User Authentication Method Based on Keystroke Dynamics and Mouse Dynamics with Scene-Irrelated Features in Hybrid Scenes
Abstract
:1. Introduction
- A user authentication method is proposed to filter scene-irrelated features in hybrid scenes to reduce the impact of scenes on user authentication;
- A user authentication method is proposed to fuse scene-irrelated features with user-related features in hybrid scenes.
2. Related Research
2.1. Feature Selection
2.1.1. Filter Methods
2.1.2. Wrapper Methods
2.1.3. Embedded Methods
2.2. Keystroke Dynamics and Mouse Dynamics
2.2.1. Keystroke Dynamics
2.2.2. Mouse Dynamics
2.2.3. Fusion of Keystroke Dynamics and Mouse Dynamics
2.3. Multiple Kernel Learning
3. Proposed Approach
- First, features are extracted from the collected user operation data (details will be provided in Section 3.1);
- Second, scene-irrelated features and user-related features are selected from the original features (details will be elaborated in Section 3.2);
- Finally, scene-irrelated features and user-related features are fused and the model is trained (details will be presented in Section 3.3).
3.1. Feature Extraction
3.1.1. Keystroke Features
- Single key features: keystroke duration (KD). Each key pair has two single key features:
- the keystroke duration of the first key (KD1);
- the keystroke duration of the second key (KD2).
- Key pair features: the diagram latency between two keys. Each key pair has six key pair features:
- down-down diagram latency (DDDL);
- down-up diagram latency (DUDL);
- up-down diagram latency (UDDL);
- up-up diagram latency (UUDL).
3.1.2. Mouse Features
- Movement speed compared with traveled distance (MSD): The average speed of all moves within different moving distances. The range of moving distance is every interval of 100 pixels within 1–800 pixels (such as 1–100 pixels, 101–200 pixels). The length of the MSD feature vector is 8;
- Average movement speed per movement direction (MDA): The average speed of all movements in different moving directions. The moving direction is divided into eight equal parts. The length of the MDA feature vector is 8;
- Average speed for different types of actions (ATA): The average mouse movement speed of different operation types. There are three operations, namely, mouse-move, drag-drop, and point-click. The length of the ATA feature vector is 3;
- Traveled distance histogram (TDH): The ratio of the number of mouse operations in different moving distances to the total number of mouse operations. The length of the TDH feature vector is 8;
- Movement direction histogram (MDH): The ratio of the number of mouse operations in different moving directions to the total number of mouse operations. The length of the MDH feature vector is 8;
- Actions types histogram (ATH): The ratio of the number of mouse operations of different operation types to the total number of mouse operations. The length of the ATH feature vector is 3;
- Movement elapsed time histogram (MTH): The ratio of the number of mouse operations in different operation durations to the total number of mouse operations. The time range is 5 time periods separated by 200 milliseconds within 1–1000 milliseconds. The length of the MTH feature vector is 5.
3.2. Feature Processing
3.2.1. Scene-Irrelated Features and User-Related Features
3.2.2. Feature Selection Method
3.3. Feature Fusion Based on MKL
4. Experiments and Result Analysis
4.1. Experiment
4.1.1. Data Collection
4.1.2. Process Description
- If user is marked as a legal user, the remaining 40 users are marked as illegal users. The legal user data are taken as positive samples, and the illegal user data are taken as negative samples;
- In order to prevent data imbalance, negative samples are under-sampled, and the same number of negative samples as positive samples are randomly obtained;
- After combining the positive and negative samples, the scene-irrelated features and user-related features are selected and fused (according to the method described in Section 3.2) to obtain the user-scene features used for classification;
- The classification models based on a support vector machine and MKL are trained and tested by 5-fold cross-validation using the obtained features.
4.1.3. Evaluation
4.2. Analysis of Experimental Results
4.2.1. Illustrate the Reduction in User Authentication Accuracy of Hybrid Scenes
4.2.2. Determine Suitable Feature Combination
4.2.3. Determining Suitable Time window
4.2.4. Determine Appropriate Kernel
- Linear plus Linear (expressed as linear_linear) use linear kernels to map user-related features and linear kernels to map scene-irrelated features;
- Linear plus RBF (expressed as linear_rbf) use linear kernels to map user-related features and RBF kernels to map scene-irrelated features;
- RBF plus Linear (expressed as rbf_linear) use RBF kernels to map user-related features and Linear kernels to map scene-irrelated features;
- RBF plus RBF (expressed as rbf_rbf) use RBF kernels to map user-related features and RBF kernels to map scene-irrelated features.
