Matthieu Lecce

Papers

Total Citations

289

H-Index

About

Matthieu Lecce is a leading researcher in robotic perception, with a focus on 3D object detection, pose estimation, and shape recovery for manipulation. His work tackles the fundamental challenge of enabling robots to perceive and interact with objects in cluttered, real-world environments—including those that are notoriously difficult for traditional computer vision systems. Lecce’s most influential contribution is his 2014 paper on single-image 3D object detection and pose estimation for grasping, which has garnered over 227 citations. This work introduced a deformable parts-based model trained on silhouette clusters, enabling robust detection and pose estimation of textureless objects from a single view—a critical capability for robotic grasping. He has also made pioneering advances in perceiving transparent objects, such as glassware, a long-standing open problem in robotics. His 2016 paper on this topic, with 50 citations, combined learned detectors with projective geometry to recover pose and shape. Additionally, Lecce addressed the practical challenge of grasping rotationally symmetric objects without known 3D models, developing a two-view method for simultaneous pose and shape recovery. Through these contributions, Lecce has significantly advanced the state of the art in robot perception, bridging the gap between vision and physical interaction.