Keystroke Dynamics for Mobile Phones: A Survey

Abstract

Biometric is the science of authenticating a user based on his physical or behavioral attributes. Keystroke dynamics is behavioral study which analyses the typing rhythm of the user. We adopted a systematic procedure for studying the state of the art in keystroke dynamics in mobile phones. We analyzed the features extracted, the classification techniques, the input text, length of the input text, number of users, hardware used and the results that each study got. We included research articles that focused on keystroke dynamics for mobile devices only. It was found that majority of the research used latency as the prominent feature. Hold time and pressure are also used in combination with latency to get improved results. The most popular classification techniques are either statistical or neural network based, although it is difficult to say which is better since the users, testing conditions and features used are different in all researches. Also the number of users that are used for taking the input are generally less than 100 which is not a good representation sample. The application of this technique is very cost effective as it does not require any extra hardware. Hence there is a need to share the datasets by researchers and develop a standard against which every researcher can compare his results. Also the environment in which the tests are performed should be uncontrolled which will give results that are more realistic and close to real deployment environment.