University of Twente Student Theses

Login

Forensic Facial Recognition Utilizing Cartoonization Segmentation

Suidgeest, Wouter (2021) Forensic Facial Recognition Utilizing Cartoonization Segmentation.

[img] PDF
2MB
Abstract:The topic of implementing automated facial recognition software has been a challenge for decades. Multiple very well working systems have been developed over the past years and are already in use. However, most of these systems lack transparency, causing them to be unusable in the field of forensic science. To assist in this field, the Facial Identification Scientific Working Group (FISWG) has described and documented a set of facial features suitable for facial recognition in forensic science, and how they need to be treated. However, these specific features are currently not sufficiently detectable by automated systems, as some features need to be segmented very precisely. In this study, a novel face recognition technology is designed, which utilizes cartoonization technology in the segmentation process, as the simplified shapes in cartoonized images could make the segmentation of FISWG features more precise. While the method seems promising from visual inspections of the individual steps alone, due to a number of currently unresolved problems in certain types of images, the cartoonization process does not yet improve the reliability of the system.
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:http://purl.utwente.nl/essays/86850
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page