VOS
摘要
In this chapter, we elaborate on the task of video object segmentation (VOS), which aims at dividing pixels of a video into disjoint subsets where each subset usually represents either a target object or the background. The VOS task has different problem settings given different input or output requirements.