Smartphone Impostor Detection with Behavioral Data Privacy and Minimalist Hardware Support

Abstract

Impostors are attackers who take over a smartphone and gain access to the legitimate user's confidential and private information. This paper proposes a defense-in-depth mechanism to detect impostors quickly with simple Deep Learning algorithms, which can achieve better detection accuracy than the best prior work which used Machine Learning algorithms requiring computation of multiple features. Different from previous work, we then consider protecting the privacy of a user's behavioral (sensor) data by not exposing it outside the smartphone. For this scenario, we propose a Recurrent Neural Network (RNN) based Deep Learning algorithm that uses only the legitimate user's sensor data to learn his/her normal behavior. We propose to use Prediction Error Distribution (PED) to enhance the detection accuracy. We also show how a minimalist hardware module, dubbed SID for Smartphone Impostor Detector, can be designed and integrated into smartphones for self-contained impostor detection. Experimental results show that SID can support real-time impostor detection, at a very low hardware cost and energy consumption, compared to other RNN accelerators.

Deep Learning and Machine Learning, Better Together Than Apart: A Review on Biometrics Mobile Authentication

“…CNN has proven to be exceptional not only in authentication contexts across both behavioral and physiological dynamics, but also as a feature extractor capable of enhancing the classification performance of many other algorithms. Most notably, hybrid authentication schemes have been observed to be a compelling and advancing architecture for classification models, resulting in higher accuracies than standalone models in many studies such as [35,67,[69][70][71].…”

“…Many studies [66][67][68] chose LSTM as a viable algorithm to pair with gait mobile authentication schemes. In study [66], a gait mobile authentication method utilizing three-axis accelerometer data was proposed.…”

“…For gait authentication, their model was able to achieve a higher than 95% accuracy for almost all subjects in pocket position and 90% or higher in handholding position. Researchers of [67] continued their previous research on their smartphone imposter detection (SID) model, focusing on deep learning algorithms and protecting user data privacy. This study utilized the WALK section of the HAPT dataset, consisting of motion sensor data collected from 30 participants with smartphones strapped to their waists.…”