MFMS: Learning Modality-Fused and Modality-Specific Features for Deepfake Detection and Localization Tasks
This paper presents a summary of the proposed solution to the AV-Deepfake1M competition. Deepfake technology is developing fast, and realistic generation techniques of audio and videos have aroused public concerns. With this background, the AV-Deepfake1M competition aims to address the problem of audio-video Deepfake and provides a large-scale dataset named AV-Deepfake1M to boost the research in this area. In this paper, we present our solutions which have achieved top performance in this competition. We also provide more detailed experiments to prove the effectiveness of the modules used in our methods.
Authors
- Bingyu Hu (ORCID: https://orcid.org/0009-0007-6482-2803)
- Qi Chu (ORCID: https://orcid.org/0000-0003-3028-0755)
- Zhe Li (ORCID: https://orcid.org/0009-0006-5881-5300)
- Weiwei Feng (ORCID: https://orcid.org/0000-0002-8761-0375)
- Changtao Miao (ORCID: https://orcid.org/0000-0002-7634-9992)
- Weibin Yao
- Man Luo (ORCID: https://orcid.org/0000-0002-5399-5545)
- Yi Zhang (ORCID: https://orcid.org/0000-0001-5463-9844)
- Jianshu Li (ORCID: https://orcid.org/0000-0001-8554-6886)
- Wenjia Deng
- Tao Gong (ORCID: https://orcid.org/0000-0003-0026-6813)
Institutions
- University of Science and Technology of China (CN)
- Beijing Jiaotong University (CN)
Publication Details
- Published
- 2024-10-26
- DOI
- https://doi.org/10.1145/3664647.3688984
- Citations
- 11
- Primary Topic
- Human Pose and Action Recognition
- Type
- article
- Field-Weighted Citation Impact
- 2.74