Deepfake detection is of fundamental importance to preserve the reliability of multimedia communications. Modern deepfake detection systems are often specialized on one or more types of manipulation but are not able to generalize. On the other hand, when properly designed, ensemble learning and fusion techniques can reduce this issue. In this paper, we exploit the complementarity of different individual classifiers and evaluate which fusion rules are best suited to increase the generalization capacity of modern deepfake detection systems. We also give some insights to designers for selecting the most appropriate approach.

Analysis of Score-Level Fusion Rules for Deepfake Detection

Concas, S
Co-primo
;
La Cava, SM
Co-primo
;
Orru', G;Marcialis, GL
;
Roli, F
2022-01-01

Abstract

Deepfake detection is of fundamental importance to preserve the reliability of multimedia communications. Modern deepfake detection systems are often specialized on one or more types of manipulation but are not able to generalize. On the other hand, when properly designed, ensemble learning and fusion techniques can reduce this issue. In this paper, we exploit the complementarity of different individual classifiers and evaluate which fusion rules are best suited to increase the generalization capacity of modern deepfake detection systems. We also give some insights to designers for selecting the most appropriate approach.
2022
deepfake; detection; fusion
File in questo prodotto:
File Dimensione Formato  
Analysis of Score-Level Fusion Rules for Deepfake Detection_compressed.pdf

accesso aperto

Descrizione: articolo online
Tipologia: versione editoriale (VoR)
Dimensione 585.93 kB
Formato Adobe PDF
585.93 kB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11584/343797
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 8
  • ???jsp.display-item.citation.isi??? 5
social impact