MBZUAI Unveils AI-Powered Virtual Avatar Technology That Works in Seconds

Technology

MBZUAI Unveils AI-Powered Virtual Avatar Technology That Works in Seconds

Cutting-Edge Innovation Enhances Virtual Communication and Visual Effects

Researchers at the Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) have introduced groundbreaking technology that allows users to create realistic digital avatars in seconds using just a webcam. The innovations, known as Voodoo XP and XMem++, were demonstrated to Khaleej Times by Professor Hao Li and PhD student Ariana Bermudez, showcasing the future of virtual communication and digital interaction.

Transforming Virtual Avatars with Voodoo XP

Voodoo XP enables real-time facial reenactment with a single photo, eliminating the need for specialized equipment. Unlike Meta’s Codec Avatar system, which requires 171 cameras and hours of processing, Voodoo XP can generate an avatar instantly using only a webcam.

“What you’re seeing is me controlling the avatar in real time, with no special hardware — just an ordinary camera,” explained Professor Li, Director of MBZUAI’s Metaverse Centre.

Bermudez emphasized how accessible the technology is:

“Even small movements like blinking or smiling are captured, making it incredibly lifelike.”

The research team behind Voodoo XP includes Phong Tran (MBZUAI student), Egor Zakharov (ETH Zurich), Long-Nhat Ho (MBZUAI student), Anh Tuan Tran (VinAI Research), Liwen Hu (Pinscreen), and Professor Li.

XMem++: Revolutionizing Video Editing and VFX

Alongside Voodoo XP, MBZUAI introduced XMem++, an advanced video object segmentation tool that streamlines video editing, augmented reality, and autonomous systems.

Bermudez explained how XMem++ improves efficiency for VFX artists working in film production:

“This is something that is tedious for many VFX artists because they usually have to refine a lot of details manually. Our tool automates that process, making editing more efficient.”

XMem++ has been integrated into industry-standard compositing software like Nuke and is now widely used in Hollywood and content creation industries. The technology, developed with contributions from Maksym Bekuzarov (MBZUAI Alumni) and Joon-Young Lee (Adobe), is also open-source and freely available.

The Future of AI-Powered Virtual Interaction

MBZUAI’s latest innovations pave the way for seamless virtual interactions, enhanced digital storytelling, and efficient video production. With Voodoo XP and XMem++, real-time avatars and AI-powered video editing are set to redefine digital experiences worldwide.