3dmax-Net: A Multi-Scale Spatial Contextual Network For 3d Point Cloud Semantic Segmentation
2018 24TH INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION (ICPR)(2018)
摘要
Semantic segmentation of 3D scenes is a fundamental problem in 3D computer vision. In this paper, we propose a deep neural network for 3D semantic segmentation of raw point clouds. A multi-scale feature learning block is first introduced to obtain informative contextual features in 3D point clouds. A global and local feature aggregation block is then extended to improve the feature learning ability of the network. Based on these strategies, a powerful architecture named 3DMAX-Net is finally provided for semantic segmentation in raw 3D point clouds. Experiments have been conducted on the Stanford large-scale 3D Indoor Spaces Dataset using only geometry information. Experimental results have clearly shown the superiority of the proposed network.
更多查看译文
关键词
3DMAX-Net,multiscale spatial contextual network,3D point cloud semantic segmentation,3D computer vision,deep neural network,raw point clouds,multiscale feature,informative contextual features,global feature aggregation block,local feature aggregation block,feature learning ability,raw 3D point clouds,Stanford large-scale 3D Indoor Spaces Dataset
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络