| Title |
Netikrų veidų generavimo tyrimas / |
| Translation of Title |
Analysis of deepfake face generation methods. |
| Authors |
Laurinavičius, Donatas |
| Full Text |
|
| Pages |
78 |
| Keywords [eng] |
facial beauty prediction ; generative adversarial networks ; deep learning |
| Abstract [eng] |
The aim of the Master‘s Thesis was to analyze the effectiveness of facial beauty prediction when using images that contain human faces. To predict facial beauty the analysis was conducted by comparing different methods. Images containing multiple faces were also analyzed. Having more than one face in an image raised problems like bad resolution, bad angle of the face and more. For each problem, a solution was proposed with an analysis of the possible outcome. To improve facial beauty predictions, it was proposed to use GAN to generate a copy of a face and then to predict facial beauty of the generated face. . |
| Dissertation Institution |
Kauno technologijos universitetas. |
| Type |
Master thesis |
| Language |
Lithuanian |
| Publication date |
2021 |