4.2.5. Verify Feasibility of SIURUA
4.2.6. Determine Fill Values for Empty Features
4.2.7. Proven Advantages of SIURUA
- MPCA [43]: Piantari et al. proposed a mouse dynamics user authentication method based on Principal Components Analysis (PCA) and SVM;
- UIKDMM [44]: Panasiuk et al. proposed a multimodal biometric user authentication system based on keystroke dynamics and mouse movements, which authenticates users through K-Nearest Neighbor (KNN) and by fusing keystroke dynamics and mouse dynamics;
- UAMKL [45]: Wang et al. proposed UAMKL, which is an AverageMKL-based keystroke dynamics and mouse dynamics fusion user authentication method;
- TEM [46]: Chen et al. proposed a multimodal biometric user authentication system based on keystroke dynamics and mouse dynamics with Context Information. The user authentication model in the system is a comparison, which fuses the SVM based on keystroke dynamics features and the NB based on mouse dynamics features by using the majority voting mechanism.
4.2.8. Comprehensive Comparison of Above Experiments
5. Conclusions and Future work
5.1. Summaries and Discussion
5.2. Future Work
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Data Availability Statement
Conflicts of Interest
References
- Gaines, R.S.; Lisowski, W.; Press, S.J.; Shapiro, N. Authentication by Keystroke Timing: Some Preliminary Results; RAND Corporation: Santa Monica, CA, USA, 1980. [Google Scholar]
- Ahmed, A.A.E.; Traore, I. A New Biometric Technology Based on Mouse Dynamics. IEEE Trans. Dependable Secur. Comput. 2007, 4, 165–179. [Google Scholar] [CrossRef]
- Ahmed, A.A.E.; Traore, I. Anomaly Intrusion Detection Based on Biometrics. In Proceedings of the Sixth Annual IEEE Systems, Man and Cybernetics (SMC) Information Assurance Workshop, West Point, NY, USA, 15–17 June 2005; IEEE: West Point, NY, USA, 2005; pp. 452–453. [Google Scholar]
- Biometric personal authentication using keystroke dynamics: A review. Appl. Soft Comput. 2011, 11, 1565–1573. [CrossRef]
- Peng, J.; Choo, K.-K.R.; Ashman, H. User Profiling in Intrusion Detection: A Review. J. Netw. Comput. Appl. 2016, 72, 14–27. [Google Scholar] [CrossRef]
- Cai, Z.; Shen, C.; Guan, X. Mitigating Behavioral Variability for Mouse Dynamics: A Dimensionality-Reduction-Based Approach. IEEE Trans. Hum.-Mach. Syst. 2014, 44, 244–255. [Google Scholar] [CrossRef]
- Quafafou, M.; Boussouf, M. Induction of Strong Feature Subsets. In Principles of Data Mining and Knowledge Discovery; Komorowski, J., Zytkow, J., Eds.; Lecture Notes in Computer Science; Springer: Berlin/Heidelberg, Germany, 1997; Volume 1263, pp. 384–392. ISBN 978-3-540-63223-8. [Google Scholar]
- Peng, H.; Long, F.; Ding, C. Feature Selection Based on Mutual Information Criteria of Max-Dependency, Max-Relevance, and Min-Redundancy. IEEE Trans. Pattern Anal. Mach. Intell. 2005, 27, 1226–1238. [Google Scholar] [CrossRef]
- Grefenstette, J.J. Genetic Algorithms and Machine Learning. In Proceedings of the Sixth Annual Conference on Computational Learning Theory, New York, NY, USA, 26–28 July 1993; pp. 3–4. [Google Scholar] [CrossRef]
- Kohavi, R.; John, G.H. Wrappers for Feature Subset Selection. Artif. Intell. 1997, 97, 273–324. [Google Scholar] [CrossRef]
- Yang, Y.; Ma, Z.; Hauptmann, A.G.; Sebe, N. Feature Selection for Multimedia Analysis by Sharing Information Among Multiple Tasks. IEEE Trans. Multimedia 2013, 15, 661–669. [Google Scholar] [CrossRef]
