3 papers
cs.CV2026
RoAD Benchmark: How LiDAR Models Fail under Coupled Domain Shifts and Label Evolution
Subeen Lee, Siyeong Lee, Namil Kim +1
For 3D perception systems to operate reliably in real-world environments, they must remain robust to evolving sensor characteristics and changes in object taxonomies. However, exis…
cs.CV2026
VERIA: Verification-Centric Multimodal Instance Augmentation for Long-Tailed 3D Object Detection
Jumin Lee, Siyeong Lee, Namil Kim +1
Long-tail distributions in driving datasets pose a fundamental challenge for 3D perception, as rare classes exhibit substantial intra-class diversity yet available samples cover th…
cs.CV2024
Just Add $100 More: Augmenting NeRF-based Pseudo-LiDAR Point Cloud for Resolving Class-imbalance Problem
Mincheol Chang, Siyeong Lee, Jinkyu Kim +1
Typical LiDAR-based 3D object detection models are trained in a supervised manner with real-world data collection, which is often imbalanced over classes (or long-tailed). To deal…