Research

VOODOO XP: Expressive One-Shot Head Reenactment for VR Telepresence

VOODOO XP: Expressive One-Shot Head Reenactment for VR Telepresence

SIGGRAPH Asia 2024

Introducing VOODOO XP: a 3D-aware one-shot head reenactment method that generates expressive facial animations from any input video and a single 2D portrait, requiring no calibration or fine-tuning. Our real-time solution ensures view consistency and identity preservation, demonstrated in monocular video and VR telepresence systems.

[Arxiv]

voodooxp

voodoovr

headreenactment

mbzuai

pinscreen

ethz

vinai

generativeai

nerf

deepfakes

3davatar

virtualreality

metaversecenter

mbzuai

VOODOO VR: One-Shot Neural Avatars for Virtual Reality

VOODOO VR: One-Shot Neural Avatars for Virtual Reality

SIGGRAPH RTL 2024

Introducing VOODOO VR: a two-way immersive telepresence solution based on the state-of-the-art one-shot facial reenactment technology.

[Project] [Arxiv]

voodoo3d

voodoovr

headreenactment

mbzuai

pinscreen

ethz

vinai

generativeai

nerf

deepfakes

3davatar

virtualreality

metaversecenter

mbzuai

VOODOO 3D: Volumetric Portrait Disentanglement for One-Shot 3D Head Reenactment

VOODOO 3D: Volumetric Portrait Disentanglement for One-Shot 3D Head Reenactment

CVPR 2024

A high-fidelity 3D-aware one-shot head reenactment technique. Our method transfers the expression of a driver to a source and produces view consistent renderings for holographic displays.

[Project] [Arxiv] [Github]

voodoo3d

headreenactment

mbzuai

pinscreen

ethz

vinai

generativeai

holographicdisplay

nerf

deepfakes

3davatar

virtualreality

metaversecenter

mbzuai

Blur2Blur: Blur Conversion for Unsupervised Image Deblurring on Unknown Domains

Blur2Blur: Blur Conversion for Unsupervised Image Deblurring on Unknown Domains

CVPR 2024

A farewell to the overfitting issue in image deblurring.

[Project] [Github]

deblurring

deeplearning

hypercut

vinai

mbzuai

HyperCUT: Video Sequence from a Single Blurry Image using Unsupervised Ordering

HyperCUT: Video Sequence from a Single Blurry Image using Unsupervised Ordering

CVPR 2023

A deep learning network that can recover a sequence of sharp images from a single blurry image

[Arxiv] [Github]

deblurring

deeplearning

hypercut

vinai

mbzuai

QC-StyleGAN – Quality Controllable Image Generation and Manipulation

QC-StyleGAN – Quality Controllable Image Generation and Manipulation

NeuRIPS 2022

A StyleGAN-based generative model that can produce pairs of clean and degraded images.

[Arxiv] [Github]

generativeai

stylegan

deblurring

denoising

superresolution

vinai

mbzuai

Exploring Image Deblurring via Encoded Blur Kernel Space

Exploring Image Deblurring via Encoded Blur Kernel Space

CVPR 2021

Introducing a method to encode blur kernel space, which can be used in many downstream applications such as image deblurring, data augmentation, etc.

[Arxiv] [Github]

deblurring

cvpr

vinai

mbzuai