978-1-4673-5637-4/13/$31.00 ©2013 IEEE  Abstract. Actual trends in HCI tend to move systems to mobile environments. Moreover, biometrics is a technology that is entering maturity, getting involved in several security architectures nowadays. Thus, migrating biometrics to mobile scenarios is a trending topic in the research community. Nevertheless, in this kind of systems, usability has been put aside in the intent of produce better performance and it could involve undesirable results. In this work a behavioural biometric modality (handwritten signature recognition) is tested in mobile environments, in order to obtain a complete usability evaluation. Users signed in an iPad with different styluses in different scenarios, correlating performance results with several usability parameters (gathered through video, notes and forms) and obtaining interesting outcomes. Keywords: Biometrics, evaluation, iPad, scenarios, stylus, usability, mobile environments. I. INTRODUCTION iometric recognition is the univocal differentiation of individuals according to special physical or behavioural attributes such as iris, fingerprint or voice. Biometric systems are common in places where high security is needed and are included many times as an accompaniment to other security techniques (smart cards, PIN codes, passwords, etc.). Accessing to a secure area (e.g. using iris or finger recognition) or recognizing passengers in an airport (by using face recognition [1] or fingerprint) are two examples. Usually biometric systems are developed trying to reach the best throughputs but many times, users’ satisfaction is forgotten. Being the final users who interact with products, dissatisfactions can involve misuses, worst results or rejecting to use the technology. There are groups who made studies about usability in biometrics like [2], [3] or [4] (HCI point of view) but there are also additional biometric modalities and usability factors that have to be covered. As long as users’ utilization of technology is moving to mobile scenarios, biometrics should be adapted to them assuring an acceptable behaviour without compromising performance. Therefore, several challenges have to be overcome, such as adapting devices, algorithms and systems architectures. Adapting biometrics to mobile environments is stirring researchers’ interest nowadays [5] and there are various works in this line (e.g. [6] or [7]). This paper continues the previous work done by authors for adapting handwritten signature verification to mobile environments [8] contributing with a comprehensive usability analysis. The algorithm applied is a DTW (Dynamic Time Warping)-based [9] modified for mobile devices. According to this, a usability evaluation of biometrics in mobile environments was done. Furthermore, signatures were collected with a capacitive tablet (Apple iPad) using different styluses (specially made for signing and drawing over mobile devices). In addition various different scenarios where users had to sign in different postures (the most common when using mobile devices) were analyzed. The iPad was chosen due to its popularity and due to be one of the most preferred devices by users [8]. In previous works users had to sign with their finger tip, while in this work they use capacitive styluses, as it was also shown that users prefer to sign with a stylus. The styluses used have different kind of tips, representing the most widely consumed. Then a combination of both factors, iPad and stylus, was tested in order to analyze performance and users’ satisfaction. The main intention of the evaluation is to obtain usability conclusions in order to improve future biometric developments. The measured features were extracted from the ISO 9241-11 [10], where usability is defined as “The extent to which a product can be used by specified users to achieve specified goals with effectiveness, efficiency, and satisfaction in a specified context of use”. Thus, effectiveness, efficiency and satisfaction are the 3 usability factors analyzed. Furthermore, it is feasible to measure learnability as another usability factor, as the evaluation has been carried out in three sessions. To complete the full usability evaluation, different kind of data was collected: the signatures set from 21 users, notes taken by the operator, satisfaction forms filled by users at the end of the evaluation and videos. The whole process was recorded for two reasons: to observe users’ behaviour and to know why wrong interactions occur. Performance results were extracted when the evaluation was completely finished and all the signatures had been collected in a database. The error rates obtained from the algorithm were the FMR (False Match Rate), the FNMR (False Non-Match Rate) and the EER (Equal Error Rate) as three of the most common error rates used in biometrics. FMR expresses the ratio of impostor users who were Usability Evaluation of Biometrics in Mobile Environments Ramon Blanco-Gonzalo 1 , Luis Diaz-Fernandez 1 , Oscar Miguel-Hurtado 2 , and Raul Sanchez-Reillo 1 1 University Carlos III of Madrid, Leganes, España, 2 Incita, Madrid, España, rbgonzal@ing.uc3m.es, ldifernan@gmail.com, oscar.miguel@incita.eu, rsreillo@ing.uc3m.es B 123