Bidirectional Feature Globalization for Few-shot Semantic Segmentation of 3D Point Cloud Scenes

by   Yongqiang Mao, et al.
University of the Chinese Academy of Sciences
Wuhan University

Few-shot segmentation of point cloud remains a challenging task, as there is no effective way to convert local point cloud information to global representation, which hinders the generalization ability of point features. In this study, we propose a bidirectional feature globalization (BFG) approach, which leverages the similarity measurement between point features and prototype vectors to embed global perception to local point features in a bidirectional fashion. With point-to-prototype globalization (Po2PrG), BFG aggregates local point features to prototypes according to similarity weights from dense point features to sparse prototypes. With prototype-to-point globalization (Pr2PoG), the global perception is embedded to local point features based on similarity weights from sparse prototypes to dense point features. The sparse prototypes of each class embedded with global perception are summarized to a single prototype for few-shot 3D segmentation based on the metric learning framework. Extensive experiments on S3DIS and ScanNet demonstrate that BFG significantly outperforms the state-of-the-art methods.


page 7

page 8


Number-Adaptive Prototype Learning for 3D Point Cloud Semantic Segmentation

3D point cloud semantic segmentation is one of the fundamental tasks for...

A Unified Point-Based Framework for 3D Segmentation

3D point cloud segmentation remains challenging for structureless and te...

GDDS: Pulmonary Bronchioles Segmentation with Group Deep Dense Supervision

Airway segmentation, especially bronchioles segmentation, is an importan...

Dense-Resolution Network for Point Cloud Classification and Segmentation

Point cloud analysis is attracting attention from Artificial Intelligenc...

Point Cloud Segmentation Using Sparse Temporal Local Attention

Point clouds are a key modality used for perception in autonomous vehicl...

Global-Local Bidirectional Reasoning for Unsupervised Representation Learning of 3D Point Clouds

Local and global patterns of an object are closely related. Although eac...

CAM/CAD Point Cloud Part Segmentation via Few-Shot Learning

3D part segmentation is an essential step in advanced CAM/CAD workflow. ...

Please sign up or login with your details

Forgot password? Click here to reset