Authors: Georgii Stanishevskii, Jakub Steczkiewicz, Tomasz Szczepanik, Sławomir Tadeja, Jacek Tabor, Przemysław Spurek

Published on: February 09, 2024

Impact Score: 8.22

Arxiv code: Arxiv:2402.06390

Summary

• What is new: Combining deepfake algorithms with Neural Radiance Fields (NeRFs) and Gaussian Splatting (GS) to create 3D avatars.
• Why this is important: Traditional methods for creating 3D avatars are either not realistic enough or require significant resources for training.
• What the research proposes: ImplicitDeepfake1 uses deepfake algorithms to modify training images, then applies NeRF and GS to these modified images to create realistic 3D avatars.
• Results: Plausible 3D deepfake-based avatars that maintain a high level of realism and detail.

Technical Details

Technological frameworks used: Neural Radiance Fields (NeRFs), Gaussian Splatting (GS)

Models used: Deepfake algorithm for image modification

Data used: Images with known camera positions for NeRF; modified images for GS.

Potential Impact

Gaming, virtual reality, avatar creation platforms, and social media could greatly benefit from these insights, while traditional avatar creation tools might face disruption.

Want to implement this idea in a business?

We have generated a startup concept here: FaceFusion.