| Popis: |
This paper explores the usability of the Ripley’s K function and the nearest neighbor distance, in the detection of clustered graphical passwords in the graphical authentication stage. For it, both tests were applied to two bases of data of 10, 000 clustered graphical passwords each, the first with graphical passwords clustered in an area lesser than the fourth part of the original image, the second clustered in an area lesser than eighth part of the image. The results show that none of these tests is effective in the detection of clustered graphical passwords, the reason of such failure is due to the short size of the spatial pattern in question, only the 5 points of the graphical password analyzed. |
