Time evolution of face recognition in accessible scenarios

Ramon Blanco-Gonzalo,Norman Poh,Raul Sanchez-Reillo,Rita Wong

doi:10.1186/s13673-015-0043-0

Ramon Blanco-Gonzalo, Norman Poh + Show 2 more

Open Access

https://doi.org/10.1186/s13673-015-0043-0

Copy DOI

Abstract

AbstractUp to now, biometric recognition has shown significant advantages as to be considered a reliable solution for security systems in mobile environments. Nevertheless, due to the short lifetime of biometrics in mobile devices, a handful of concerns regarding usability and accessibility need to be covered in order to meet users’ requirements. This work is focused on analysing the usability and accessibility of a face recognition system used by visually impaired people, focusing on the time spent in the process, which is a critical aspect. Specifically, we cover different key questions including which kind of feedback is more useful for visually impaired users and beneficial for performance and how is the performance evolution in contrast with the time spent in the recognition. Our findings suggest that several parameters improve along with the time spent in the process, including performance. The audio feedback provided in real time involves also better performance and user experience than instructions given previously.

Highlights

Biometrics in mobile environments The amount of sensitive data that needs to be protected, at institutions or companies levels and for ordinary people, is increasing exponentially [10] [23]
It could be thought that biometric recognition is challenging for disabled people but we show in this work that a correct design can make the process easy for everyone
The fact that the results in E2 outperform those of E3 points out that the audio feedback could be more useful than the previous instructions

Summary

Introduction

Biometrics in mobile environments The amount of sensitive data that needs to be protected, at institutions or companies levels and for ordinary people, is increasing exponentially [10] [23]. It is common to use the smartphone to access bank accounts [20], make payments or handle important information in general [7, 8], which leads to the necessity of increase the security in those devices [17, 24]. Biometric recognition is suggested to be embedded in mobile devices for many reasons. The second one is that for some biometric modalities, the capture device is already included within the mobile device (e.g., camera for face recognition, touch screen for handwritten signature recognition, microphone for speaker recognition, or the inclusion of some swipe sensors for fingerprint verification). Handwritten signature, voice and face recognition has been suggested as the most suitable modalities [19]

Methods

Results

Conclusion