Exemplar-Based Open-Set Panoptic Segmentation Network
Jaedong Hwang*†
Seoung Wug Oh
Joon-Young Lee
Bohyung Han
Seoul National University
Adobe Research
CVPR 2021
* This work was done during an internship at Adobe Research


We extend panoptic segmentation to the open-world and introduce an open-set panoptic segmentation~(OPS) task. This task requires performing panoptic segmentation for not only known classes but also unknown ones that have not been acknowledged during training. We investigate the practical challenges of the task and construct a benchmark on top of an existing dataset, COCO. In addition, we propose a novel exemplar-based open-set panoptic segmentation network (EOPSN) inspired by exemplar theory. Our approach identifies a new class based on exemplars, which are identified by clustering and employed as pseudo-ground-truths. The size of each class increases by mining new exemplars based on the similarities to the existing ones associated with the class. We evaluate EOPSN on the proposed benchmark and demonstrate the effectiveness of our proposals. The primary goal of our work is to draw the attention of the community to the recognition in the open-world scenarios.


Open-Set Panoptic Segmentation results on COCO

Qualitative Results

More Results


Paper and Code

Jaedong Hwang, Seoung Wug Oh, Joon-Young Lee, Bohyung Han.
Exemplar-Based Open-Set Panoptic Segmentation Network
Proc. Computer Vision and Pattern Recognition (CVPR). 2021.
[PDF] [Code]