Global Multimedia Deepfake Detection Challenge 2024: Innovative Technological Advancements

The Global Multimedia Deepfake Detection Challenge 2024 concluded with 2,200 participants from 26 countries. The event, co-hosted by the INCLUSION Conference and Ant Group, aimed to advance deepfake detection technology. Participants showcased innovative models in two tracks: image and audio/video, achieving notable success in improving recognition accuracy.

The Global Multimedia Deepfake Detection Challenge 2024 has wrapped up, attracting over 2,200 competitors from 26 nations. Jointly hosted by the INCLUSION Conference and Ant Group, and organized by Ant Digital Technologies, the event highlighted the transformative potential of technology for societal benefit.

Kicking off in July 2024, the Challenge invited participants to create, test, and refine more precise detection models against diverse deepfake threats, fostering advanced recognition strategies. The competition featured an image track and an audio/video track, presenting multifaceted face forgery methods within its datasets.

A team from Hong Kong and Macau led by Dr. Wu Haiwei achieved a 97.038% accuracy rate in the Image Track, while Tang Yongwei excelled in the Audio/Video Track with state-of-the-art algorithms for detecting manipulated audio. The Challenge emphasized integrating industry, academia, and research, preparing practitioners for real-world applications.

On August 30th, the committee launched an open-source initiative, encouraging participants to share their work on GitHub. The initiative saw enthusiastic participation, with nearly 100 participants and teams keen to join. For more details on the INCLUSION Conference, visit inclusionconf.com.

(With inputs from agencies.)