Robin Magnet
Postdoc, Inria Paris

I am currently a postdoc with Jean Feydy in the HeKA group at Inria Paris, where I work on shape and volume analysis with medical imaging, applied to bone fracture detection.

Before that, I obtained my PhD under the supervision of Maks Ovsjanikov at École Polytechnique in France (in the GeomeriX Team at LIX laboratory). During my PhD, I studied scalable shape comparison using functional maps and spectral methods.

I am also the maintainer of the pyfmaps (pyFM) library, which provides simple bindings for several powerful functional maps methods.

Email : robin.magnet@inria.fr

News

I released a python implementation of the Reversible Harmonic Maps paper. Check it out here

I started a postdoc at Inria Paris in the Heka Team.

I defended my PhD thesis !

My package PyFM has been released on PyPI in version 1.0.0.

Our paper Scalable and Simplified Functional Map Learning with Maks Ovsjanikov has been accepted at CVPR 2024.

I did a Research Scientist Internship at Meta Reality Labs in Pittsburgh

see more

Our paper Assessing craniofacial growth and form without landmarks: A new automatic approach based on spectral methods with Kevin Bloch, Maxime Taverne, Simone Melzi, Maya Geoffroy, Roman H. Khonsari and Maks Ovsjanikov has been published in the Journal of Morphology.

Our paper Scalable and Efficient Functional Map Computations on Dense Meshes with Maks Ovsjanikov has been accepted at Eurographics 2023.

I starded a position of Teaching Assistant for class INF631 - Analysis and Deep Learning on Geometric Data (2022-2023) at Ecole Polytechnique

I starded a position of Teaching Assistant for class INF573 - Image Analysis and Computer Vision (2022-2023) at Ecole Polytechnique

Our paper Smooth Non-Rigid Shape Matching via Effective Dirichlet Energy Optimization has received the best paper award at 3DV 2022

Our paper Smooth Non-Rigid Shape Matching via Effective Dirichlet Energy Optimization with Jing Ren, Olga Sorkine-Hornung and Maks Ovsjanikov has been accepted at 3DV 2022 as an oral presentation

I starded a position of Teaching Assistant for class INF473V - Modal d'informatique - Deep Learning in Computer Vision (2021-2022) at Ecole Polytechnique

I starded a position of Teaching Assistant for class INF573 - Image Analysis and Computer Vision (2021-2022) at Ecole Polytechnique

Our paper DWKS : A Local Descriptor of Deformations Between Meshes and Point Clouds with Maks Ovsjanikov has been accepted at ICCV 2021


Research

Robust Spectral Methods for Shape Analysis and Deformation Assessment

Robin Magnet
Ph.D. dissertation, Institut Polytechnique de Paris, 2024..

[compressed] [pdf]

Scalable and Simplified Functional Map Learning

Robin Magnet, Maks Ovsjanikov
Proc. International Conference on Computer Vision and Pattern Recognition (CVPR), 2024.

[pdf] [suppl]

Assessing craniofacial growth and form without landmarks: A new automatic approach based on spectral methods

Robin Magnet, Kevin Bloch, Maxime Taverne, Simone Melzi, Maya Geoffroy, Roman H. Khonsari, Maks Ovsjanikov
Journal of Morphology, 2023

[pdf]

Scalable and Efficient Functional Map Computations on Dense Meshes

Robin Magnet, Maks Ovsjanikov
Proc. Eurographics, 2023 (Oral)

[pdf]

Smooth Non-Rigid Shape Matching via Effective Dirichlet Energy Optimization

Robin Magnet, Jing Ren, Olga Sorkine-Hornung, Maks Ovsjanikov
International Conference on 3D Vision (3DV), 2022 (Oral)
Best paper award

[pdf] [supplementary]

DWKS : A Local Descriptor of Deformations Between Meshes and Point Clouds

Robin Magnet, Maks Ovsjanikov
Proc. International Conference on Computer Vision (ICCV), 2021

[pdf] [supplementary]


Code

Python bindings for functional maps related computations.

A lightweight library for a memory scalable unified representation of correspondence

Python bindings for Discrete Optimization and Smooth Discrete Optimization.