Qualificação de doutorado do aluno Pedro Silva, dia 30/03/2021, as 14 horas.

Qualificação de doutorado do aluno Pedro Silva, dia 30/03/2021, as 14 horas.

Título: Rethink and Redesign Deep Learning for Biometrics.
Resumo: The concept of biometric systems is a well-known subject in the human being's daily life. The efforts to enhance those systems are increasing every year due to robustness and the ease of their use. However, the system based on a biometric modality does not have a performance close to perfection in non-cooperative environments when compared to cooperative ones, and complexity is observed in the approaches proposed in the literature. Due to that fact, new studies are developed to enhance biometric-based systems performance by creating new manners of teaching a machine learning algorithm and new representations. Currently, several researchers are directing their efforts to develop new metric learning approaches for deep learning architectures for a wide range of problems, including biometrics. In this work, the proposal is to develop a loss function that uses concepts from biometrics. A new representation for data is also presented, which will be useful for fusing biometric modalities and reducing the footprint of current deep networks. This new representation based on Hilbert Curve will be used as a new layer for deep learning models. The objective is to convert 2D images to 1D signals and fusion this new representation with an ECG signal at feature-level and adjust the metric learning for two or more modalities at the same time during the model training. Besides, it is also addressed in this work an open problem in literature of how to create synthetic databases from two or more other databases. The goal is to improve the accuracy of biometric systems, with both approaches proposed in this work.
Orientador: Prof. Dr. Gladston Moreira.
Coorientador: Prof. Dr. Eduardo Luz.
Banca: Prof. Dr. Rodrigo Silva (UFOP-PPGCC); Prof. Dr. Luciano Oliveira (UFBA); Prof. Dr.Thiago Santos (UFES).
Data: 30/03/2021 (Terça-feira).
Horário: 14hs.
Link: https://meet.google.com/fzx-vzmc-fei

Departamento de Computação  |  ICEB  |  Universidade Federal de Ouro Preto
Campus Universitário Morro do Cruzeiro  |  CEP 35400-000  |  Ouro Preto - MG, Brasil
Telefone: +55 31 3559-1692  |  decom@ufop.edu.br