SAIL-VOS: Semantic Amodal Instance Level Video Object Segmentation – A Synthetic Dataset and Baselines | IEEE Conference Publication | IEEE Xplore