Research Article
Analysis of the Number of Sides of Voronoi Polygons in PassPoint
@INPROCEEDINGS{10.1007/978-3-030-69839-3_13, author={Lisset Su\^{a}rez-Plasencia and Joaqu\^{\i}n A. Herrera-Mac\^{\i}as and Carlos M. Leg\^{o}n-P\^{e}rez and Raisa Socorro-LLanes and Omar Rojas and Guillermo Sosa-G\^{o}mez}, title={Analysis of the Number of Sides of Voronoi Polygons in PassPoint}, proceedings={Computer Science and Health Engineering in Health Services. 4th EAI International Conference, COMPSE 2020, Virtual Event, November 26, 2020, Proceedings}, proceedings_a={COMPSE}, year={2021}, month={7}, keywords={Passpoint Voronoi polygons Entropy}, doi={10.1007/978-3-030-69839-3_13} }
- Lisset Suárez-Plasencia
Joaquín A. Herrera-Macías
Carlos M. Legón-Pérez
Raisa Socorro-LLanes
Omar Rojas
Guillermo Sosa-Gómez
Year: 2021
Analysis of the Number of Sides of Voronoi Polygons in PassPoint
COMPSE
Springer
DOI: 10.1007/978-3-030-69839-3_13
Abstract
The probabilistic distribution of the characteristics of Voronoi polygons has been extensively studied due to its many areas of application. In various works that differ in the number of polygons generated and in the size of their regions, it is concluded that the expected value of the characteristic number of sides of Voronoi polygons is equal to 6. In this work, this characteristic in the polygons generated by the graphical passwords of the graphical authentication system is studied. Its distribution is estimated and it is shown that the expected value of the number of sides of the Voronoi polygons in this scenario differs from previous works. The effectiveness of this feature is evaluated to detect weak graphical passwords made up of grouped dots. They are to be detected by estimating the entropy of the number of sides and by the expected value of the number of sides. It is concluded that the distribution of the number of sides in this scenario does not the 3-parameter gamma distribution reported in previous work or any of 61 distributions that were tested, and that the entropy and the expected value of the number of sides are not efficient for the detection of weak graphical passwords of formed by 5 grouped points.