Publications
Publications
Take a further look into our breakthroughs
Take a further look into our breakthroughs
Take a further look into our breakthroughs
December 15, 2025
Comparing 3D Data Representations for Keypoint Estimation in Humanoid Characters
Comparing 3D Data Representations for Keypoint Estimation in Humanoid Characters
JE. Lefèvre, T. Cheynel, T. Daniel
JE. Lefèvre, T. Cheynel, T. Daniel
This work investigates how the use of different 3D data representations affects the performance of learning-based models for the task of predicting keypoints from meshes of humanoid characters, used for rigging.
This work investigates how the use of different 3D data representations affects the performance of learning-based models for the task of predicting keypoints from meshes of humanoid characters, used for rigging.
February 28, 2025
ReConForM: Real-time Contact-aware Motion Retargeting for more Diverse Character Morphologies
ReConForM: Real-time Contact-aware Motion Retargeting for more Diverse Character Morphologies
T. Cheynel, T. Rossi, B. Bellot-Gurlet, D. Rohmer, MP. Cani
T. Cheynel, T. Rossi, B. Bellot-Gurlet, D. Rohmer, MP. Cani
Preserving semantics, in particular in terms of contacts, is a key challenge when retargeting motion between characters of different morphologies.
Preserving semantics, in particular in terms of contacts, is a key challenge when retargeting motion between characters of different morphologies.
December 2, 2024
Rethinking motion keyframe extraction: A greedy procedural approach using a neural control rig
Rethinking motion keyframe extraction: A greedy procedural approach using a neural control rig
T. Cheynel, O. El Khalifi, B. Bellot-Gurlet, D. Rohmer, MP. Cani
T. Cheynel, O. El Khalifi, B. Bellot-Gurlet, D. Rohmer, MP. Cani
3D animators traditionally use a "pose to pose" approach, whereas motion capture (MoCap) tools generate a pose for every frame, making the motion challenging to edit.
3D animators traditionally use a "pose to pose" approach, whereas motion capture (MoCap) tools generate a pose for every frame, making the motion challenging to edit.
November 15, 2023
Sparse Motion Semantics for Contact-Aware Retargeting
Sparse Motion Semantics for Contact-Aware Retargeting
T. Cheynel, T. Rossi, B. Bellot-Gurlet, D. Rohmer, MP. Cani
T. Cheynel, T. Rossi, B. Bellot-Gurlet, D. Rohmer, MP. Cani
This paper presents a method for retargeting motion onto a character with a completely different skeleton and mesh.
Sparse Motion Semantics for Contact-Aware Retargeting
T. Cheynel, T. Rossi, B. Bellot-Gurlet, D. Rohmer, MP. Cani
This paper presents a method for retargeting motion onto a character with a completely different skeleton and mesh.
November 15, 2023
Rethinking motion keyframe extraction: A greedy procedural approach using a neural control rig
T. Cheynel, O. El Khalifi, B. Bellot-Gurlet, D. Rohmer, MP. Cani
3D animators traditionally use a "pose to pose" approach, whereas motion capture (MoCap) tools generate a pose for every frame, making the motion challenging to edit.
November 15, 2023
ReConForM: Real-time Contact-aware Motion Retargeting for more Diverse Character Morphologies
T. Cheynel, T. Rossi, B. Bellot-Gurlet, D. Rohmer, MP. Cani
Preserving semantics, in particular in terms of contacts, is a key challenge when retargeting motion between characters of different morphologies.
February 28, 2025
ReConForM: Real-time Contact-aware Motion Retargeting for more Diverse Character Morphologies
T. Cheynel, T. Rossi, B. Bellot-Gurlet, D. Rohmer, MP. Cani
Preserving semantics, in particular in terms of contacts, is a key challenge when retargeting motion between characters of different morphologies.
February 28, 2025
Comparing 3D Data Representations for Keypoint Estimation in Humanoid Characters
JE. Lefèvre, T. Cheynel, T. Daniel
This work investigates how the use of different 3D data representations affects the performance of learning-based models for the task of predicting keypoints from meshes of humanoid characters, used for rigging.
