1 paper
Tim Meinhardt, Matt Feiszli, Yuchen Fan +2
Until recently, the Video Instance Segmentation (VIS) community operated under the common belief that offline methods are generally superior to a frame by frame online processing.…