RegionNet: Region-feature-enhanced 3D Scene Understanding Network with Dual Spatial-aware Discriminative Loss
Lists
Only you see this
Abstract
Neural networks have recently achieved impressive success in semantic and instance segmentation on 2D images. However, their capabilities have not been fully explored to address semantic instance segmentation on unstructured 3D point cloud data. Digging into the regional feature representation to boost point cloud comprehension, we propose a region-feature-enhanced structure consisting of adaptive regional feature complementary (ARFC) module and affinity-based regional relational reasoning (AR <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> ) module. The ARFC module aims to complement low-level features of sparse regions adaptively. The AR <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> module emphasizes on mining the potential reasoning relationships between high-level features based on affinity. Both the ARFC and AR <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> modules are plug-and-play. Besides, a novel dual spatial-aware discriminative loss is proposed to improve the discrimination of instance embedding. Our proposal-free point cloud instance segmentation network (RegionNet) equipped with the region-feature-enhanced structure and dual spatial-aware discriminative loss achieves state-of-the-art performance on S3DIS dataset and ScanNet-v2 dataset.
BibTeX
@inproceedings{zhang2020regionnet,
author = {Guanghui Zhang and Dongchen Zhu and Xiaoqing Ye and Wenjun Shi and Minghong Chen and Jiamao Li and Xiaolin Zhang},
title = {{RegionNet: Region-feature-enhanced 3D Scene Understanding Network with Dual Spatial-aware Discriminative Loss}},
booktitle = {IEEE/RSJ International Conference on Intelligent Robots and Systems},
year = {2020},
doi = {10.1109/iros45743.2020.9340725},
}Loading related content...