Article
Computer Science, Information Systems
Abbas Acar, Hidayet Aksu, A. Selcuk Uluagac, Kemal Akkaya
Summary: Continuous authentication, such as Wearable-Assisted Continuous Authentication (WACA), improves the reliability of user identity verification throughout a session, overcoming limitations of conventional one-time login systems. WACA utilizes sensor-based keystroke dynamics to transparently and periodically compare authentication data with the initial user profile, ensuring the current user is the same as the initially logged-in user. Empirical evaluation shows WACA's feasibility, low error rate, minimal computational overhead, and high accuracy in identifying insider threats and resisting powerful adversaries. The usability and efficiency of WACA propose potentially transformative implications in the authentication field.
IEEE TRANSACTIONS ON MOBILE COMPUTING
(2021)
Article
Computer Science, Information Systems
Luis de-Marcos, Jose-Javier Martinez-Herraiz, Javier Junquera-Sanchez, Carlos Cilleruelo, Carmen Pages-Arevalo
Summary: Continuous authentication (CA) is a process of verifying user identity regularly without their active participation. By training machine learning classifiers with data like typing events, predictions of user identity can be returned with a small number of key events and measurements. Among the various classifiers tested, ensemble algorithms, particularly GBC, showed the best statistical results for the CA mobile keystroke classification problem.
Article
Chemistry, Multidisciplinary
Najwa Altwaijry
Summary: This paper investigates the influence of typing language on keystroke dynamics. The results show that keystroke dynamics is influenced by the language being used, and bilingual users need to have two profiles created.
APPLIED SCIENCES-BASEL
(2023)
Article
Multidisciplinary Sciences
Nahuel Gonzalez, Enrique P. Calot, Jorge S. Ierache, Waldo Hasperue
Summary: This study compared heavy-tailed and positively skewed candidate distributions for fitting timing histograms in keystroke dynamics profiles. Results were summarized in three ways: counting best fit frequency, measuring average information content, and ranking distributions by hypothesis rejection frequency. The log-logistic distribution emerged as the clear winner among all candidates, surpassing others under three evaluation criteria.
Article
Multidisciplinary Sciences
Suliman A. Alsuhibany, Afnan S. Almuqbil
Summary: This paper examines users' typing behavior on different tablet devices with varying touch keyboard sizes and empirically assesses the impact on free-text KD authentication. The results provide interesting research directions for studying the effect of dynamic keystroke typing-pad on user security and trust authentication analysis.
SCIENTIFIC REPORTS
(2022)
Article
Chemistry, Analytical
Xiujuan Wang, Yutong Shi, Kangfeng Zheng, Yuyang Zhang, Weijie Hong, Siwei Cao
Summary: In this article, a user authentication method called SIURUA is proposed, which utilizes scene-irrelated features and user-related features for user identification. By fusing these two types of features, the method improves the accuracy of user authentication.
Article
Computer Science, Information Systems
Yutong Shi, Xiujuan Wang, Kangfeng Zheng, Siwei Cao
Summary: This paper presents a novel method for user authentication using keystroke dynamics and mouse dynamics, which utilizes heterogeneous domain adaptation to handle complex real-environment HCI data. Experimental results demonstrate the effectiveness of the proposed method.
MULTIMEDIA SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Orcan Alpar
Summary: Investigating new intelligent solutions for user identification and authentication is vital for enhancing password security, with frequency-based solutions showing better retention of unique biometric characteristics without the risk of convergence.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Computer Science, Artificial Intelligence
Andres Mafla, Ruben Tito, Sounak Dey, Lluis Gomez, Marcal Rusinol, Ernest Valveny, Dimosthenis Karatzas
Summary: In this study, the task of scene text retrieval is addressed by proposing a single shot CNN architecture for predicting bounding boxes and building compact representations of spotted words. Experimental results demonstrate that the proposed model outperforms previous state-of-the-art while offering significant increase in processing speed and unmatched expressiveness.
PATTERN RECOGNITION
(2021)
Article
Computer Science, Information Systems
Vitalijus Gurcinas, Juozas Dautartas, Justinas Janulevicius, Nikolaj Goranin, Antanas Cenys
Summary: This paper investigates keystroke-injection attacks and their potential risks, proposing the use of deep learning to bypass security measures in keystroke-based authentication systems. The results show the effectiveness of this approach, but caution against deploying it in real-world scenarios due to high security risks associated with adaptive keystroke injection and impersonated payload.
Article
Engineering, Electrical & Electronic
Bilal Taha, Sherif Nagib Abbas Seha, Dae Yon Hwang, Dimitrios Hatzinakos
Summary: Eye movement is used as a new behavioral modality for continuous driver authentication. An end-to-end learning model based on LSTM and dense networks is proposed to generate embeddings using only eye movement data. The model demonstrates excellent authentication performance with short frame lengths (as low as 30 seconds) compared to existing methods.
IEEE JOURNAL OF SELECTED TOPICS IN SIGNAL PROCESSING
(2023)
Article
Computer Science, Information Systems
Soobin Lee, Seungjae Lee, Eunkyoung Park, Jongshill Lee, In Young Kim
Summary: With the rapid development of networking and computing technology, the security of personal information stored on smart devices has become increasingly important. This study proposes a sensor compensation algorithm and new 2D cyclogram features to improve user authentication performance in gait recognition. Experimental results show the proposed algorithm can transform unstable sensor signals into stable anatomical coordinates and effectively discriminate individual gait patterns using 2D cyclogram features, achieving high accuracy rates for gait authentication.
Review
Computer Science, Information Systems
Riseul Ryu, Soonja Yeom, Soo-Hyung Kim, David Herbert
Summary: Building safeguards against illegitimate access and authentication is crucial for system security. Continuous multimodal biometric authentication systems have been proposed as a reliable solution to address the challenges in existing user authentication schemes. However, there is a lack of critical analysis on current progress in the field, highlighting the need for further research and development in this area.
Article
Computer Science, Information Systems
Haitian Yang, Xiang Meng, Xuan Zhao, Yan Wang, Yuejun Liu, Xiaoyu Kang, Jiahui Shen, Weiqing Huang
Summary: With the rapid development of digitalization, information security becomes more and more important. User authentication is a significant line of defense. Recently, researchers have applied learning-based models to user authentication, and in this paper, the authors propose contents and keystroke dual attention networks with pre-trained models for continuous authentication.
COMPUTERS & SECURITY
(2023)
Article
Computer Science, Hardware & Architecture
Chenhao Lin, Jingyi He, Chao Shen, Qi Li, Qian Wang
Summary: Behavioral biometrics has been widely studied and applied in real-world scenarios for human authentication. However, there is almost no attempt to identify behavior patterns in cross-scenario settings, which are common in practice and require immediate attention. This paper defines and investigates cross-scenario behavioral biometrics authentication using keystroke dynamics. The proposed approach, CrossBehaAuth, utilizes a temporal-aware learning mechanism and local Gaussian data augmentation to improve authentication accuracy in cross-scenario settings and achieve comparable performance in single-scenario authentication tasks.
IEEE TRANSACTIONS ON DEPENDABLE AND SECURE COMPUTING
(2023)