COCO-OLAC

Emerging

1papers using it

2024first seen

COCO-OLAC is a large-scale dataset derived from the COCO dataset, containing images manually labeled with three levels of perceived occlusion, used to evaluate the impact of occlusion on panoptic segmentation and image understanding.

🔎 Find this dataset

Papers using COCO-OLAC (1)

COCO-OLAC: A Benchmark for Occluded Panoptic Segmentation and Image Understanding2024