Björn Michele

About

I am a Ph.D. student at the valeo.ai and the IRISA OBELIX lab, working on domain adaptation for 3D data. My Ph.D. advisor is Prof. Nicolas Courty who co-advises me with Alexandre Boulch. I also collaborate closely with Gilles Puy, Tuan-Hung Vu, and Renaud Marlet in my research.

I graduated from the Institute Polytechnique Paris with a Master's degree and a thesis on "Zero-shot learning for 3D data". My final year was the M2 Mathematics, Vision, and Apprenticeship (MVA) after completing the Computer Science M1 at the Ecole Polytechnique. I earned my Bachelor's degree from the Baden-Württemberg Cooperative State University in a joint study program with the Robert Bosch GmbH.

If you are looking for a comprehensive overview of publications on domain adaptation and domain generalization for 3D data, check out my curated list: Awesome Domain Adaptation and Generalization for 3D.

Research interests: Computer vision and frugal learning.

Publications

Björn Michele, Alexandre Boulch, Tuan-Hung Vu, Gilles Puy, Renaud Marlet, Nicolas Courty (2024). Train Till You Drop: Towards Stable and Robust Source-free Unsupervised 3D Domain Adaptation. ECCV, 2024.

Paul Berg, Björn Michele, Minh-Tan Phan, Laetitia Chapel, Nicolas Courty (2024). Horospherical Learning with Smart Prototypes. Oral @ BMVC, 2024.

Björn Michele, Alexandre Boulch, Gilles Puy, Tuan-Hung Vu, Renaud Marlet, Nicolas Courty (2024). SALUDA: Surface-based Automotive Lidar Unsupervised Domain Adaptation. Spotlight @ 3DV, 2024.

Alexandre Boulch, Corentin Sautier, Björn Michele, Gilles Puy, Renaud Marlet (2023). ALSO: Automotive Lidar Self-supervision by Occupancy estimation. CVPR, 2023.