Our paper, Articulate3D, has been accepted for publication at ICCV 2025.

The work presents a method for recovering animatable 3D human models from monocular RGB video. Articulate3D introduces a canonical representation and a reconstruction framework that jointly capture detailed surface geometry and articulation across a wide range of motions and subjects.

Authors:
Anna-Maria Halacheva, Jan-Nico Zaech, Xi Wang, Danda Pani Paudel, Luc Van Gool

The project page, including code and supplementary materials, is available at:
https://insait-institute.github.io/articulate3d.github.io/