Dual-level Adaptive Self-Labeling for Novel Class Discovery in Point Cloud Segmentation

TL;DR

This paper introduces a dual-level self-labeling approach for point cloud segmentation that adaptively handles class imbalance and leverages spatial context, significantly improving novel class discovery performance.

Contribution

It proposes a novel adaptive self-labeling strategy and dual-level representation to better discover and segment novel classes in point clouds, addressing class imbalance and spatial context issues.

Findings

Outperforms state-of-the-art on SemanticKITTI and SemanticPOSS datasets.

Effectively handles class imbalance in point cloud segmentation.

Enhances semantic segmentation accuracy with dual-level spatial context.

Abstract

We tackle the novel class discovery in point cloud segmentation, which discovers novel classes based on the semantic knowledge of seen classes. Existing work proposes an online point-wise clustering method with a simplified equal class-size constraint on the novel classes to avoid degenerate solutions. However, the inherent imbalanced distribution of novel classes in point clouds typically violates the equal class-size constraint. Moreover, point-wise clustering ignores the rich spatial context information of objects, which results in less expressive representation for semantic segmentation. To address the above challenges, we propose a novel self-labeling strategy that adaptively generates high-quality pseudo-labels for imbalanced classes during model training. In addition, we develop a dual-level representation that incorporates regional consistency into the point-level classifier learning, reducing noise in generated segmentation. Finally, we conduct extensive experiments on two widely used datasets, SemanticKITTI and SemanticPOSS, and the results show our method outperforms the state of the art by a large margin.