Self-Supervised Deep Learning on Point Clouds by Reconstructing Space

TL;DR

This paper introduces a self-supervised learning method for point clouds where the network learns to reconstruct rearranged parts, enabling better feature learning and reducing the need for labeled data in 3D tasks.

Contribution

It presents a novel self-supervised task for point clouds that improves downstream classification and reduces reliance on labeled datasets.

Findings

Outperforms existing unsupervised methods in object classification

Pre-training with our method enhances supervised model performance

Significantly improves sample efficiency in training

Abstract

Point clouds provide a flexible and natural representation usable in countless applications such as robotics or self-driving cars. Recently, deep neural networks operating on raw point cloud data have shown promising results on supervised learning tasks such as object classification and semantic segmentation. While massive point cloud datasets can be captured using modern scanning technology, manually labelling such large 3D point clouds for supervised learning tasks is a cumbersome process. This necessitates methods that can learn from unlabelled data to significantly reduce the number of annotated samples needed in supervised learning. We propose a self-supervised learning task for deep learning on raw point cloud data in which a neural network is trained to reconstruct point clouds whose parts have been randomly rearranged. While solving this task, representations that capture semantic properties of the point cloud are learned. Our method is agnostic of network architecture and outperforms current unsupervised learning approaches in downstream object classification tasks. We show experimentally, that pre-training with our method before supervised training improves the performance of state-of-the-art models and significantly improves sample efficiency.