COARSE3D: Class-Prototypes for Contrastive Learning in Weakly-Supervised 3D Point Cloud Segmentation
Preprint
Abstract
Annotation of large-scale 3D data is notoriously cumbersome and costly. As an alternative, weakly-supervised learning alleviates such a need by reducing the annotation by several order of magnitudes. We propose COARSE3D, a novel architecture-agnostic contrastive learning strategy for 3D segmentation. Since contrastive learning requires rich and diverse examples as keys and anchors, we leverage a prototype memory bank capturing class-wise global dataset information efficiently into a small number of prototypes acting as keys. An entropy-driven sampling technique then allows us to select good pixels from predictions as anchors. Experiments on three projection-based backbones show we outperform baselines on three challenging real-world outdoor datasets, working with as low as 0.001% annotations.Keywords
All Related Versions
This publication has 63 references indexed in Scilit:
- 4DContrast: Contrastive Learning with Dynamic Correspondences for 3D Scene Understanding2022
- Language-Grounded Indoor 3D Semantic Segmentation in the Wild2022
- Attribute Prototype Network for Any-Shot LearningInternational Journal of Computer Vision, 2022
- SalsaNext: Fast, Uncertainty-Aware Semantic Segmentation of LiDAR Point Clouds2020
- PointContrast: Unsupervised Pre-training for 3D Point Cloud Understanding2020
- Searching Efficient 3D Architectures with Sparse Point-Voxel Convolution2020
- SqueezeSegV3: Spatially-Adaptive Convolution for Efficient Point-Cloud Segmentation2020
- Improving Generalization via Scalable Neighborhood Component Analysis2018
- Hard Negative Mining for Metric Learning Based Zero-Shot Classification2016
- Improving Semantic Embedding Consistency by Metric Learning for Zero-Shot Classiffication2016