December 15, 2025
![](data:image/svg+xml,<svg display="block" role="presentation" viewBox="0 0 50 50" xmlns="http://www.w3.org/2000/svg"><path d="M 22.49 22 L 13.695 9.178 L 13.71 9.19 L 21.64 0 L 18.99 0 L 12.53 7.48 L 7.4 0 L 0.45 0 L 8.661 11.971 L 8.66 11.97 L 0 22 L 2.65 22 L 9.832 13.678 L 15.54 22 Z M 6.35 2 L 18.69 20 L 16.59 20 L 4.24 2 Z" fill="rgb(255, 255, 255)" height="22px" id="fqKB6LiGl" transform="translate(13.88 14)" width="22.490000000000002px"/></svg>)
![](data:image/svg+xml,<svg display="block" role="presentation" viewBox="0 0 50 50" xmlns="http://www.w3.org/2000/svg"><path d="M 6 0 C 7.591 0 9.117 0.632 10.243 1.757 C 11.368 2.883 12 4.409 12 6 L 12 13 L 8 13 L 8 6 C 8 5.47 7.789 4.961 7.414 4.586 C 7.039 4.211 6.53 4 6 4 C 5.47 4 4.961 4.211 4.586 4.586 C 4.211 4.961 4 5.47 4 6 L 4 13 L 0 13 L 0 6 C 0 4.409 0.632 2.883 1.757 1.757 C 2.883 0.632 4.409 0 6 0 Z" fill="rgb(255, 255, 255)" height="13px" id="vKyABfctg" transform="translate(23 21)" width="12px"/><path d="M 4 0 L 0 0 L 0 12 L 4 12 Z" fill="rgb(255, 255, 255)" height="12px" id="RQzQ7KrNa" transform="translate(15 22)" width="4px"/><path d="M 2 4 C 3.105 4 4 3.105 4 2 C 4 0.895 3.105 0 2 0 C 0.895 0 0 0.895 0 2 C 0 3.105 0.895 4 2 4 Z" fill="rgb(255, 255, 255)" height="4px" id="QsdvYOgEL" transform="translate(15 15)" width="4px"/></svg>)
![](data:image/svg+xml,<svg display="block" role="presentation" viewBox="0 0 50 50" xmlns="http://www.w3.org/2000/svg"><path d="M 21.541 2.42 C 21.422 1.945 21.18 1.511 20.839 1.159 C 20.499 0.808 20.071 0.553 19.601 0.42 C 17.881 0 11.001 0 11.001 0 C 11.001 0 4.121 0 2.401 0.46 C 1.93 0.593 1.503 0.848 1.162 1.199 C 0.821 1.551 0.579 1.985 0.461 2.46 C 0.146 4.206 -0.008 5.976 0.001 7.75 C -0.011 9.537 0.143 11.321 0.461 13.08 C 0.591 13.54 0.839 13.958 1.179 14.294 C 1.519 14.631 1.939 14.874 2.401 15 C 4.121 15.46 11.001 15.46 11.001 15.46 C 11.001 15.46 17.881 15.46 19.601 15 C 20.071 14.867 20.499 14.612 20.839 14.261 C 21.18 13.909 21.422 13.475 21.541 13 C 21.853 11.268 22.007 9.51 22.001 7.75 C 22.012 5.963 21.858 4.179 21.541 2.42 Z" fill="rgb(255, 255, 255)" height="15.46px" id="FuH4e4plM" stroke-dasharray="" stroke-linecap="round" stroke-linejoin="round" stroke-width="2" stroke="rgb(255, 255, 255)" transform="translate(14 17.2)" width="22.00113913890833px"/><path d="M 0 6.54 L 5.75 3.27 L 0 0 Z" fill="rgb(0, 0, 0)" height="6.539999999999999px" id="RfLEo_tLX" transform="translate(22.75 21.68)" width="5.75px"/></svg>)