文章基本信息

标题：RGBD Based Generative Adversarial Network for 3D Semantic Scene Completion
本地全文：下载
作者：Jiahao Wang ; Ling Pei ; Danping Zou 等
期刊名称：Computer Science & Information Technology
电子版ISSN：2231-5403
出版年度：2020
卷号：10
期号：1
页码：117-125
DOI：10.5121/csit.2020.100111
出版社：Academy & Industry Research Collaboration Center (AIRCC)
摘要：3D scene understanding is of importance since it is a reflection about the real-world scenario.The goal of our work is to complete the 3d semantic scene from an RGB-D image. The state-ofthe-artmethods have poor accuracy in the face of complex scenes. In addition, other existing3D reconstruction methods use depth as the sole input, which causes performance bottlenecks.We introduce a two-stream approach that uses RGB and depth as input channels to a novelGAN architecture to solve this problem. Our method demonstrates excellent performance onboth synthetic SUNCG and real NYU dataset. Compared with the latest method SSCNet, weachieve 4.3% gains in Scene Completion (SC) and 2.5% gains in Semantic Scene Completion(SSC) on NYU dataset.
关键词：Scene Completion; Semantic Segmentation; Generation Adversarial Network; RGB;D