Generalized Few-Shot Point Cloud Segmentation Via Geometric Words
This addresses a practical limitation in point cloud segmentation for applications like robotics and autonomous driving, though it is incremental as it builds on existing few-shot learning paradigms.
The paper tackles the problem of few-shot point cloud segmentation where models struggle to adapt to new classes while maintaining accuracy on base classes, proposing geometric words and prototypes to achieve better generalization without forgetting old classes, with experiments on S3DIS and ScanNet showing superior performance over baselines.
Existing fully-supervised point cloud segmentation methods suffer in the dynamic testing environment with emerging new classes. Few-shot point cloud segmentation algorithms address this problem by learning to adapt to new classes at the sacrifice of segmentation accuracy for the base classes, which severely impedes its practicality. This largely motivates us to present the first attempt at a more practical paradigm of generalized few-shot point cloud segmentation, which requires the model to generalize to new categories with only a few support point clouds and simultaneously retain the capability to segment base classes. We propose the geometric words to represent geometric components shared between the base and novel classes, and incorporate them into a novel geometric-aware semantic representation to facilitate better generalization to the new classes without forgetting the old ones. Moreover, we introduce geometric prototypes to guide the segmentation with geometric prior knowledge. Extensive experiments on S3DIS and ScanNet consistently illustrate the superior performance of our method over baseline methods. Our code is available at: https://github.com/Pixie8888/GFS-3DSeg_GWs.