- Tibshirani, R. Regression Shrinkage and Selection Via the Lasso. J. R. Stat. Soc. Ser. B (Methodol.) 1996, 58, 267–288. [Google Scholar] [CrossRef]
- Hoerl, E.; Kennard, W. Ridge Regression: Biased Estimation for Nonorthogonal Problems. Technometrics 2000, 42, 7. [Google Scholar] [CrossRef]
- Nie, F.; Cai, X.; Huang, H.; Ding, C. Efficient and Robust Feature Selection via Joint 2,1-Norms Minimization. In Proceedings of the 23rd International Conference on Neural Information Processing Systems, Vancouver British Columbia, Canada, 6–9 December 2010; pp. 1813–1821. [Google Scholar]
- Pisani, P.H.; Lorena, A.C. A Systematic Review on Keystroke Dynamics. J. Braz. Comput. Soc. 2013, 19, 573–587. [Google Scholar] [CrossRef]
- Azevedo, G.L.F.B.G.; Cavalcanti, G.D.C.; Carvalho Filho, E.C.B. An Approach to Feature Selection for Keystroke Dynamics Systems Based on PSO and Feature Weighting. In Proceedings of the 2007 IEEE Congress on Evolutionary Computation, Singapore, 25–28 September 2007; pp. 3577–3584. [Google Scholar]
- Alsultan, A.; Warwick, K.; Wei, H. Improving the Performance of Free-Text Keystroke Dynamics Authentication by Fusion. Appl. Soft Comput. 2018, 70, 1024–1033. [Google Scholar] [CrossRef]
- Epp, C.; Lippold, M.; Mandryk, R.L. Identifying Emotional States Using Keystroke Dynamics. In Proceedings of the Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, Vancouver, BC, Canada, 7–12 May 2011; ACM: Vancouver, BC, Canada; pp. 715–724. [Google Scholar]
- Antal, M.; Szabó, L.Z.; László, I. Keystroke Dynamics on Android Platform. Procedia Technol. 2015, 19, 820–826. [Google Scholar] [CrossRef]
- Tsai, C.-J.; Huang, P.-H. Keyword-Based Approach for Recognizing Fraudulent Messages by Keystroke Dynamics. Pattern Recognit. 2020, 98, 107067. [Google Scholar] [CrossRef]
- Ayotte, B.; Banavar, M.; Hou, D.; Schuckers, S. Fast Free-Text Authentication via Instance-Based Keystroke Dynamics. IEEE Trans. Biom. Behav. Identity Sci. 2020, 2, 377–387. [Google Scholar] [CrossRef]
- Tewari, A.; Verma, P. An Improved User Identification Based on Keystroke-Dynamics and Transfer Learning. WEB 2022, 19, 5369–5387. [Google Scholar] [CrossRef]
- Lu, X.; Zhang, S.; Hui, P.; Lio, P. Continuous Authentication by Free-Text Keystroke Based on CNN and RNN. Comput. Secur. 2020, 96, 101861. [Google Scholar] [CrossRef]
- Higashino, J. Signature Verification System on Neuro-Computer. In Proceedings of the 11th IAPR International Conference on Pattern Recognition. Vol. IV. Conference D: Architectures for Vision and Pattern Recognition, The Hague, Netherlands, 30 August–1 September 1992; pp. 517–521. [Google Scholar]
- Everitt, R.A.J.; McOwan, P.W. Java-Based Internet Biometric Authentication System. IEEE Trans. Pattern Anal. Mach. Intell. 2003, 25, 1166–1172. [Google Scholar] [CrossRef]
- Feher, C.; Elovici, Y.; Moskovitch, R.; Rokach, L.; Schclar, A. User Identity Verification via Mouse Dynamics. Inf. Sci. 2012, 201, 19–36. [Google Scholar] [CrossRef]
- Kasprowski, P.; Harezlak, K. Fusion of Eye Movement and Mouse Dynamics for Reliable Behavioral Biometrics. Pattern Anal. Appl. 2018, 21, 91–103. [Google Scholar] [CrossRef] [Green Version]
