publications
List of publications.
2025
- JournalScanmove: Motion prediction and transfer for unregistered body meshesThomas Besnier, Sylvain Arguillère, and Mohamed DaoudiComputers & Graphics, 2025
Unregistered surface meshes, especially raw 3D scans, present significant challenges for automatic computation of plausible deformations due to the lack of established point-wise correspondences and the presence of noise in the data. In this paper, we propose a new, rig-free, data-driven framework for motion prediction and transfer on such body meshes. Our method couples a robust motion embedding network with a learned per-vertex feature field to generate a spatio-temporal deformation field, which drives the mesh deformation. Extensive evaluations, including quantitative benchmarks and qualitative visuals on tasks such as walking and running, demonstrate the effectiveness and versatility of our approach on challenging unregistered meshes.
@article{BESNIER2025104409, title = {Scanmove: Motion prediction and transfer for unregistered body meshes}, journal = {Computers & Graphics}, volume = {132}, pages = {104409}, year = {2025}, issn = {0097-8493}, doi = {https://doi.org/10.1016/j.cag.2025.104409}, url = {https://www.sciencedirect.com/science/article/pii/S009784932500250X}, author = {Besnier, Thomas and Arguillère, Sylvain and Daoudi, Mohamed}, keywords = {3D computer vision, Unregistered meshes, Motion transfer, Human body motion prediction}, category = {journal-articles}, }
2024
- PreprintBeyond Fixed Topologies: Unregistered Training and Comprehensive Evaluation Metrics for 3D Talking HeadsFederico Nocentini, Thomas Besnier, Claudio Ferrari, and 3 more authors2024
@misc{nocentini2024fixedtopologiesunregisteredtraining, title = {Beyond Fixed Topologies: Unregistered Training and Comprehensive Evaluation Metrics for 3D Talking Heads}, author = {Nocentini, Federico and Besnier, Thomas and Ferrari, Claudio and Arguillere, Sylvain and Berretti, Stefano and Daoudi, Mohamed}, year = {2024}, eprint = {2410.11041}, archiveprefix = {arXiv}, primaryclass = {cs.CV}, url = {https://arxiv.org/abs/2410.11041}, category = {preprints} } - PreprintPartial Non-rigid Deformations and interpolations of Human Body SurfacesThomas Besnier, Emery Pierson, Sylvain Arguillere, and 1 more author2024
@misc{besnier2024partialnonrigiddeformationsinterpolations, title = {Partial Non-rigid Deformations and interpolations of Human Body Surfaces}, author = {Besnier, Thomas and Pierson, Emery and Arguillere, Sylvain and Daoudi, Mohamed}, year = {2024}, eprint = {2412.02306}, archiveprefix = {arXiv}, primaryclass = {cs.CV}, url = {https://arxiv.org/abs/2412.02306}, category = {preprints} } - Conference
ScanTalk: 3D Talking Heads from Unregistered ScansFederico Nocentini, Thomas Besnier, Claudio Ferrari, and 3 more authors2024@misc{nocentini2024scantalk3dtalkingheads, title = {ScanTalk: 3D Talking Heads from Unregistered Scans}, author = {Nocentini, Federico and Besnier, Thomas and Ferrari, Claudio and Arguillere, Sylvain and Berretti, Stefano and Daoudi, Mohamed}, year = {2024}, booktitle = {Proceedings of the IEEE/CVF European Conference on Computer Vision (ECCV)}, eprint = {2403.10942}, archiveprefix = {arXiv}, primaryclass = {cs.CV}, category = {conference-papers} }
2023
- Journal
Toward Mesh-Invariant 3D Generative Deep Learning with Geometric MeasuresThomas Besnier, Sylvain Arguillère, Emery Pierson, and 1 more authorComputers & Graphics, 20233D generative modeling is accelerating as the technology allowing the capture of geometric data is developing. However, the acquired data is often inconsistent, resulting in unregistered meshes or point clouds. Many generative learning algorithms require correspondence between each point when comparing the predicted shape and the target shape. We propose an architecture able to cope with different parameterizations, even during the training phase. In particular, our loss function is built upon a kernel-based metric over a representation of meshes using geometric measures such as currents and varifolds. The latter allows to implement an efficient dissimilarity measure with many desirable properties such as robustness to resampling of the mesh or point cloud. We demonstrate the efficiency and resilience of our model with a generative learning task of human faces.
@article{BESNIER2023, title = {Toward Mesh-Invariant 3D Generative Deep Learning with Geometric Measures}, journal = {Computers & Graphics}, volume = {115}, pages = {309-320}, year = {2023}, issn = {0097-8493}, doi = {https://doi.org/10.1016/j.cag.2023.06.027}, url = {https://www.sciencedirect.com/science/article/pii/S009784932300122X}, author = {Besnier, Thomas and Arguillère, Sylvain and Pierson, Emery and Daoudi, Mohamed}, category = {journal-articles}, keywords = {3D generative model, Unsupervised learning, Geometric measures} } - Conference
A Function Space Perspective on Stochastic Shape EvolutionElizabeth Baker, Thomas Besnier, and Stefan SommerIn Image Analysis, 2023Modelling randomness in shape data, for example, the evolution of shapes of organisms in biology, requires stochastic models of shapes. This paper presents a new stochastic shape model based on a description of shapes as functions in a Sobolev space. Using an explicit orthonormal basis as a reference frame for the noise, the model is independent of the parameterisation of the mesh. We define the stochastic model, explore its properties, and illustrate examples of stochastic shape evolutions using the resulting numerical framework.
@inproceedings{Baker_Besnier_2022, author = {Baker, Elizabeth and Besnier, Thomas and Sommer, Stefan}, editor = {Gade, Rikke and Felsberg, Michael and K{\"a}m{\"a}r{\"a}inen, Joni-Kristian}, title = {A Function Space Perspective on Stochastic Shape Evolution}, booktitle = {Image Analysis}, year = {2023}, publisher = {Springer Nature Switzerland}, address = {Cham}, pages = {278--292}, isbn = {978-3-031-31438-4}, category = {conference-papers} }
