Video Object Segmentation
The problem of video object segmentation (VOS) has a variety of important applications, including object boundary estimation for grasping, autonomous driving, surveillance and video editing. The task is to predict pixel-accurate masks of the region occupied by a specific target object, in every frame of a given video sequence. Our research work focuses on the semi-supervised setting, where a target ground truth mask is provided in the first frame. The main difficulty is to effectively handle appearance changes and similar background objects, while maintaining accurate segmentation. We strive to develop novel deep learning-based VOS models that can hangle large appearance changes, are fast, easily trainable and remains highly effective in cases of limited training data. .