全景分割研究综述

在计算机视觉领域, 全景分割是一个新颖且重要的研究主题, 它是机器感知、自动驾驶等新兴前沿技术的基石, 具有十分重要的研究意义. 本文综述了基于深度学习的全景分割研究的最新进展, 首先总结了全景分割任务的基本处理流程, 然后对已发表的全景分割工作基于其网络结构特点进行分类, 并进行了全面的介绍与分析, 最后对全景分割任务目前面临的问题以及未来的发展趋势做出了分析, 并针对所面临的问题提出了一些切实可行的解决思路.

A Survey of Panoptic Segmentation Methods

In the field of computer vision,panoptic segmentation is a novel and important research topic.It is the cornerstone of emerging leading-edge technologies such as machine perception and autonomous driving,and has very important significance of research.This paper reviews the research progress of panoptic segmentation based on deep learning methods,summarizes the basic processing flow of panoptic segmentation,and then classifies the already published panoptic segmentation works based on the characteristics of its network structure,and makes a more comprehensive introduction and analysis.Finally,the current problems and future development trends of the panoptic segmentation task are analyzed,and some potential solutions are provided for solving the existing problems.