Description

Cette thèse vise le développement de méthodes avancées de perception 3D LiDAR pour la navigation autonome de robots en environnements ouverts et déstructurés. Elle s’appuie sur la segmentation sémantique en vocabulaire ouvert (c’est-à-dire exploitant le vocabulaire des LLM pour Large Language Models) et sur le transfert de connaissances issues de VLM pour Visual Language Models vers la 3D afin de rendre possible la compréhension de scènes 3D en terrain naturel.

Compétences requises

- Diplôme niveau Master 2 (BAC+5) - Rigueur scientifique et autonomie - Connaissances en Vision 3D, Machine Learning, Deep Learning - Bon niveau de programmation en Python, PyTorch et CUDA

Bibliographie

- Holger Caesar, Varun Bankiti, Alex H. Lang, Sourabh Vora, Venice Erin Liong, Qiang Xu, Anush Krishnan, Yu Pan, Giancarlo Baldan, and Oscar Beijbom. nuscenes: A multimodal dataset for autonomous driving. In IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2020.
- Peng Jiang, Philip Osteen, Maggie Wigness, and Srikanth Saripalli. Rellis-3d dataset: Data, benchmarks and analysis. In IEEE International Conference on Robotics and Automation (ICRA), 2021.
- Karim Knaebel, Kadir Yilmaz, Daan de Geus, Alexander Hermans, David Adrian, Timm Linder, and Bastian Leibe. DINO in the Room: Leveraging 2D Foundation Models for 3D Segmentation. arXiv e-prints, page arXiv:2503.18944, March 2025. arXiv:2503.18944, doi:10.48550/arXiv.2503.18944.
- Youquan Liu, Lingdong Kong, Jun Cen, Runnan Chen, Wenwei Zhang, Liang Pan, Kai Chen, and Ziwei Liu. Segment any point cloud sequences by distilling vision foundation models. In International Conference on Neural Information Processing Systems (NeurIPS), 2023.
- Peter Mortimer, Raphael Hagmanns, Miguel Granero, Thorsten Luettel, Janko Petereit, and Hans- Joachim Wuensche. The goose dataset for perception in unstructured environments. In IEEE International Conference on Robotics and Automation (ICRA), 2024.
- Gilles Puy, Spyros Gidaris, Alexandre Boulch, Oriane Siméoni, Corentin Sautier, Patrick Pérez, Andrei Bursuc, and Renaud Marlet. Three pillars improving vision foundation model distillation for lidar. In IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024.
- Nermin Samet, Gilles Puy, and Renaud Marlet. LOSC: LiDAR Open-voc Segmentation Consolidator. arXiv e-prints, page arXiv:2507.07605, July 2025. arXiv:2507.07605, doi:10.48550/arXiv.2507.07605. Oriane Siméoni, Huy V. Vo, Maximilian Seitzer, Federico Baldassarre, Maxime Oquab, Cijo Jose, Vasil
- Khalidov, Marc Szafraniec, Seungeun Yi, Michaël Ramamonjisoa, Francisco Massa, Daniel Haziza, Luca Wehrstedt, Jianyuan Wang, Timothée Darcet, Théo Moutakanni, Leonel Sentana, Claire Roberts, Andrea Vedaldi, Jamie Tolan, John Brandt, Camille Couprie, Julien Mairal, Hervé Jégou, Patrick Labatut, and Piotr Bojanowski. DINOv3. arXiv e-prints, page arXiv:2508.10104, August 2025. arXiv:2508.10104, doi:10.48550/arXiv.2508.10104.
- Chandan Yeshwanth, Yueh-Cheng Liu, Matthias Nießner, and Angela Dai. Scannet++: A high-fidelity dataset of 3d indoor scenes. In IEEE/CVF International Conference on Computer Vision (ICCV), 2023.
- Yujia Zhang, Xiaoyang Wu, Yunhan Yang, Xianzhe Fan, Han Li, Yuechen Zhang, Zehao Huang, Naiyan Wang, and Hengshuang Zhao. Utonia: Toward One Encoder for All Point Clouds. arXiv e-prints, page arXiv:2603.03283, March 2026. arXiv:2603.03283.

Mots clés

LiDAR, Perception 3D, Robotique, VLM, Distillation, Terrain naturel