3 papers
cs.CV2026
Rethinking Transfer Learning for Industrial Inspection: DINOv3 vs. ImageNet Pretraining Across RGB and X-ray Tasks
Mehdi Gharbage, Céline Teulière, Pierre Bouges +1
Vision foundation models pretrained on web-scale data have recently shown strong transfer capabilities on many downstream tasks, but their effectiveness for industrial visual inspe…
cs.CV2026
Multi-View Projection for Unsupervised Domain Adaptation in 3D Semantic Segmentation
Andrew Caunes, Thierry Chateau, Vincent Fremont
3D semantic segmentation plays a pivotal role in autonomous driving and road infrastructure analysis, yet state-of-the-art 3D models are prone to severe domain shift when deployed…
cs.CV2025
3D Can Be Explored In 2D: Pseudo-Label Generation for LiDAR Point Clouds Using Sensor-Intensity-Based 2D Semantic Segmentation
Andrew Caunes, Thierry Chateau, Vincent Frémont
Semantic segmentation of 3D LiDAR point clouds, essential for autonomous driving and infrastructure management, is best achieved by supervised learning, which demands extensive ann…