Self-Supervised Learning for Domain Adaptation on Point-Clouds

Abstract

Self-supervised learning (SSL) is a technique for learning useful representations from unlabeled data. It has been applied effectively to domain adaptation (DA) on images and videos. It is still unknown if and how it can be leveraged for domain adaptation in 3D perception problems. Here we describe the first study of SSL for DA on point clouds. We introduce a new family of pretext tasks, Deformation Reconstruction, inspired by the deformations encountered in sim-to-real transformations. In addition, we propose a novel training procedure for labeled point cloud data motivated by the MixUp method called Point cloud Mixup (PCM). Evaluations on domain adaptations datasets for classification and segmentation, demonstrate a large improvement over existing and baseline methods.

References in corpus (16)

• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• Prototypical Networks for Few-shot Learning
• Deep Domain Confusion: Maximizing for Domain Invariance
• Learning Representations and Generative Models for 3D Point Clouds
• ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenes
• Unsupervised Domain Adaptation through Self-Supervision
• Self-supervised Domain Adaptation for Computer Vision Tasks
• Universal Domain Adaptation through Self Supervision
• Improve Unsupervised Domain Adaptation with Mixup Training
• Unpaired Point Cloud Completion on Real Scans using Adversarial Training
• PointDAN: A Multi-Scale 3D Domain Adaption Network for Point Cloud Representation
• Virtual Mixup Training for Unsupervised Domain Adaptation
• Complete & Label: A Domain Adaptation Approach to Semantic Segmentation of LiDAR Point Clouds
• MortonNet: Self-Supervised Learning of Local Features in 3D Point Clouds
• Improving Semantic Analysis on Point Clouds via Auxiliary Supervision of Local Geometric Priors
• Unsupervised Feature Learning for Point Cloud by Contrasting and Clustering With Graph Convolutional Neural Network

Cited by in corpus (10)

• Point-BERT: Pre-training 3D Point Cloud Transformers with Masked Point Modeling
• A Review of Single-Source Deep Unsupervised Visual Domain Adaptation
• See Eye to Eye: A Lidar-Agnostic 3D Detection Framework for Unsupervised Multi-Target Domain Adaptation
• Prototypical Cross-domain Self-supervised Learning for Few-shot Unsupervised Domain Adaptation
• A Learnable Self-supervised Task for Unsupervised Domain Adaptation on Point Clouds
• SF-UDA${}^{3D}$: Source-Free Unsupervised Domain Adaptation for LiDAR-Based 3D Object Detection
• From Local Structures to Size Generalization in Graph Neural Networks
• Auxiliary Learning by Implicit Differentiation
• Multi-source Few-shot Domain Adaptation
• R-AGNO-RPN: A LIDAR-Camera Region Deep Network for Resolution-Agnostic Detection