University of Twente Student Theses


Differentiating between facial landmark shifts due to pose variation and morphing

Kesteren, L. van (2022) Differentiating between facial landmark shifts due to pose variation and morphing.

[img] PDF
Abstract:Face recognition systems (FRSs) provide accurate biometric authentication, although some security risks have been discovered. In most cases, the FRSs are vulnerable to morphing attacks. In a morphing attack, the pictures of two subjects are morphed into one, resulting in an artificial image of a face that resembles both subjects. Multiple techniques have been developed to detect morphing attacks, however none of the existing algorithms have the necessary accuracy to prevent them. One technique consists of comparing the facial landmarks of the possibly morphed image with the bona fide image. Shifting of these landmarks could indicate the image is morphed. However, the shifts can also be caused by pose variation. This research aims to improve the performance of the aforementioned technique by differentiating landmark shifts due to pose variation and due to morphing to a greater extent. By including pose variation in the data performance can be increased significantly.
Item Type:Essay (Bachelor)
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:54 computer science
Programme:Computer Science BSc (56964)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page