Avatar Generation
Avatar generation focuses on creating realistic and animatable 3D human (and sometimes animal) models from various input modalities, including text descriptions, images, and videos. Current research heavily utilizes diffusion models, often combined with neural radiance fields (NeRFs) or parametric body models like SMPL-X, to achieve high-fidelity results and control over attributes like pose, expression, and clothing. This field is significant due to its potential applications in diverse areas such as virtual and augmented reality, gaming, film, and robotics, driving advancements in both computer graphics and artificial intelligence.
Papers
MagicMirror: Fast and High-Quality Avatar Generation with a Constrained Search Space
Armand Comas-Massagué, Di Qiu, Menglei Chai, Marcel Bühler, Amit Raj, Ruiqi Gao, Qiangeng Xu, Mark Matthews, Paulo Gotardo, Octavia Camps, Sergio Orts-Escolano, Thabo Beeler
HAHA: Highly Articulated Gaussian Human Avatars with Textured Mesh Prior
David Svitov, Pietro Morerio, Lourdes Agapito, Alessio Del Bue