University of Twente Student Theses


Morphing Attack Detection (MAD) based on the location of the corneal specular highlights

Streapco, Lia (2023) Morphing Attack Detection (MAD) based on the location of the corneal specular highlights.

[img] PDF
Abstract:Generative Adversarial Networks (GAN) are tools that allow for the creation of realistic synthetic images. Their development in the past years has seen significant growth. New methodologies are emerging, including the one examined in this research, face morphing. However, the spread of GAN- generated morphed images has also made the world vulnerable, due to their realism, it is a challenge to distinguish them from the real ones. This research paper proposes a way of detecting GAN-generated morph of portrait pictures by utilizing the specular corneal highlights. The brightest point from the highlight is extracted, and this data is then processed through a logistic regression algorithm to determine the authenticity of the portrait picture. The research investigates the effectiveness of the proposed approach by utilizing two different methods of iris extraction, thus creating two different datasets. The results demonstrate promising outcomes and offer a path for further development.
Item Type:Essay (Bachelor)
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:54 computer science
Programme:Business & IT BSc (56066)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page