Missions

As part of the binational project "HuMoCar: Realistic Human Models for Care Robots for Aged People" (October 2021 - October 2025), we aim to push the current limits of robot vision in human cognition by making its vision-intelligence robust to large variations and to occlusion. The project also aims to enhance the capability of understanding certain human-object interactions by developing a photo-realistic, physics-aware 4D human model. The specific goal within this framework is to develop a predictive/generative model for dressed human and clothes in motion through deep learning over annotated datasets. Several downstream tasks will be defined and implemented, involving various conditional generations. Geometric deep learning models such as GCN and PointNet will be deployed to process 3D surface data. The project will take place within the MLMS research team (Machine Learning, Modeling & Simulation, https://mlms.icube.unistra.fr/), located at the hospital site of the laboratory.

Activities

- Research and development on the aforementioned theme: Predictive/generative model of clothing and dressed humans. This involves predicting the shape and physical behavior of dressed individuals, taking into account the laws of physics and 3D simulation.
- Technical management and collaboration with other researchers (PhD candidates, post-docs, and/or permanent researchers) participating in the project: This involves organizing the work among team members, assigning responsibilities, while ensuring the technical coherence of the project.
- Possible supervision of Master students.

Skills

− PhD in Computer Science (2022 or later).
− Skills in programming, designing efficient algorithms, and writing articles.
− Solid knowledge and experience in numerical simulation and deep learning.
− English level C or higher.
− Team player.

Work Context

ICube Laboratory (The Engineering science, computer science and imaging laboratory) at the University of Strasbourg is a leading research center in Computer Science, with more than 300 permanent researchers. Within the framework of a bi-national (with South Korea) and tri-institutional project (HuMoCar : Realistic Human Models for Care Robots for Aged People, oct. 2021 - oct. 2025), with CNRS, INRIA and ETRI, we aim to push the current limits of robot vision in human cognition by care-robots in the in-house situation. Our specific goal is to make the performance of the vision-intelligence robust to large variations (in body shapes, motions,..) to occlusion (cloth, furniture, wall,..), and capable of understanding the interaction by developing a photo-realistic, physics-aware 4D human model.

Additional Information

Local candidates will be given priority.