Juliette Bertrand

I relocated to New York and am looking for opportunities. If you want to reach out, feel free to email me. I am particularly interested in opportunities in video understanding for sport analysis, video generation or the impact of distribution shifts in real world settings, specifically for 3D/4D reconstruction.

Previously, I was a research engineer at the Thoth team collaborating with Professor Julien Mairal , Dr. James Hollingsworth and Dr. Sophie Giffard-Roisin . Previously, I was a researcher at the Czech Technical University in Prague, member of the VRG group and collaborating with Professor Giorgos Tolias and Dr Yannis Kalantidis at Naverlabs Europe . During my time with them, I focused on the challenging task of representation learning and matching for video understanding, particularly when dealing with limited annotated data.

• re-identifying soccer players through jersey number recognition;
• real-time ball tracking in low-resolution sports videos;
• predicting the primary topic of webpages from limited, noisy annotations (built from scratch and deployed in production for the 2018 EuroCup);
• detecting and reducing the influence of traffic spikes webpage view counts (deployed within the forecast tool);
• understanding daily human activities in home environments (Mobile Mii) ;
• inferring disparity maps around individuals' faces.

In 2013, I successfully completed my master's degree at Ecole Centrale Lyon, and I also graduated with distinction in Image and Signal Processing from the University of Lyon.

Outside of work, I enjoy hiking, dancing, acting, traveling, and embracing new cultures. I have a deep interest for cinema and the performing arts, to the extent that I even collaborated with a group of musical aficionados to create a musical entirely from scratch.

Email  /  Google Scholar  /  Twitter  /  Github

News

• [2025-11] - Presented MicroFlow as a poster at the NaverLabs 4th AI for Robotics workshop (poster).
• [2025-09] - Presented our work at the IS-Terre machine learning for geoscience seminar (slides).
• [2025-04] - Our work on Domain-Specific Optical Flow for Ground Deformation Estimation in Seismic Events is available on arXiv and Hugging Face.
• [2024-01] - Join the Thoth team at Inria as a research engineer.
• [2023-11] - Presented Test-time Training for Matching-based Video Object Segmentation as a poster at the NaverLabs 3rd AI for Robotics workshop
• [2023-09] - Our paper got accepted at NeurIPS.
• [2023-09] - Presented our work about limited data in video understanding at the Imagine Lab seminar (slides).
• [2023-05] - Our paper got accepted at the L3D-IVU workshop paper at CVPR.
• [2023-02] - Our paper got accepted at SCIA.
• [2022-07] - Our team won the essay competition award at ICVSS.

Blog

Coming soon.

Publications / Preprints

	MicroFlow: Domain-Specific Optical Flow for Ground Deformation Estimation in Seismic Events Juliette Bertrand, Sophie Giffard Roisin, James Hollingsworth, Julien Mairal Preprint, avril 2025 project page / arXiv / code / poster
	Test-time Training for Matching-based Video Object Segmentation Juliette Bertrand*, Giorgos Kordopatis-Zilos*, Yannis Kalantidis, Giorgos Tolias NeurIPS, 2023 open review / poster
	Rethinking matching-based few-shot action recognition Juliette Bertrand, Yannis Kalantidis, Giorgos Tolias SCIA, 2023 (Oral Presentation) L3D-IVU@CVPR, 2023 project page / arXiv / code

Many thanks to Jon Barron for providing the template for this page.