University of Twente Student Theses
Exploring the GANformer for Face Generation : investigating the segmentation and smile augmentation potential
Ferla, R.I. (2022) Exploring the GANformer for Face Generation : investigating the segmentation and smile augmentation potential.
|
PDF
97MB |
| Abstract: | Advancing the research in face applications is limited by proprietary databases and increasing data protection regulations, synthetically generated databases may provide a solution. In this work the GANformer, a hybrid generative image model, is explored for this application. While only trained for unconditioned face generation like many other models, this works shows the potential of two use cases. First, the unique implementation of the attention is examined for the application of segmentation. Results indicate segmenting behaviour is present, though post-processing is needed before its implementation in synthetic databases. Second, real labeled faces are reconstructed in latent space to find latent directions describing disentangled attributes. This concept is brought in practice by augmenting neutral to smiling faces, but could be applied on other expressions and attributes as well. In both the segmentation and the smile augmentation the results indicate that the GANformer is able to be used for multiple applications in synthetic database generation. This work can be use as basis as it opens up two directions for further research. |
| Item Type: | Essay (Master) |
| Faculty: | EEMCS: Electrical Engineering, Mathematics and Computer Science |
| Subject: | 53 electrotechnology, 54 computer science |
| Programme: | Electrical Engineering MSc (60353) |
| Link to this item: | https://purl.utwente.nl/essays/90496 |
| 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