Face forgery in videos is an increasing threat due to an easy access to manipulation tools. This work propose a Deep Learning approach to generate a mask showing the region of a tampered face. A Conditional Generative Adversarial Network is used,a total of 16 features are extracted from the video and used as input of the network. This method is tested on videos from YouTube and a dataset from Google and Jigsaw. The results are extremely promising, although there is still room for improvements.
Face forgery detection using conditional GAN
Rossi, Gianmaria
2020/2021
Abstract
Face forgery in videos is an increasing threat due to an easy access to manipulation tools. This work propose a Deep Learning approach to generate a mask showing the region of a tampered face. A Conditional Generative Adversarial Network is used,a total of 16 features are extracted from the video and used as input of the network. This method is tested on videos from YouTube and a dataset from Google and Jigsaw. The results are extremely promising, although there is still room for improvements.File in questo prodotto:
| File | Dimensione | Formato | |
|---|---|---|---|
|
gianmaria_rossi_tesi.pdf
Open Access dal 21/12/2022
Dimensione
2.12 MB
Formato
Adobe PDF
|
2.12 MB | Adobe PDF | Visualizza/Apri |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
Utilizza questo identificativo per citare o creare un link a questo documento:
https://hdl.handle.net/20.500.12608/28846