DCARN: Deep Context Aware Recurrent Neural Network for Semantic Segmentation of Large Scale Unstructured 3D Point Cloud
- PDF / 3,125,544 Bytes
- 24 Pages / 439.37 x 666.142 pts Page_size
- 38 Downloads / 199 Views
DCARN: Deep Context Aware Recurrent Neural Network for Semantic Segmentation of Large Scale Unstructured 3D Point Cloud Saba Mehmood1 · Muhammad Shahzad1,2 · Muhammad Moazam Fraz1,3 Accepted: 4 October 2020 © Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract Semantic segmentation of large unstructured 3D point clouds is important problem for 3D object recognition which in turn is essential to solving more complex tasks such as scene understanding. The problem is highly challenging owing to large scale of data, varying point density and localization errors of 3D points. Nevertheless, with recent successes of deep neural network architectures to solve complex 2D perceptual problems, several researchers have shown interest to translate the developed 2D networks to 3D point cloud segmentation by a prior voxelization step for an explicit neighborhood representation. However, such a 3D grid representation loses the fine details and inherent structure due to quantization artifacts. For this purpose, this paper proposes an approach to performing semantic segmentation of 3D point clouds by exploiting the idea of super-point based graph construction. The proposed architecture is composed of two cascaded modules including a light-weight representation learning module which uses unsupervised geometric grouping to partition the large-scale unstructured 3D point cloud and a deep context aware sequential network based on long short memory units and graph convolutions with embedding residual learning for semantic segmentation. The proposed model is evaluated on two standard benchmark datasets and achieves competitive performance with the existing state-of-the-art datasets. The code and the obtained results have been made public at https://github.com/saba155/DCARN. Keywords 3D point clouds · Residual learning · Semantic segmentation · Mean shift clustering · Density-based spatial clustering of applications with noise (DBSCAN) · Hierarchical density-based spatial clustering of applications with noise (HDBSCAN) · Long short term memory · Edge conditioned convolution
B
Muhammad Moazam Fraz [email protected] Muhammad Shahzad [email protected]
1
National University of Sciences and Technology (NUST), Islamabad, Pakistan
2
Deep Learning Laboratory, National Center of Articial Intelligence, Islamabad 44000, Pakistan
3
The Alan Turing Institute, British Library, London NW1 2DB, UK
123
S. Mehmood et al.
1 Introduction Scene understanding is an important problem particularly in the field of robotics and computer vision having wide range of applications including navigation and perception, autonomous driving, augmented/virtual reality, surveillance and monitoring, etc. Among a variety of sensors, 3D point clouds produced using different sensor modalities, including multi-view images [1], laser scanning or LiDAR [2,3], synthetic aperture radars [4,5] etc., allows to efficiently represent the 3D characteristics of the objects within the scene. An essential step towards 3D point cloud base
Data Loading...
