Different passwords for different websites. Changing them all regularly. Memorising them all. And for all your caution, sites get breached and data is compromised, making logging into websites and accessing services frequently both complicated and insecure. Biometrics seem to offer robust solutions. Voice recognition for home banking, facial recognition embedded in mobile applications, finger print scans to get into your phone – all serve to speed up our interaction with technology.
But, although they liberate us from the tyranny of passwords, these security techniques can be circumvented. Technology to create fully fleshed out security measures requires a comprehensive array of feature sets from which to extract consistent characteristics leading towards identification of the subject. The more variety in the type of data gleaned, the better chance of creating accurate profiles making the augmentation of security more realistic.
The EU-funded AMBER (Enhanced Mobile Biometrics) project has done just this. The project has shown it can identify gender by breaking down gestures and analysing the way in which users swipe screens using multiple datasets. In a recently published paper
the team detail the software and protocol used for data collection, the feature set extracted and subsequent machine learning analysis. The results of this exploratory analysis have confirmed the possibility of sex prediction from the swipe gesture data, obtaining an encouraging 78% accuracy rate using swipe gesture data from two different directions.
The swipe gestures data were captured using a Samsung GT-I9100 ‘Galaxy S2’ smartphone and the researchers focussed on 14 parametres such as average speed, arc distances, angles to start and end, area and length. Participants were instructed to operate the smartphone one-handed (according to their preference) in portrait orientation, using the thumb of the same hand to interact with the screen.
Such data is known as soft biometrics, which can be used to enhance interaction and boost security. Soft-biometrics traits are defined as ‘anatomical or behavioural characteristics that provide some information about the identity of a person, but do not provide sufficient evidence to precisely determine the identity.’ Hard biometrics include finger-print, iris and face recognition.
The team is developing software that can not only identify gender from the way someone swipes the screen but also takes into account the orientation in which they hold the phone and the way the phone moves when it is carried. The beauty of their concept is that the sensors needed to provide the data are already integrated into most phones and tablets.
They explain the predicted soft-biometrics traits can allow touchscreen computers-based systems to tailor their interaction to better suit the user''s characteristics. In addition, this information could also improve the performance of continuous authentication biometrics systems deployed in touchscreen devices.
For more information, please see: project website