOSIGNAL ANALYSIS, CONVOLUTIONAL NEURAL NETWORKS, DEEP LEARNING, DEEP NEURAL NETWORKS, EMBEDDED SYSTEMS, ENERGY EFFICIENCY, LOW POWER, MICROCONTROLLERS, PPG
Reference persons DANIELE JAHIER PAGLIARI
External reference persons Alessio Burrello (Politecnico di Torino)
Research Groups DAUIN - GR-06 - ELECTRONIC DESIGN AUTOMATION - EDA, ELECTRONIC DESIGN AUTOMATION - EDA, GR-06 - ELECTRONIC DESIGN AUTOMATION - EDA
Thesis type EXPERIMENTAL, SOFTWARE DEVELOPMENT
Description Biometric identification systems based on wearable devices are developing quickly in recent years However, the involved machine learning algorithms need dedicated training on each subject in order to obtain good accuracy. In particular, biometric recognition based on machine learning and photoplethysmography (PPG) signals are currently based on the recognition of a subject already seen during training.
The goal of this thesis is to develop a model that does not need further training in order to be extended to new subjects, but only a small amount of reference data.
In particular, the candidate will have to train neural networks able to project the data of a new subject, never seen during training, onto a multi-dimensional space, in which the biometric trace is perfectly distinguishable from all other subjects.
The final product of the thesis will be the validation of the proposed method on public datasets, and the implementation of the complete application on a wearable device.
Interested candidates must send an email to daniele.jahier@polito.it attaching their CV and exams' transcript with scores.
Required skills Required skills include C and Python programming, and a minimal familiarity with machine/deep learning concepts and the corresponding models.
Notes Thesis in collaboration with Prof. Luca Benini’s research group at the University of Bologna and ETH Zurich. The thesis can be carried out either in Torino or in one of the other two universities.
Deadline 30/11/2023
PROPONI LA TUA CANDIDATURA