Papers
2
Total Citations
50
H-Index
2
About
Renaud Marlet is a leading researcher in computer vision and robotics, with a core focus on 3D object understanding, pose estimation, and geometric deep learning. His most influential work addresses the fundamental challenge of enabling machines to perceive and interact with arbitrary objects in unconstrained environments. Marlet’s landmark 2019 paper, “Pose from Shape: Deep Pose Estimation for Arbitrary 3D Objects” (36 citations), introduced a paradigm-shifting approach that eliminates the need for category-specific training or canonical object orientations, allowing deep networks to estimate the pose of any 3D object from a single image. This work has been widely recognized for its potential to generalize across diverse real-world scenarios. In his 2018 study, “Virtual Training for a Real Application: Accurate Object-Robot Relative Localization without Calibration” (14 citations), Marlet tackled the practical challenge of robotic manipulation by demonstrating how synthetic training data can enable precise object-robot localization using uncalibrated cameras—a critical step toward autonomous systems that operate without tedious manual setup. His contributions bridge the gap between theoretical computer vision and applied robotics, with a strong emphasis on robustness, generalization, and real-world deployment. Marlet’s research continues to influence the development of more flexible, data-efficient perception systems.
Research Focus
Key Achievements
Top Papers
- 1Pose from Shape: Deep Pose Estimation for Arbitrary 3D Objects36 citations · 2019
- 2