Proceedings of the 3rd International Conference on Public Management and Big Data Analysis, PMBDA 2023, December 15–17, 2023, Nanjing, China

Research Article

Multi-Modal Solution: Deepfake Detection and the Source Identification

Download131 downloads
Cite
BibTeX Plain Text
  • @INPROCEEDINGS{10.4108/eai.15-12-2023.2345299,
    author={Yahan  Zheng and Xu  Zhou and Cheng  Chen and Jingwen  Hu},
    title={Multi-Modal Solution: Deepfake Detection and the Source Identification},
    proceedings={Proceedings of the 3rd International Conference on Public Management and Big Data Analysis, PMBDA 2023, December 15--17, 2023, Nanjing, China},
    publisher={EAI},
    proceedings_a={PMBDA},
    year={2024},
    month={5},
    keywords={deepfake detection; multi-modal; mmmu-ba},
    doi={10.4108/eai.15-12-2023.2345299}
}
  • Yahan Zheng
    Xu Zhou
    Cheng Chen
    Jingwen Hu
    Year: 2024
    Multi-Modal Solution: Deepfake Detection and the Source Identification
    PMBDA
    EAI
    DOI: 10.4108/eai.15-12-2023.2345299
Yahan Zheng1,*, Xu Zhou2, Cheng Chen3, Jingwen Hu4
  • 1: Sichuan University
  • 2: Zhejiang University of Technology
  • 3: Donghua University
  • 4: KangChiao International School
*Contact email: 2020141440150@stu.scu.edu.cn

Abstract

Deepfake technology has recently raised significant concerns due to its potential for manipulating and misusing multimedia content. In response to this issue, researchers have been exploring novel approaches for deepfake detection. In this study, we propose a multimodal analysis framework that combines visual, audio, and textual modalities to determine if an unknown video is a fake one and to identify the source identity in manipulated media. By leveraging the complementary information from multiple modalities, our approach aims to enhance the accuracy and robustness of deepfake detection.

Keywords
deepfake detection; multi-modal; mmmu-ba
Published
2024-05-13
Publisher
EAI
http://dx.doi.org/10.4108/eai.15-12-2023.2345299
Copyright © 2023–2024 EAI