Elisabetta Fedele

I am a Computer Science PhD student at ETH Zürich, working on 3D Computer Vision and Machine Learning, under the supervision of Prof. Marc Pollefeys and Prof. Bernhard Schölkopf. I am also a Doctoral Fellow at the ETH AI Center.

My general goal is to ground World Models in compact 3D representations, making them more efficient, controllable, and interactive.

I am currently based in Menlo Park, where I am a Research Scientist Intern at Meta. Before that, I spent six months as a research intern in Leonidas Guibas' lab at Stanford, and in 2023 I was a research intern at EPFL, in Alessandro Chiesa's lab.

Email  /  Twitter  /  CV  /  Google Scholar  /  Linkedin  /  Github

profile photo
News

Selected Research

SuperFlex: Deformable Superquadrics for Point Cloud Decomposition
Gabriel Tavernini*, Elisabetta Fedele*, Tiago Novello, Leonidas Guibas, Marc Pollefeys, Francis Engelmann
ECCV 2026
project page / arXiv

SuperFlex improves over SuperDec in terms of reconstruction quality, primitive expressivity (introduces bending and tapering), and robustness to occluded point clouds.


SpaceControl: Introducing Test-Time Spatial Control to 3D Generative Modeling
Elisabetta Fedele*, Francis Engelmann*, Ian Huang, Or Litany, Marc Pollefeys, Leonidas Guibas
ICLR 2026
project page / paper

SpaceControl allows to control 3D generative models with explicit spatial guidance using arbitrary geometries.


SuperDec: 3D Scene Decomposition with Superquadric Primitives
Elisabetta Fedele, Boyang Sun, Leonidas Guibas, Marc Pollefeys, Francis Engelmann
ICCV 2025   (Oral Presentation)
project page / arXiv

SuperDec allows to represent arbitrary 3D scenes with a compact and modular set of superquadric primitives.


OpenMask3D: Open-Vocabulary 3D Instance Segmentation
Ayça Takmaz*, Elisabetta Fedele*, Robert W. Sumner, Marc Pollefeys, Federico Tombari,
Francis Engelmann
NeurIPS 2023
project page / arXiv / code

OpenMask3D is a zero-shot approach to perform 3D Instance Segmentation with Open-Vocabulary queries.

Others

FABRIC: Personalizing Diffusion Models with Iterative Feedback
Dimitri Von Rütte*, Elisabetta Fedele*, Jonathan Thomm*, Lukas Wolf*
ECCVW 2024
project page / arXiv / code

FABRIC is a training-free approach that conditions the diffusion process on a set of feedback images, applicable to a wide range of popular diffusion models.


A Time-Space Tradeoff for the Sumcheck Prover
Alessandro Chiesa*, Elisabetta Fedele*, Giacomo Fenzi*, Andrew Zitek-Estrada*
ZKProof Workshop 6, Berlin 2024
blog post / ePrint / code

Blendy is a new concretely efficient algorithm for multilinear sumcheck proving which uses linear time and sublinear memory.

web counter