Self-Contrastive Learning with Hard Negative Sampling for Self-supervised Point Cloud Learning

TL;DR

This paper introduces a self-contrastive learning method for point cloud analysis that leverages self-similar patches within a single cloud and actively learns hard negatives, achieving state-of-the-art results in segmentation and classification.

Contribution

It proposes a novel self-contrastive approach using intra-cloud self-similarity and hard negative mining for improved self-supervised point cloud learning.

Findings

Achieves state-of-the-art performance on benchmark datasets.

Effectively captures local and nonlocal geometric features.

Improves transfer learning for point cloud classification.

Abstract

Point clouds have attracted increasing attention. Significant progress has been made in methods for point cloud analysis, which often requires costly human annotation as supervision. To address this issue, we propose a novel self-contrastive learning for self-supervised point cloud representation learning, aiming to capture both local geometric patterns and nonlocal semantic primitives based on the nonlocal self-similarity of point clouds. The contributions are two-fold: on the one hand, instead of contrasting among different point clouds as commonly employed in contrastive learning, we exploit self-similar point cloud patches within a single point cloud as positive samples and otherwise negative ones to facilitate the task of contrastive learning. On the other hand, we actively learn hard negative samples that are close to positive samples for discriminative feature learning. Experimental results show that the proposed method achieves state-of-the-art performance on widely used benchmark datasets for self-supervised point cloud segmentation and transfer learning for classification.