University of Twente Student Theses
Spot the difference : Using Procrustes analysis on semantic features to enhance the performance of Facial Recognition Systems
Santen, A. van (2022) Spot the difference : Using Procrustes analysis on semantic features to enhance the performance of Facial Recognition Systems.
|
PDF
6MB |
| Abstract: | Difficult cases within facial recognition include so-called doppelg ̈angers or lookalikes. Earlier research shows that these cases are difficult for both open source and commercial software. This research explores the use of the Procrustes analysis on different semantic features to increase the performance of the commercial software FaceVACS. The Hochschule Darmstadt-doppelganger (HDA-database) and a number of collected images are used. Two experiments are performed using this database. Experiment A looks at the most prominent semantic features using feature compositions described by the Facial Identification Scientific Working Group (FISWG). This showed that hte nose, eyes, and eyebrows are the most prominent semantic features. Experiment B calculates the Procrustes distance between these features, and will try to improve performance of FaceVACS using both score level fusion and decision level fusion. However, neither of these methods yield better results. Investigation of different individuals does show the application of this research by improving the individual results, although this was not significant either. |
| Item Type: | Essay (Bachelor) |
| Faculty: | EEMCS: Electrical Engineering, Mathematics and Computer Science |
| Subject: | 54 computer science |
| Programme: | Technology and Liberal Arts & Sciences BSc (50427) |
| Link to this item: | https://purl.utwente.nl/essays/93202 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page