Boosting Multi-Modal Unsupervised Domain Adaptation for LiDAR Semantic Segmentation by Self-Supervised Depth Completion

LiDAR semantic segmentation is receiving increased attention due to its deployment in autonomous driving applications. As LiDARs come often with other sensors such as RGB cameras, multi-modal approaches for this task have been developed, which however suffer from the domain shift problem deep learning approaches. To address this, we propose a novel Unsupervised Domain Adaptation (UDA) technique segmentation. Unlike previous works field, leverage depth completion an auxiliary align features extracted 2D images across domains, and powerful data augmentation LiDARs. We validate our method on three popular UDA benchmarks achieve better performances than competitors.