Institutional Repository of Vilnius University
Slaptažodžiais grįsta autentifikacija naudojant biometriką /
Abstract
dc:descriptionPassword-based authentication remains one of the most widely used authentication methods, but security vulnerabilities in this method, such as weak passwords, careless user behavior, and attacks that exploit password weaknesses, indicate that it is not sufficiently secure to protect user data from unauthorized access. Multi-factor authentication is used to strengthen password-based authentication, but such solutions require additional user actions or specialized equipment, which can complicate the usual login process and negatively affect the user experience. Therefore, it is important to look for a solution that could not only strengthen authentication, but also change the user’s usual login process as little as possible, thus maintaining a balance of convenience and security. This paper examines biometric authentication methods, from which the keystroke dynamics method is selected as an additional layer of protection according to selected criteria. Then, the structure of the authentication model is designed and its main components are implemented. The designed model is built in accordance with the principles of classical biometric authentication: keyboard events are collected, features are formed from them using a digraph method, which are then transferred to the comparison component. The comparison component uses the one-class nearest neighbor method, and the decision-making component uses statistical thresholding methods - percentile, standard deviation, median absolute deviation and two-stage thresholding methodology ”2T”. The effectiveness of the model is experimentally evaluated using a publicly available dataset, additionally analyzing the influence of threshold parameter values and user template size on the final authentication results. The results showed that the designed model can be successfully integrated into the usual login process and operates in the internal part of the system, however, the dynamics of keystrokes based on statistical methods is not accurate enough to ensure reliable authentication results while maintaining a balance between security and convenience. Even in the case of the most balanced method, the error rates obtained are at a noticeable level, therefore the effectiveness of such an additional layer of protection could be further improved by searching for more accurate comparison or thresholding methods.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Lauruška, Augustas,
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD308337468&prefLang=en_US