- Gao, L.; Lian, Y.; Yang, H.; Xin, R.; Yu, Z.; Chen, W.; Liu, W.; Zhang, Y.; Zhu, Y.; Xu, S.; et al. Continuous Authentication of Mouse Dynamics Based on Decision Level Fusion. In Proceedings of the 2020 International Wireless Communications and Mobile Computing (IWCMC), Limassol, Cyprus, 15–19 June 2020; pp. 210–214. [Google Scholar]
- Antal, M.; Fejer, N.; Buza, K. SapiMouse: Mouse Dynamics-Based User Authentication Using Deep Feature Learning. In Proceedings of the 2021 IEEE 15th International Symposium on Applied Computational Intelligence and Informatics (SACI), Timisoara, Romania, 19–21 May 2021; pp. 61–66. [Google Scholar]
- Hu, T.; Niu, W.; Zhang, X.; Liu, X.; Lu, J.; Liu, Y. An Insider Threat Detection Approach Based on Mouse Dynamics and Deep Learning. Secur. Commun. Netw. 2019, 2019, 1–12. [Google Scholar] [CrossRef]
- Bailey, K.O.; Okolica, J.S.; Peterson, G.L. User Identification and Authentication Using Multi-Modal Behavioral Biometrics. Comput. Secur. 2014, 43, 77–89. [Google Scholar] [CrossRef]
- Mondal, S.; Bours, P. A Study on Continuous Authentication Using a Combination of Keystroke and Mouse Biometrics. Neurocomputing 2016, 230, 1–22. [Google Scholar] [CrossRef]
- Earl, S.; Campbell, J.; Buckley, O. Identifying Soft Biometric Features from a Combination of Keystroke and Mouse Dynamics. In Proceedings of the Advances in Human Factors in Robots, Unmanned Systems and Cybersecurity, Washington DC, USA, 25–29 July 2021; Zallio, M., Raymundo Ibañez, C., Hernandez, J.H., Eds.; Springer International Publishing: Cham, Switzerland, 2021; pp. 184–190. [Google Scholar]
- Boser, B.E.; Guyon, I.M.; Vapnik, V.N. A Training Algorithm for Optimal Margin Classifiers. In Proceedings of the Fifth Annual Workshop on Computational Learning Theory-COLT ’92, Pittsburgh, PA, USA, 27–29 July 1992; ACM Press: Pittsburgh, PA, USA, 1992; pp. 144–152. [Google Scholar]
- Belanche, L.A.; Tosi, A. Averaging of Kernel Functions. Neurocomputing 2013, 112, 19–25. [Google Scholar] [CrossRef]
- Kloft, M.; Brefeld, U.; Sonnenburg, S.; Zien, A. Non-Sparse Regularization and Efficient Training with Multiple Kernels; EECS Department, University of California: Berkeley, CA, USA, 2010. [Google Scholar]
- Xu, Z.; Jin, R.; Yang, H.; King, I.; Lyu, M.R. Simple and Efficient Multiple Kernel Learning by Group Lasso. In Proceedings of the 27th International Conference on International Conference on Machine Learning, Haifa, Israel, 21–24 June 2010; pp. 1175–1182. [Google Scholar]
- Do, H.; Kalousis, A.; Woznica, A.; Hilario, M. Margin and Radius Based Multiple Kernel Learning. In Machine Learning and Knowledge Discovery in Databases; Buntine, W., Grobelnik, M., Mladenić, D., Shawe-Taylor, J., Eds.; Lecture Notes in Computer Science; Springer: Berlin/Heidelberg, Germany, 2009; Volume 5781, pp. 330–343. ISBN 978-3-642-04179-2. [Google Scholar]
- Aiolli, F.; Donini, M. EasyMKL: A Scalable Multiple Kernel Learning Algorithm. Neurocomputing 2015, 169, 215–224. [Google Scholar] [CrossRef]
- Teh, P.S.; Teoh, A.B.J.; Ong, T.S.; Neo, H.F. Statistical Fusion Approach on Keystroke Dynamics. In Proceedings of the 2007 Third International IEEE Conference on Signal-Image Technologies and Internet-Based System, Shanghai, China, 16–18 December 2022; IEEE: Shanghai, China, 2007; pp. 918–923. [Google Scholar]
- Tsimperidis, I.; Yucel, C.; Katos, V. Age and Gender as Cyber Attribution Features in Keystroke Dynamic-Based User Classification Processes. Electronics 2021, 10, 835. [Google Scholar] [CrossRef]
