图像语义分割方法综述  

A Review on Image Semantic Segmentation

作　　者：胡涛 李卫华[1] 秦先祥[1] HU Tao;LI Wei-hua;QIN Xian-xiang(Information and Navigation College,Air Force Engineering University,Xi'an 710077,China)

机构地区：[1]空军工程大学信息与导航学院

出　　处：《测控技术》2019年第7期8-12,共5页Measurement & Control Technology

基　　金：陕西省自然科学基础研究计划(2018JM4029);国家自然科学基金(41601436,61403414,61703423)

摘　　要：图像语义分割是一种将图像分割成一系列具有特定语义类别属性的区域的方法,目前已成为当前图像理解分析和计算机视觉等领域的热点研究内容。在广泛文献调研的基础上,简述了图像语义分割方法的发展历程,从驱动类型的角度将当前主流的图像语义分割方法分为基于模型驱动的方法和基于数据驱动的方法,并分别进行了介绍,总结归纳了图像语义分割方法的性能评价标准,指出了图像语义分割方法的主要发展趋势。Image semantic segmentation is a method of segmenting an image into a series of regions with specific semantic category attributes.It has become a hot research topic in the fields of image understanding analysis and computer vision.On the basis of extensive literature research,the development history of image semantic segmentation method was briefly introduced.The current mainstream image semantic segmentation methods were divided into model-driven method and data-driven method from the perspective of driver type,and intr-duced separately.The performance evaluation criteria of image semantic segmentation method was summarized,and the main development trend of image semantic segmentation method was pointed out.

关 键 词：图像语义分割 卷积神经网络 图像标记 综述 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]