Few-shot 3D LiDAR Semantic Segmentation for Autonomous Driving
This addresses the challenge of novel objects and lack of annotations in autonomous driving, though it is incremental as it builds on existing few-shot learning approaches.
The paper tackles the problem of few-shot 3D LiDAR semantic segmentation for autonomous driving by proposing a method that predicts both novel and base classes simultaneously, achieving effectiveness as demonstrated on the SemanticKITTI dataset.
In autonomous driving, the novel objects and lack of annotations challenge the traditional 3D LiDAR semantic segmentation based on deep learning. Few-shot learning is a feasible way to solve these issues. However, currently few-shot semantic segmentation methods focus on camera data, and most of them only predict the novel classes without considering the base classes. This setting cannot be directly applied to autonomous driving due to safety concerns. Thus, we propose a few-shot 3D LiDAR semantic segmentation method that predicts both novel classes and base classes simultaneously. Our method tries to solve the background ambiguity problem in generalized few-shot semantic segmentation. We first review the original cross-entropy and knowledge distillation losses, then propose a new loss function that incorporates the background information to achieve 3D LiDAR few-shot semantic segmentation. Extensive experiments on SemanticKITTI demonstrate the effectiveness of our method.