Abstract:With the proliferation of Lidar sensors and 3D vision cameras, 3D point cloud analysis has attracted significant attention in recent years. After the success of the pioneer work PointNet, deep learning-based methods have been increasingly applied to various tasks, including 3D point cloud segmentation and 3D object classification. In this paper, we propose a novel 3D point cloud learning network, referred to as Dynamic Point Feature Aggregation Network (DPFA-Net), by selectively performing the neighborhood feature aggregation with dynamic pooling and an attention mechanism. DPFA-Net has two variants for semantic segmentation and classification of 3D point clouds. As the core module of the DPFA-Net, we propose a Feature Aggregation layer, in which features of the dynamic neighborhood of each point are aggregated via a self-attention mechanism. In contrast to other segmentation models, which aggregate features from fixed neighborhoods, our approach can aggregate features from different neighbors in different layers providing a more selective and broader view to the query points, and focusing more on the relevant features in a local neighborhood. In addition, to further improve the performance of the proposed semantic segmentation model, we present two novel approaches, namely Two-Stage BF-Net and BF-Regularization to exploit the background-foreground information. Experimental results show that the proposed DPFA-Net achieves the state-of-the-art overall accuracy score for semantic segmentation on the S3DIS dataset, and provides a consistently satisfactory performance across different tasks of semantic segmentation, part segmentation, and 3D object classification. It is also computationally more efficient compared to other methods.

Semantic Feature Mining for 3D Object Classification and Segmentation.

Associate Semantic-Instance Segmentation of 3D Point Clouds Based on Local Feature Extraction

Using Multi-Scale and Hierarchical Deep Convolutional Features for 3D Semantic Classification of TLS Point Clouds.

3D Semantic Segmentation Using Deep Learning for Large-Scale Indoor Point Cloud

PointMS: Semantic Segmentation for Point Cloud Based on Multi-scale Directional Convolution

Pointwise Geometric and Semantic Learning Network on 3D Point Clouds.

Superpoint-guided Semi-supervised Semantic Segmentation of 3D Point Clouds

A Point-Based Deep Learning Network for Semantic Segmentation of MLS Point Clouds

A Multi-scale Network for Semantic Segmentation of 3D Point Clouds

Fuzzy Neighborhood Learning for Deep 3-D Segmentation of Point Cloud.

Deep Projective 3D Semantic Segmentation

Voxel-based 3D Point Cloud Semantic Segmentation: Unsupervised Geometric and Relationship Featuring vs Deep Learning Methods

Semantic Segmentation of 3D Point Clouds Based on High Precision Range Search Network

Background-Aware 3D Point Cloud Segmentationwith Dynamic Point Feature Aggregation

Using Deep Learning in Semantic Classification for Point Cloud Data

Semantic Classification of 3D Point Clouds with Multiscale Spherical Neighborhoods

Mining Local Geometric Structure for Large-Scale 3D Point Clouds Semantic Segmentation

Spherical Interpolated Convolutional Network with Distance-Feature Density for 3D Semantic Segmentation of Point Clouds

3D Semantic Segmentation of Large-Scale Point-Clouds in Urban Areas Using Deep Learning

Spherical Interpolated Convolutional Network With Distance-Feature Density for 3-D Semantic Segmentation of Point Clouds

Deep Learning-Based 3D Instance and Semantic Segmentation: A Review