Learning Hierarchical Orthogonal Prototypes for Generalized Few-Shot 3D Point Cloud Segmentation
This work addresses the stability-plasticity trade-off in few-shot learning for 3D segmentation, offering a solution to prevent base-class forgetting when adapting to novel classes, which is incremental but improves performance in a domain-specific context.
The paper tackles the challenge of generalized few-shot 3D point cloud segmentation, where adapting to novel classes from few annotations can degrade base-class performance, by proposing HOP3D, a framework that uses hierarchical orthogonal prototypes and an entropy-based regularizer to achieve robust adaptation without forgetting, outperforming state-of-the-art methods on datasets like ScanNet200 and ScanNet++ in 1-shot and 5-shot settings.
Generalized few-shot 3D point cloud segmentation aims to adapt to novel classes from only a few annotations while maintaining strong performance on base classes, but this remains challenging due to the inherent stability-plasticity trade-off: adapting to novel classes can interfere with shared representations and cause base-class forgetting. We present HOP3D, a unified framework that learns hierarchical orthogonal prototypes with an entropy-based few-shot regularizer to enable robust novel-class adaptation without degrading base-class performance. HOP3D introduces hierarchical orthogonalization that decouples base and novel learning at both the gradient and representation levels, effectively mitigating base-novel interference. To further enhance adaptation under sparse supervision, we incorporate an entropy-based regularizer that leverages predictive uncertainty to refine prototype learning and promote balanced predictions. Extensive experiments on ScanNet200 and ScanNet++ demonstrate that HOP3D consistently outperforms state-of-the-art baselines under both 1-shot and 5-shot settings. The code is available at https://fdueblab-hop3d.github.io/.