Distribution Guidance Network for Weakly Supervised Point Cloud Semantic Segmentation

TL;DR

This paper introduces DGNet, a novel network that uses distribution guidance with moVMF to improve weakly supervised point cloud segmentation, achieving state-of-the-art results by aligning feature space distributions.

Contribution

The paper proposes a new distribution guidance approach using moVMF for weakly supervised point cloud segmentation, with a novel network architecture and distribution alignment strategy.

Findings

DGNet outperforms existing methods on multiple datasets.

The moVMF distribution effectively models feature space in weak supervision.

Distribution alignment improves segmentation accuracy.

Abstract

Despite alleviating the dependence on dense annotations inherent to fully supervised methods, weakly supervised point cloud semantic segmentation suffers from inadequate supervision signals. In response to this challenge, we introduce a novel perspective that imparts auxiliary constraints by regulating the feature space under weak supervision. Our initial investigation identifies which distributions accurately characterize the feature space, subsequently leveraging this priori to guide the alignment of the weakly supervised embeddings. Specifically, we analyze the superiority of the mixture of von Mises-Fisher distributions (moVMF) among several common distribution candidates. Accordingly, we develop a Distribution Guidance Network (DGNet), which comprises a weakly supervised learning branch and a distribution alignment branch. Leveraging reliable clustering initialization derived from the weakly supervised learning branch, the distribution alignment branch alternately updates the parameters of the moVMF and the network, ensuring alignment with the moVMF-defined latent space. Extensive experiments validate the rationality and effectiveness of our distribution choice and network design. Consequently, DGNet achieves state-of-the-art performance under multiple datasets and various weakly supervised settings.