SmartCAMPP - Smartphone-based continuous authentication leveraging motion sensors with privacy preservation

Hernández–Álvarez, Luis; Fuentes, José María de; González-Manzano, Lorena; Encinas, Luis Hernández

doi:10.1016/j.patrec.2021.04.013

Cited by 19 publications

(14 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the context of user biometric authentication, both ML and DL have been shown to be useful tools. Both techniques have produced good results when using different biometric features, including: (1) behavioral characteristics with gyroscopes, accelerometers, and magnetometers [ 7 , 8 ], (2) physical attributes such as facial [ 9 ], ocular [ 10 ], or fingerprint [ 11 ] recognition, and (3) physiological signals such as electroencephalograms (EEG) [ 12 ] or electrocardiograms [ 13 ]. It is important to note the difference between the user identification problem and the user authentication problem.…”

Section: Introductionmentioning

confidence: 99%

EEG Authentication System Based on One- and Multi-Class Machine Learning Classifiers

Hernández–Álvarez

Barbierato

Caputo

et al. 2022

Sensors

Self Cite

View full text Add to dashboard Cite

In the current Information Age, it is usual to access our personal and professional information, such as bank account data or private documents, in a telematic manner. To ensure the privacy of this information, user authentication systems should be accurately developed. In this work, we focus on biometric authentication, as it depends on the user’s inherent characteristics and, therefore, offers personalized authentication systems. Specifically, we propose an electrocardiogram (EEG)-based user authentication system by employing One-Class and Multi-Class Machine Learning classifiers. In this sense, the main novelty of this article is the introduction of Isolation Forest and Local Outlier Factor classifiers as new tools for user authentication and the investigation of their suitability with EEG data. Additionally, we identify the EEG channels and brainwaves with greater contribution to the authentication and compare them with the traditional dimensionality reduction techniques, Principal Component Analysis, and χ2 statistical test. In our final proposal, we elaborate on a hybrid system resistant to random forgery attacks using an Isolation Forest and a Random Forest classifiers, obtaining a final accuracy of 82.3%, a precision of 91.1% and a recall of 75.3%.

show abstract

Section: Introductionmentioning

confidence: 99%

EEG Authentication System Based on One- and Multi-Class Machine Learning Classifiers

Hernández–Álvarez

Barbierato

Caputo

et al. 2022

Sensors

Self Cite

View full text Add to dashboard Cite

show abstract

“…Although the current methods based on traditional machine learning or deep learning have made some exciting progress, they still suffer some limitations. First, sensory data from impostors (negative samples) are needed to train the continuous authentication model (binary classification or multi-classification) [1,6,11,12,15,21,39,47], since the distribution of negative training data from diverse attackers are unknown, it is a difficult problem to solve in a real-world scenario. Besides, sharing of other smartphone users' biometric data (negative samples) may lead to the leakage of biometric data.…”

Section: Introductionmentioning

confidence: 99%

Relative Attention-based One-Class Adversarial Autoencoder for Continuous Authentication of Smartphone Users

Hu¹,

Zhang²,

You³

et al. 2022

Preprint

View full text Add to dashboard Cite

Behavioral biometrics-based continuous authentication is a promising authentication scheme, which uses behavioral biometrics recorded by built-in sensors to authenticate smartphone users throughout the session. However, current continuous authentication methods suffer some limitations: 1) behavioral biometrics from impostors are needed to train continuous authentication models. Since the distribution of negative samples from diverse attackers are unknown, it is a difficult problem to solve in real-world scenarios; 2) most deep learningbased continuous authentication methods need to train two models to improve authentication performance. A deep learning model for deep feature extraction, and a machine learning-based classifier for classification; 3) weak capability of capturing users' behavioral patterns leads to poor authentication performance. To solve these issues, we propose a relative attention-based one-class adversarial autoencoder for continuous authentication of smartphone users. First, we propose a one-class adversarial autoencoder to learn latent representations of legitimate users' behavioral patterns, which is trained only with legitimate smartphone users' behavioral biometrics. Second, we present the relative attention layer to capture richer contextual semantic representation of users' behavioral patterns, which modifies the standard self-attention mechanism using convolution projection instead of linear projection to perform the attention maps. Experimental results demonstrate that we can achieve superior performance of 1.05% EER, 1.09% EER, and 1.08% EER with a high authentication frequency (0.7s) on three public datasets.CCS Concepts: • Security and privacy → Biometrics; • Human-centered computing → Collaborative and social computing;

show abstract

“…These channels have been demonstrated in several studies to be vulnerable and leak the information of the patients [37,38]. Especially harmful is the scenario in which the transmitted data are not encrypted and can be directly interpreted [39].…”

mentioning

confidence: 99%

“…This scenario can be conducted by substituting the original app by a malicious one, specifically designed by the adversary to perform concrete actions. Likewise, leaks of sensors readings can be used to infer the activity the patient was executing at a certain time [39].…”

mentioning

confidence: 99%

See 1 more Smart Citation

Security Threats and Cryptographic Protocols for Medical Wearables

et al. 2022

Self Cite

View full text Add to dashboard Cite

In the past few years, the use of several medical devices is increasing. This paper will pay attention to a device developed to get measures of the temperature of diabetic foot. These wearables usually do not have cryptographic protocols to guarantee data security. This study analyzes the existing security in these devices, and simulate malware propagation taking into account the vulnerabilities and lack of security in these highly-constrained interconnected devices. A simulation of malware spreading in a network made by 10 and 15 individuals with 6 and 34 sensors each one, respectively, is included in this study. To avoid such attacks, a lightweight cryptographic protocol could be a satisfactory solution. Considering the quick development of quantum computers, several current cryptographic protocols have been compromised.

show abstract

SmartCAMPP - Smartphone-based continuous authentication leveraging motion sensors with privacy preservation

Cited by 19 publications

References 20 publications

EEG Authentication System Based on One- and Multi-Class Machine Learning Classifiers

EEG Authentication System Based on One- and Multi-Class Machine Learning Classifiers

Relative Attention-based One-Class Adversarial Autoencoder for Continuous Authentication of Smartphone Users

Security Threats and Cryptographic Protocols for Medical Wearables

Contact Info

Product

Resources

About