- Vinh, N.X.; Epps, J.; Bailey, J. Information Theoretic Measures for Clusterings Comparison: Is a Correction for Chance Necessary? In Proceedings of the 26th Annual International Conference on Machine Learning-ICML ’09, Montreal, QC, Canada, 14–18 June 2009; ACM Press: Montreal, QC, Canada, 2009; pp. 1–8. [Google Scholar]
- Munir; Piantari, E.; Firman, F.N. Dimensional Reduction in Behavioral Biometrics Authentication System. In Proceedings of the ICETIT 2019, Delhi, India, 21–22 June 2019; Singh, P.K., Panigrahi, B.K., Suryadevara, N.K., Sharma, S.K., Singh, A.P., Eds.; Lecture Notes in Electrical Engineering. Springer International Publishing: Cham, Switzerland, 2020; Volume 605, pp. 984–992, ISBN 978-3-030-30576-5. [Google Scholar]
- Panasiuk, P.; Szymkowski, M.; Dąbrowski, M.; Saeed, K. A Multimodal Biometric User Identification System Based on Keystroke Dynamics and Mouse Movements. In Computer Information Systems and Industrial Management; Saeed, K., Homenda, W., Eds.; Lecture Notes in Computer Science; Springer International Publishing: Cham, Switzerland, 2016; Volume 9842, pp. 672–681. ISBN 978-3-319-45377-4. [Google Scholar]
- Wang, X.; Zheng, Q.; Zheng, K.; Wu, T. User Authentication Method Based on MKL for Keystroke and Mouse Behavioral Feature Fusion. Secur. Commun. Netw. 2020, 2020, 1–14. [Google Scholar] [CrossRef]
- Chen, L.; Zhong, Y.; Ai, W.; Zhang, D. Continuous Authentication Based on User Interaction Behavior. In Proceedings of the 2019 7th International Symposium on Digital Forensics and Security (ISDFS), Barcelos, Portugal, 10–12 June 2019; IEEE: Barcelos, Portugal, 2019; pp. 1–6. [Google Scholar]
Feature | |||||
---|---|---|---|---|---|
Feature | |||||
---|---|---|---|---|---|
458 | |||||
3 |
Sums | |||||
---|---|---|---|---|---|
Sums |
Sums | |||||
---|---|---|---|---|---|
Sums |
Algorithm | SVM | DT | NB | LR | SIURUA |
---|---|---|---|---|---|
Building | |||||
Authentication |
Algorithm | aAcc | aPrec | aTPR | aFPR(aFAR) | aFRR |
---|---|---|---|---|---|
SVM | 0.752 | 0.754 | 0.794 | 0.290 | 0.206 |
DT | 0.712 | 0.714 | 0.716 | 0.291 | 0.284 |
NB | 0.598 | 0.632 | 0.561 | 0.352 | 0.439 |
LR | 0.758 | 0.742 | 0.790 | 0.282 | 0.210 |
SVM_400 | 0.766 | 0.762 | 0.798 | 0.266 | 0.202 |
MPCA [43] | 0.739 | 0.735 | 0.770 | 0.290 | 0.230 |
UIKDMM [44] | 0.674 | 0.674 | 0.705 | 0.335 | 0.295 |
UAMKL [45] | 0.773 | 0.779 | 0.788 | 0.243 | 0.212 |
TEM [46] | 0.722 | 0.713 | 0.782 | 0.333 | 0.217 |
SIURUA | 0.840 | 0.841 | 0.850 | 0.169 | 0.150 |
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |
© 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Wang, X.; Shi, Y.; Zheng, K.; Zhang, Y.; Hong, W.; Cao, S. User Authentication Method Based on Keystroke Dynamics and Mouse Dynamics with Scene-Irrelated Features in Hybrid Scenes. Sensors 2022, 22, 6627. https://doi.org/10.3390/s22176627
Wang X, Shi Y, Zheng K, Zhang Y, Hong W, Cao S. User Authentication Method Based on Keystroke Dynamics and Mouse Dynamics with Scene-Irrelated Features in Hybrid Scenes. Sensors. 2022; 22(17):6627. https://doi.org/10.3390/s22176627
Chicago/Turabian StyleWang, Xiujuan, Yutong Shi, Kangfeng Zheng, Yuyang Zhang, Weijie Hong, and Siwei Cao. 2022. "User Authentication Method Based on Keystroke Dynamics and Mouse Dynamics with Scene-Irrelated Features in Hybrid Scenes" Sensors 22, no. 17: 6627. https://doi.org/10.3390/s22176627