Use of autoencoders as a generalization method for anti-spoofing classification

Watman Outon, Tomás Ernesto (2019). Use of autoencoders as a generalization method for anti-spoofing classification. Thesis (Master thesis), E.T.S. de Ingenieros Informáticos (UPM).

Description

Title: Use of autoencoders as a generalization method for anti-spoofing classification
Author/s:
  • Watman Outon, Tomás Ernesto
Contributor/s:
  • Segovia Pérez, Francisco Javier
Item Type: Thesis (Master thesis)
Masters title: Data Science
Date: June 2019
Subjects:
Faculty: E.T.S. de Ingenieros Informáticos (UPM)
Department: Lenguajes y Sistemas Informáticos e Ingeniería del Software
Creative Commons Licenses: Recognition - No derivative works - Non commercial

Full text

[img]
Preview
PDF - Requires a PDF viewer, such as GSview, Xpdf or Adobe Acrobat Reader
Download (2MB) | Preview

Abstract

Computer vision is one of the hot topics nowadays when talking about artificial intelligence, covering a wide range of problems under its domain, from document verification to object detection. One of these problems consists of detecting spoofing attacks, masquerading as a registered user in a face-recognition system, gaining illegitimate access and advantages, used in face-recognition payments, phone unlocking, workers checking –in… When training models focused on solving this kind of problems, one of the main concerns is the data we have to deal with due to its small variety, leading to generalization problems. Some of the main approaches that try to tackle these overfitting issues include dropout techniques, model variations or adding some kind of regularization. While in general the way of adding regularization is by introducing a regularization parameter in the objective function, in this thesis we will be presenting another equivalent regularization algorithm.

More information

Item ID: 62937
DC Identifier: http://oa.upm.es/62937/
OAI Identifier: oai:oa.upm.es:62937
Deposited by: Biblioteca Facultad de Informatica
Deposited on: 08 Jul 2020 08:52
Last Modified: 08 Jul 2020 08:52
  • Logo InvestigaM (UPM)
  • Logo GEOUP4
  • Logo Open Access
  • Open Access
  • Logo Sherpa/Romeo
    Check whether the anglo-saxon journal in which you have published an article allows you to also publish it under open access.
  • Logo Dulcinea
    Check whether the spanish journal in which you have published an article allows you to also publish it under open access.
  • Logo de Recolecta
  • Logo del Observatorio I+D+i UPM
  • Logo de OpenCourseWare UPM