Deep Learning Markov Random Field for Semantic Segmentation.

IEEE Transactions on Pattern Analysis and Machine Intelligence(2018)

引用 189|浏览189
暂无评分
摘要
Semantic segmentation tasks can be well modeled by Markov Random Field (MRF). This paper addresses semantic segmentation by incorporating high-order relations and mixture of label contexts into MRF. Unlike previous works that optimized MRFs using iterative algorithm, we solve MRF by proposing a Convolutional Neural Network (CNN), namely Deep Parsing Network (DPN), which enables deterministic end-t...
更多
查看译文
关键词
Image segmentation,Semantics,Markov random fields,Videos,Computational modeling,Neural networks,Computer architecture
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要