💃 IDOL: Unified Dual-Modal Latent Diffusion for Human-Centric Joint Video-Depth Generation

Yuanhao Zhai¹, Kevin Lin², Linjie Li², Chung-Ching Lin², Jianfeng Wang², Zhengyuan Yang², David Doermann¹, Junsong Yuan¹, Zicheng Liu³, Lijuan Wang²

¹State University of New Yort at Buffalo   |   ²Microsoft  |   ³Advanced Micro Devices

European Conference on Computer Vision (ECCV) 2024

 

TL;DR: Our IDOL enables human-centric joint video-depth generation, which could be rendered into realistic 2.5 videos.

All code and checkpoints will be released soon!