视频对象分割方法(可用作实验对比):FST,FSEG,SFL,LVO,ARP
1、FST
(ICCV2013: Fast object segmentation in unconstrained video)
两个主要阶段:
(1)Efficient initial foreground estimation有效的初始前景估计
(2)Foreground-background labelling refinement前景-背景标签细化
官方代码链接:http://groups.inf.ed.ac.uk/calvin/software.html
2、FSEG: FunctionSeg
(CVPR2017: FusionSeg: Learning to combine motion and appearance for fully automatic segmentation of generic objects in videos)
官方代码链接:https://github.com/suyogduttjain/fusionseg
3、SFL: SegFlow
(ICCV2017: SegFlow: Joint Learning for Video Object Segmentation and Optical Flow)
官方代码链接:https://github.com/JingchunCheng/SegFlow
4、LVO
(ICCV2017: Learning Video Object Segmentation with Visual Memory)
5、ARP: Augmentation and Reduction Process
(CVPR2017: Primary Object Segmentation in Videos Based on Region Augmentation and Reduction)