Welcome
I am associate professor of Computer Science at the Ensimag (School of Computer Science and Applied Mathematics, INP Grenoble University), and a researcher at the INRIA Grenoble Rhône-Alpes, France, with the Morpheo team since 2011.
I obtained my PhD from the Institut National Polytechnique de Grenoble in 2005 with the INRIA MOVI team. I started my professional career as a postdoctoral research assistant at the University of North Carolina's Computer Vision Group in 2006, and as assistant professor at the University of Bordeaux, with the IPARLA team, INRIA Bordeaux Sud-Ouest. My expertise and interest is in the field of computer vision, dynamic 3D modeling and 4D spatio-temporal modeling from multiple views, 3D interaction. Full CV available.
News
-
PAMI 2021 article accepted: A Visual Approach to Measure Cloth-Body and Cloth-Cloth Friction
Abdullah-Haroon Rasheed, Victor Romero, Florence Bertails-Descoubes, Stefanie Wuhrer, Jean-Sébastien Franco, Arnaud Lazarus -
IJCV 2021 article accepted: Volume Sweeping: Learning Photoconsistency for Multi-View Shape Reconstruction
Vincent Leroy, Jean-Sébastien Franco, Edmond Boyer -
TVCG 2020 article accepted: Mesh Denoising with Facet Graph Convolutions
Matthieu Armando, Jean-Sébastien Franco, Edmond Boyer -
ACCV 2020 paper accepted: Reconstructing Human Body Mesh from Point Clouds by Adversarial GP Network
Boyao Zhou, Jean-Sébastien Franco, Federica Bogo, Bugra Tekin, Edmond Boyer -
CVPR 2020 oral accepted: Learning to Measure the Static Friction Coefficient in Cloth Contact
Abdullah-Haroon Rasheed, Victor Romero, Florence Bertails-Descoubes, Stefanie Wuhrer, Jean-Sébastien Franco, Arnaud Lazarus -
ICCV 2019 paper accepted: Moulding Humans: Non-parametric 3D Human Shape Estimation from Single Images
Valentin Gabeur, Jean-Sébastien Franco, Xavier Martin, Cordelia Schmid, Gregory Rogez -
3DV 2019 paper accepted: Adaptive Mesh Texture for Multi-View Appearance Modeling
Matthieu Armando, Jean-Sébastien Franco, Edmond Boyer - 50% Delegation, September 2018-2020: awarded delegation teaching exemption grant for pursuit of research
-
PAMI 2018 article: Tracking-by-Detection of 3D Human Shapes: from Surfaces to Volumes
Chun Hao Huang, Benjamin Allain, Edmond Boyer, Jean-Sébastien Franco, Federico Tombari, Nassir Navab, Slobodan Ilic -
ECCV 2018 paper accepted: Shape Reconstruction Using Volume Sweeping and Learned Photoconsistency
Vincent Leroy, Jean-Sébastien Franco, Edmond Boyer -
ECCV 2018 paper accepted: Analyzing Clothing Layer Deformation Statistics of 3D Human Motions
Jinlong Yang, Jean-Sébastien Franco, Franck Hétroy-Wheeler, Stefanie Wuhrer -
ICCV 2017 paper accepted: Multi-View Dynamic Shape Refinement Using Local Temporal Integration
Vincent Leroy, Jean-Sébastien Franco, Edmond Boyer -
CVPR 2017 paper accepted: Surface Motion Capture Transfer with Gaussian Process Regression
Adnane Boukhayma, Jean-Sébastien Franco, Edmond Boyer - QuickCSG, our new boolean computation software, released.
Research Interests
4D Temporal Alignment and Shape Tracking

Tracking Under Clothing

Deformable Shape Analysis


Multi-View Segmentation


Temporal Appearance Representations and Superresolution
Fast Booleans on Meshes

QuickCSG: Fast Variadic Polyhedral Booleans
Efficient Silhouette-based Modeling


Probabilistic Occupancy Analysis

Scene Flow


Interactive Vision Systems

3D Telepresence with Multi-Camera Systems
QuickCSG
We are releasing binaries of our new software, QuickCSG, which can compute visual hulls as well as arbitrary boolean combinations on N polyhedral solids. Current tests on our 68-camera Kinovis platform show that QuickCSG outperforms our previous visual hull computation software EPVH by a factor of 20x on 40 cameras or more. Our experimental tests also show that QuickCSG significantly outperforms all currently known available boolean CSG codes in terms of speed on two or more inputs. A publication is currently submitted for review but a preliminary version is available as research report.
EPVH
EPVH is one of the projects developed during my thesis. I receive a number of requests for providing the code or binary. I can no longer distribute either, as EPVH is now exclusively distributed by 4DViews, and built-in 4DViews software suites. If you have datasets on which you would like to perform comparisons with EPVH for research purposes, I can process them and provide the resulting data for static objects or sequences. However, you are encouraged to use QuickCSG instead, which significantly outperforms EPVH.