Lorenzo Bertoni

École Polytechnique Fédérale de Lausanne

Papers

Total Citations

H-Index

About

Lorenzo Bertoni is a leading researcher in computer vision, with a primary focus on human pose estimation and semantic keypoint detection for autonomous systems. His most influential contributions center on developing efficient, bottom-up methods for multi-person pose estimation that are particularly well-suited for real-world applications like self-driving cars and delivery robots. Bertoni is the creator of the PifPaf and OpenPifPaf frameworks, which introduce innovative "composite fields"—specifically Part Intensity Fields (PIF) for localizing body parts and Part Association Fields (PAF) for associating them into full poses. His 2021 paper, "OpenPifPaf: Composite Fields for Semantic Keypoint Detection and Spatio-Temporal Association," extends this work to general keypoint detection and tracking, earning 22 citations. His foundational 2019 work, "PifPaf: Composite Fields for Human Pose Estimation," has garnered 20 citations and remains a benchmark in the field. Bertoni's work is notable for its practical impact on urban mobility, enabling robust perception in dynamic environments. His open-source implementations have become widely adopted tools for researchers and engineers working on embodied AI and autonomous navigation.