Autenticação usando Sinais Biométricos: Fundamentos, Aplicações e Desafios

Abstract

Our systems and data (e.g., websites, smartphones, safes, cars, banks, airports) are protected by traditional authentication methods. However, the growing concern about information security and the deployment of smart cities are demanding efforts to find solutions that prevent theft, loss, unauthorized copy, or the forgery of keys, tokens, and passwords. The Internet of Things (IoT) enabled a large increase of personal data collected and published in the Internet. In this context, biometric authentication has earned more and more place as a security solution because they demand, in general, physical presence, vitality, and they are hard to falsify. Among the biometric signals used in academic or commercial products, we mention: iris, face, the palm of hand, fingerprints, walking, voice, electrocardiogram (ECG), electroencephalogram (EEG), and photoplethysmogram (PPG). This chapter brings the state-of-the-art about the use of several biometric signals in an authentication system, contextualizing the applications already developed and the challenges they faced when coexisting with Smart Cities and the Internet of Things. Each type of biometric signal has its own challenges for data acquisition, cost, feature selection, and method to implement the classification. In addition, this chapter presents a review of the machine learning techniques used in biometric systems. The proper choice of the identification technique and classification directly influences results, costs and also the required amount of input data, and the quality of the captured data.

Autenticação Biométrica Baseada em PPG e ECG utilizando Aprendizado Profundo

“…Dentre os principais sinais vitais considerados junto ao estado da arte em autenticação, os sinais Fotopletismografia (PPG) e Eletrocardiograma (ECG) possuem grande potencial de evolução [6]. Consequentemente, houve aumento em pesquisas relacionadas ao uso destes sinais como biometrias para a identificação de usuários [7].…”