LogoSP: Local-global Grouping of Superpoints for Unsupervised Semantic Segmentation of 3D Point Clouds

TL;DR

LogoSP introduces a novel unsupervised 3D semantic segmentation method that leverages global pattern grouping of superpoints in the frequency domain to achieve state-of-the-art results without human labels.

Contribution

It proposes a new approach that combines local and global features through frequency domain grouping to improve unsupervised 3D semantic segmentation.

Findings

Outperforms existing unsupervised methods by large margins

Achieves state-of-the-art performance on multiple datasets

Global patterns learned are meaningful for 3D semantics

Abstract

We study the problem of unsupervised 3D semantic segmentation on raw point clouds without needing human labels in training. Existing methods usually formulate this problem into learning per-point local features followed by a simple grouping strategy, lacking the ability to discover additional and possibly richer semantic priors beyond local features. In this paper, we introduce LogoSP to learn 3D semantics from both local and global point features. The key to our approach is to discover 3D semantic information by grouping superpoints according to their global patterns in the frequency domain, thus generating highly accurate semantic pseudo-labels for training a segmentation network. Extensive experiments on two indoor and an outdoor datasets show that our LogoSP surpasses all existing unsupervised methods by large margins, achieving the state-of-the-art performance for unsupervised 3D semantic segmentation. Notably, our investigation into the learned global patterns reveals that they truly represent meaningful 3D semantics in the absence of human labels during training.