Saliency detection using a deep conditional random field network
作者：
Highlights：
• Designed a multi-scale backward optimization network which can hold both rich inherent features from shallower layers and semantic features from deeper layers, then high-level features are transmitted backward to guiding low-level features.
• Deep CRF network is introduced to form the relationships between adjacent pixels which is crucial to improve the quality of saliency maps.
• All modules can be conveniently embedded into other Convolutional Neural Networks for feature and relation representation.
摘要
•Designed a multi-scale backward optimization network which can hold both rich inherent features from shallower layers and semantic features from deeper layers, then high-level features are transmitted backward to guiding low-level features.•Deep CRF network is introduced to form the relationships between adjacent pixels which is crucial to improve the quality of saliency maps.•All modules can be conveniently embedded into other Convolutional Neural Networks for feature and relation representation.
论文关键词：Saliency detection,Conditional random field,Convolutional neural network
论文评审过程：Received 24 February 2019, Revised 2 February 2020, Accepted 8 February 2020, Available online 19 February 2020, Version of Record 5 March 2020.
论文官网地址：https://doi.org/10.1016/j.patcog.2020.107266
