基于弱监督学习的图像语义分割方法综述

Review of Image Semantic Segmentation Methods Based on Weakly Supervised Learning

为了解决全监督语义分割网络训练成本高的问题,研究者们提出了基于弱监督学习下的语义分割方法。文章对弱监督学习的语义分割方法进行综述,并介绍了语义分割领域常用的数据集和评价指标,最后提出了弱监督语义分割的发展方向。

In order to solve the problem of high training cost of fully supervised semantic segmentation network,researchers proposed a semantic segmentation method based on weakly supervised learning.This paper summarizes the semantic segmentation methods of weakly supervised learning,introduces the data sets and evaluation indexes commonly used in the field of semantic segmentation,and finally puts forward the development direction of weakly supervised semantic segmentation.