Holistic Segmentation

Panoptic segmentation methods assign a known class to each pixel given in input. Even for state-of-the-art approaches, this inherently enforces decisions that systematically lead to wrong predictions for unknown objects that are not part of the training categories. However, in safety-critical settings, robustness against out-of-distribution samples and corner cases is crucial to avoid dangerous consequences. Since real-world datasets cannot contain enough data points to properly sample the long tail of the underlying distribution, models must be able to deal with unknown and unseen scenarios as well. Previous methods targeted this issue by re-identifying already seen unlabeled objects. In this work, we propose the necessary step to extend segmentation with a new task which we term holistic segmentation. The aim of holistic segmentation is to identify and separate objects of unseen unknown categories into instances, without any prior knowledge about them, while performing panoptic segmentation of known classes. We tackle this new problem with U3HS, which finds unknowns as highly uncertain regions, and clusters their corresponding instance-aware embeddings into individual objects. By doing so, for the first time in panoptic segmentation with unknown objects, our U3HS is not trained with unknown categories, reducing assumptions and leaving the settings as unconstrained as in real-life scenarios. Extensive experiments on publicly available data from Cityscapes and Lost&Found demonstrate the effectiveness of U3HS for the new challenging task of holistic segmentation.