Citation

Ng, Jin Yang and Chong, Siew Chin and Wee, Kuok Kwee (2024) Enhancing Deepfake Detection for Public Awareness, Law Enforcement and Legal Verification. In: 2024 International Conference on Information Technology Research and Innovation (ICITRI), 05-06 September 2024, Jakarta, Indonesia.

Text g.pdf - Published Version Restricted to Repository staff only Download (643kB)

Abstract

In recent years, as you can see the emergence of deepfake technology had raised worrying concerns in various sectors, including politics, the media industry and most importantly cybersecurity. As deep fakes they can manipulate audio and video content created using deep learning techniques. This research aims to investigate and discover the best methods to detect and uncover deep fake content which also addresses the needs for a reliable deepfake method. The study’s approach, which integrates multimodal data fusion and leverages NAS, achieved outstanding results, reaching 99.37% accuracy after 15 epochs. This performance surpasses benchmarks set by traditional CNNs and earlier GAN-based methods. Experiments were conducted using two datasets, FaceForensics++ and Celeb-DF (v2), further validating the robustness of the method. These findings underscore the effectiveness of automated deep learning in combating digital misinformation

Item Type:	Conference or Workshop Item (Paper)
Uncontrolled Keywords:	Deepfake, Neural Architecture Search, Machine Learning, Deep Learning, Cybersecurity
Subjects:	Q Science > QA Mathematics > QA71-90 Instruments and machines
Divisions:	Faculty of Information Science and Technology (FIST)
Depositing User:	Ms Nurul Iqtiani Ahmad
Date Deposited:	04 Nov 2024 01:50
Last Modified:	04 Nov 2024 01:50
URII:	http://shdl.mmu.edu.my/id/eprint/13101

Downloads

Edit (login required)