Mathieu Gonzalez

Institut de Recherche Technologique B-com

Papers

Total Citations

H-Index

About

Mathieu Gonzalez is a computer vision researcher specializing in 6 degrees-of-freedom (6DoF) object pose estimation, with a focus on developing efficient deep learning solutions for robotics and augmented reality applications. His work addresses one of the field's most persistent challenges: accurately determining an object's 3D position and orientation from visual data, particularly in cluttered real-world environments. Gonzalez's most recognized contribution, **L6DNet** (2020, 16 citations), introduced a lightweight hybrid pipeline capable of achieving robust and precise pose estimation from single RGB-D images while requiring only small training datasets — a significant practical advantage over data-hungry alternatives. Building on this foundation, his subsequent work **YOLOff** (2020, 4 citations) further refined the approach by learning geometric offsets directly, improving robustness in complex, occluded scenes. A recurring theme across his research is the design of hybrid architectures that combine data-driven learning with geometric reasoning, striking a balance between accuracy and computational efficiency. His contributions are particularly relevant for researchers and engineers working on robotic manipulation, industrial automation, and mixed reality systems, where precise spatial understanding of objects is critical to real-world deployment.