Raoul de Charette

Institut national de recherche en sciences et technologies du numérique, ITS (United Kingdom)

Papers

Total Citations

H-Index

About

Raoul de Charette is a researcher specializing in 3D scene understanding, autonomous perception, and LiDAR-based environment modeling — areas critical to the advancement of autonomous vehicles and robotics. His work focuses on enabling machines to construct rich, accurate representations of complex urban environments from sparse sensor data. Among his most notable contributions is the introduction of **Panoptic Scene Completion (PSC)** through his 2024 paper *PaSCo*, which extends the widely studied Semantic Scene Completion paradigm by incorporating instance-level information, yielding significantly more detailed 3D scene understanding. This work has already garnered 25 citations, reflecting strong community interest. His earlier *LMSCNet* (2020) demonstrated that lightweight, multiscale 2D-to-3D architectures could achieve competitive semantic scene completion from LiDAR inputs without the computational overhead of purely volumetric approaches. His foundational work on statistical occupancy grid modeling from range sensors further underscores his breadth across both classical probabilistic methods and modern deep learning techniques. De Charette's research trajectory — from principled probabilistic 3D reconstruction to panoptic deep learning — reflects a researcher steadily pushing the boundaries of how autonomous systems perceive and interpret their surroundings, making his work highly relevant to students and practitioners in robotics and computer vision alike.